|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
|
- <!DOCTYPE html>
- <html class="writer-html5" lang="en" >
- <head>
- <meta charset="utf-8" />
- <meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>super_gradients.training.transforms.transforms — SuperGradients 3.0.3 documentation</title>
- <link rel="stylesheet" href="../../../../_static/pygments.css" type="text/css" />
- <link rel="stylesheet" href="../../../../_static/css/theme.css" type="text/css" />
- <link rel="stylesheet" href="../../../../_static/graphviz.css" type="text/css" />
- <link rel="stylesheet" href="../../../../_static/custom.css" type="text/css" />
- <!--[if lt IE 9]>
- <script src="../../../../_static/js/html5shiv.min.js"></script>
- <![endif]-->
-
- <script data-url_root="../../../../" id="documentation_options" src="../../../../_static/documentation_options.js"></script>
- <script src="../../../../_static/jquery.js"></script>
- <script src="../../../../_static/underscore.js"></script>
- <script src="../../../../_static/_sphinx_javascript_frameworks_compat.js"></script>
- <script src="../../../../_static/doctools.js"></script>
- <script src="../../../../_static/sphinx_highlight.js"></script>
- <script src="../../../../_static/js/theme.js"></script>
- <link rel="index" title="Index" href="../../../../genindex.html" />
- <link rel="search" title="Search" href="../../../../search.html" />
- </head>
- <body class="wy-body-for-nav">
- <div class="wy-grid-for-nav">
- <nav data-toggle="wy-nav-shift" class="wy-nav-side">
- <div class="wy-side-scroll">
- <div class="wy-side-nav-search" >
- <a href="../../../../index.html" class="icon icon-home"> SuperGradients
- </a>
- <div role="search">
- <form id="rtd-search-form" class="wy-form" action="../../../../search.html" method="get">
- <input type="text" name="q" placeholder="Search docs" />
- <input type="hidden" name="check_keywords" value="yes" />
- <input type="hidden" name="area" value="default" />
- </form>
- </div>
- </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
- <p class="caption" role="heading"><span class="caption-text">Welcome To SuperGradients</span></p>
- <ul>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html">Version 3 is out! Notebooks have been updated!</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#build-with-supergradients">Build with SuperGradients</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#quick-installation">Quick Installation</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#what-s-new">What’s New</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#coming-soon">Coming soon</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#table-of-content">Table of Content</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#getting-started">Getting Started</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#advanced-features">Advanced Features</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#installation-methods">Installation Methods</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#implemented-model-architectures">Implemented Model Architectures</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#documentation">Documentation</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#contributing">Contributing</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#citation">Citation</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#community">Community</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#license">License</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#deci-platform">Deci Platform</a></li>
- </ul>
- <p class="caption" role="heading"><span class="caption-text">Technical Documentation</span></p>
- <ul>
- <li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.common.html">Common package</a></li>
- <li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.training.html">Training package</a></li>
- </ul>
- </div>
- </div>
- </nav>
- <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
- <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
- <a href="../../../../index.html">SuperGradients</a>
- </nav>
- <div class="wy-nav-content">
- <div class="rst-content">
- <div role="navigation" aria-label="Page navigation">
- <ul class="wy-breadcrumbs">
- <li><a href="../../../../index.html" class="icon icon-home"></a> »</li>
- <li><a href="../../../index.html">Module code</a> »</li>
- <li>super_gradients.training.transforms.transforms</li>
- <li class="wy-breadcrumbs-aside">
- </li>
- </ul>
- <hr/>
- </div>
- <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
- <div itemprop="articleBody">
-
- <h1>Source code for super_gradients.training.transforms.transforms</h1><div class="highlight"><pre>
- <span></span><span class="kn">import</span> <span class="nn">collections</span>
- <span class="kn">import</span> <span class="nn">math</span>
- <span class="kn">import</span> <span class="nn">random</span>
- <span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">Optional</span><span class="p">,</span> <span class="n">Union</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">,</span> <span class="n">Sequence</span><span class="p">,</span> <span class="n">Dict</span>
- <span class="kn">from</span> <span class="nn">PIL</span> <span class="kn">import</span> <span class="n">Image</span><span class="p">,</span> <span class="n">ImageFilter</span><span class="p">,</span> <span class="n">ImageOps</span>
- <span class="kn">from</span> <span class="nn">torchvision</span> <span class="kn">import</span> <span class="n">transforms</span> <span class="k">as</span> <span class="n">transforms</span>
- <span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
- <span class="kn">import</span> <span class="nn">cv2</span>
- <span class="kn">from</span> <span class="nn">super_gradients.common.abstractions.abstract_logger</span> <span class="kn">import</span> <span class="n">get_logger</span>
- <span class="kn">from</span> <span class="nn">super_gradients.training.utils.detection_utils</span> <span class="kn">import</span> <span class="n">get_mosaic_coordinate</span><span class="p">,</span> <span class="n">adjust_box_anns</span><span class="p">,</span> <span class="n">xyxy2cxcywh</span><span class="p">,</span> <span class="n">cxcywh2xyxy</span><span class="p">,</span> <span class="n">DetectionTargetsFormat</span>
- <span class="n">image_resample</span> <span class="o">=</span> <span class="n">Image</span><span class="o">.</span><span class="n">BILINEAR</span>
- <span class="n">mask_resample</span> <span class="o">=</span> <span class="n">Image</span><span class="o">.</span><span class="n">NEAREST</span>
- <span class="n">logger</span> <span class="o">=</span> <span class="n">get_logger</span><span class="p">(</span><span class="vm">__name__</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegmentationTransform</span><span class="p">:</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="o">*</span><span class="n">args</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span>
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__dict__</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s2">"{"</span><span class="p">,</span> <span class="s2">"("</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s2">"}"</span><span class="p">,</span> <span class="s2">")"</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegResize</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">h</span><span class="p">,</span> <span class="n">w</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">h</span> <span class="o">=</span> <span class="n">h</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">w</span> <span class="o">=</span> <span class="n">w</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">resize</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">w</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">h</span><span class="p">),</span> <span class="n">image_resample</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">resize</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">w</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">h</span><span class="p">),</span> <span class="n">mask_resample</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">class</span> <span class="nc">SegRandomFlip</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Randomly flips the image and mask (synchronously) with probability 'prob'.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">):</span>
- <span class="k">assert</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">prob</span> <span class="o"><=</span> <span class="mf">1.0</span><span class="p">,</span> <span class="sa">f</span><span class="s2">"Probability value must be between 0 and 1, found </span><span class="si">{</span><span class="n">prob</span><span class="si">}</span><span class="s2">"</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="k">if</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">transpose</span><span class="p">(</span><span class="n">Image</span><span class="o">.</span><span class="n">FLIP_LEFT_RIGHT</span><span class="p">)</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">transpose</span><span class="p">(</span><span class="n">Image</span><span class="o">.</span><span class="n">FLIP_LEFT_RIGHT</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">class</span> <span class="nc">SegRescale</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Rescales the image and mask (synchronously) while preserving aspect ratio.</span>
- <span class="sd"> The rescaling can be done according to scale_factor, short_size or long_size.</span>
- <span class="sd"> If more than one argument is given, the rescaling mode is determined by this order: scale_factor, then short_size,</span>
- <span class="sd"> then long_size.</span>
- <span class="sd"> Args:</span>
- <span class="sd"> scale_factor: rescaling is done by multiplying input size by scale_factor:</span>
- <span class="sd"> out_size = (scale_factor * w, scale_factor * h)</span>
- <span class="sd"> short_size: rescaling is done by determining the scale factor by the ratio short_size / min(h, w).</span>
- <span class="sd"> long_size: rescaling is done by determining the scale factor by the ratio long_size / max(h, w).</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scale_factor</span><span class="p">:</span> <span class="n">Optional</span><span class="p">[</span><span class="nb">float</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span> <span class="n">short_size</span><span class="p">:</span> <span class="n">Optional</span><span class="p">[</span><span class="nb">int</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span> <span class="n">long_size</span><span class="p">:</span> <span class="n">Optional</span><span class="p">[</span><span class="nb">int</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span> <span class="o">=</span> <span class="n">scale_factor</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="o">=</span> <span class="n">short_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">long_size</span> <span class="o">=</span> <span class="n">long_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">check_valid_arguments</span><span class="p">()</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">w</span><span class="p">,</span> <span class="n">h</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">size</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span>
- <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">short_size</span> <span class="o">=</span> <span class="nb">min</span><span class="p">(</span><span class="n">w</span><span class="p">,</span> <span class="n">h</span><span class="p">)</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="o">/</span> <span class="n">short_size</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">long_size</span> <span class="o">=</span> <span class="nb">max</span><span class="p">(</span><span class="n">w</span><span class="p">,</span> <span class="n">h</span><span class="p">)</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">long_size</span> <span class="o">/</span> <span class="n">long_size</span>
- <span class="n">out_size</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">scale</span> <span class="o">*</span> <span class="n">w</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">scale</span> <span class="o">*</span> <span class="n">h</span><span class="p">)</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">out_size</span><span class="p">,</span> <span class="n">image_resample</span><span class="p">)</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">out_size</span><span class="p">,</span> <span class="n">mask_resample</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">check_valid_arguments</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span> <span class="ow">is</span> <span class="kc">None</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="ow">is</span> <span class="kc">None</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">long_size</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Must assign one rescale argument: scale_factor, short_size or long_size"</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span> <span class="o"><=</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Scale factor must be a positive number, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">scale_factor</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">short_size</span> <span class="o"><=</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Short size must be a positive number, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">short_size</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">long_size</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">long_size</span> <span class="o"><=</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Long size must be a positive number, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">long_size</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegRandomRescale</span><span class="p">:</span>
- <span class="sd">"""</span>
- <span class="sd"> Random rescale the image and mask (synchronously) while preserving aspect ratio.</span>
- <span class="sd"> Scale factor is randomly picked between scales [min, max]</span>
- <span class="sd"> Args:</span>
- <span class="sd"> scales: scale range tuple (min, max), if scales is a float range will be defined as (1, scales) if scales > 1,</span>
- <span class="sd"> otherwise (scales, 1). must be a positive number.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">scales</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="mf">0.5</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">)):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scales</span> <span class="o">=</span> <span class="n">scales</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">check_valid_arguments</span><span class="p">()</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">w</span><span class="p">,</span> <span class="n">h</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">size</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="n">out_size</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">scale</span> <span class="o">*</span> <span class="n">w</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">scale</span> <span class="o">*</span> <span class="n">h</span><span class="p">)</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">out_size</span><span class="p">,</span> <span class="n">image_resample</span><span class="p">)</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">out_size</span><span class="p">,</span> <span class="n">mask_resample</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">check_valid_arguments</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Check the scale values are valid. if order is wrong, flip the order and return the right scale values.</span>
- <span class="sd"> """</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">,</span> <span class="n">collections</span><span class="o">.</span><span class="n">abc</span><span class="o">.</span><span class="n">Iterable</span><span class="p">):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span> <span class="o"><=</span> <span class="mi">1</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scales</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scales</span> <span class="o">=</span> <span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o"><</span> <span class="mi">0</span> <span class="ow">or</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o"><</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"SegRandomRescale scale values must be positive numbers, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">></span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">1</span><span class="p">]:</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scales</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">scales</span>
- <span class="k">class</span> <span class="nc">SegRandomRotate</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Randomly rotates image and mask (synchronously) between 'min_deg' and 'max_deg'.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">min_deg</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="o">-</span><span class="mi">10</span><span class="p">,</span> <span class="n">max_deg</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">10</span><span class="p">,</span> <span class="n">fill_mask</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">int</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">]</span> <span class="o">=</span> <span class="mi">0</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">min_deg</span> <span class="o">=</span> <span class="n">min_deg</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_deg</span> <span class="o">=</span> <span class="n">max_deg</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span> <span class="o">=</span> <span class="n">fill_mask</span>
- <span class="c1"># grey color in RGB mode</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span> <span class="o">=</span> <span class="p">(</span><span class="n">fill_image</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">check_valid_arguments</span><span class="p">()</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">deg</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">min_deg</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">max_deg</span><span class="p">)</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">rotate</span><span class="p">(</span><span class="n">deg</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">image_resample</span><span class="p">,</span> <span class="n">fillcolor</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span><span class="p">)</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">rotate</span><span class="p">(</span><span class="n">deg</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">mask_resample</span><span class="p">,</span> <span class="n">fillcolor</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">check_valid_arguments</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span> <span class="o">=</span> <span class="n">_validate_fill_values_arguments</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegCropImageAndMask</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Crops image and mask (synchronously).</span>
- <span class="sd"> In "center" mode a center crop is performed while, in "random" mode the drop will be positioned around</span>
- <span class="sd"> random coordinates.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">crop_size</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">],</span> <span class="n">mode</span><span class="p">:</span> <span class="nb">str</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> :param crop_size: tuple of (width, height) for the final crop size, if is scalar size is a</span>
- <span class="sd"> square (crop_size, crop_size)</span>
- <span class="sd"> :param mode: how to choose the center of the crop, 'center' for the center of the input image,</span>
- <span class="sd"> 'random' center the point is chosen randomally</span>
- <span class="sd"> """</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span> <span class="o">=</span> <span class="n">crop_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">=</span> <span class="n">mode</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">check_valid_arguments</span><span class="p">()</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">w</span><span class="p">,</span> <span class="n">h</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">size</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s2">"random"</span><span class="p">:</span>
- <span class="n">x1</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">w</span> <span class="o">-</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
- <span class="n">y1</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">h</span> <span class="o">-</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">x1</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="nb">round</span><span class="p">((</span><span class="n">w</span> <span class="o">-</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span> <span class="o">/</span> <span class="mf">2.0</span><span class="p">))</span>
- <span class="n">y1</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="nb">round</span><span class="p">((</span><span class="n">h</span> <span class="o">-</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span> <span class="o">/</span> <span class="mf">2.0</span><span class="p">))</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">crop</span><span class="p">((</span><span class="n">x1</span><span class="p">,</span> <span class="n">y1</span><span class="p">,</span> <span class="n">x1</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">y1</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]))</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">mask</span><span class="o">.</span><span class="n">crop</span><span class="p">((</span><span class="n">x1</span><span class="p">,</span> <span class="n">y1</span><span class="p">,</span> <span class="n">x1</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">y1</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]))</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">check_valid_arguments</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="ow">not</span> <span class="ow">in</span> <span class="p">[</span><span class="s2">"center"</span><span class="p">,</span> <span class="s2">"random"</span><span class="p">]:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Unsupported mode: found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">mode</span><span class="si">}</span><span class="s2">, expected: 'center' or 'random'"</span><span class="p">)</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">,</span> <span class="n">collections</span><span class="o">.</span><span class="n">abc</span><span class="o">.</span><span class="n">Iterable</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o"><=</span> <span class="mi">0</span> <span class="ow">or</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o"><=</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Crop size must be positive numbers, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegRandomGaussianBlur</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Adds random Gaussian Blur to image with probability 'prob'.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">):</span>
- <span class="k">assert</span> <span class="mf">0.0</span> <span class="o"><=</span> <span class="n">prob</span> <span class="o"><=</span> <span class="mf">1.0</span><span class="p">,</span> <span class="s2">"Probability value must be between 0 and 1"</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="k">if</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">filter</span><span class="p">(</span><span class="n">ImageFilter</span><span class="o">.</span><span class="n">GaussianBlur</span><span class="p">(</span><span class="n">radius</span><span class="o">=</span><span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()))</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">class</span> <span class="nc">SegPadShortToCropSize</span><span class="p">(</span><span class="n">SegmentationTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Pads image to 'crop_size'.</span>
- <span class="sd"> Should be called only after "SegRescale" or "SegRandomRescale" in augmentations pipeline.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">crop_size</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">],</span> <span class="n">fill_mask</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">int</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">]</span> <span class="o">=</span> <span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> :param crop_size: tuple of (width, height) for the final crop size, if is scalar size is a</span>
- <span class="sd"> square (crop_size, crop_size)</span>
- <span class="sd"> :param fill_mask: value to fill mask labels background.</span>
- <span class="sd"> :param fill_image: grey value to fill image padded background.</span>
- <span class="sd"> """</span>
- <span class="c1"># CHECK IF CROP SIZE IS A ITERABLE OR SCALAR</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span> <span class="o">=</span> <span class="n">crop_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span> <span class="o">=</span> <span class="n">fill_mask</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span> <span class="o">=</span> <span class="nb">tuple</span><span class="p">(</span><span class="n">fill_image</span><span class="p">)</span> <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">fill_image</span><span class="p">,</span> <span class="n">Sequence</span><span class="p">)</span> <span class="k">else</span> <span class="n">fill_image</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">check_valid_arguments</span><span class="p">()</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span>
- <span class="n">w</span><span class="p">,</span> <span class="n">h</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">size</span>
- <span class="c1"># pad images from center symmetrically</span>
- <span class="k">if</span> <span class="n">w</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="ow">or</span> <span class="n">h</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]:</span>
- <span class="n">padh</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="n">h</span><span class="p">)</span> <span class="o">/</span> <span class="mi">2</span> <span class="k">if</span> <span class="n">h</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="k">else</span> <span class="mi">0</span>
- <span class="n">pad_top</span><span class="p">,</span> <span class="n">pad_bottom</span> <span class="o">=</span> <span class="n">math</span><span class="o">.</span><span class="n">ceil</span><span class="p">(</span><span class="n">padh</span><span class="p">),</span> <span class="n">math</span><span class="o">.</span><span class="n">floor</span><span class="p">(</span><span class="n">padh</span><span class="p">)</span>
- <span class="n">padw</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">-</span> <span class="n">w</span><span class="p">)</span> <span class="o">/</span> <span class="mi">2</span> <span class="k">if</span> <span class="n">w</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="k">else</span> <span class="mi">0</span>
- <span class="n">pad_left</span><span class="p">,</span> <span class="n">pad_right</span> <span class="o">=</span> <span class="n">math</span><span class="o">.</span><span class="n">ceil</span><span class="p">(</span><span class="n">padw</span><span class="p">),</span> <span class="n">math</span><span class="o">.</span><span class="n">floor</span><span class="p">(</span><span class="n">padw</span><span class="p">)</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">ImageOps</span><span class="o">.</span><span class="n">expand</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="n">border</span><span class="o">=</span><span class="p">(</span><span class="n">pad_left</span><span class="p">,</span> <span class="n">pad_top</span><span class="p">,</span> <span class="n">pad_right</span><span class="p">,</span> <span class="n">pad_bottom</span><span class="p">),</span> <span class="n">fill</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span><span class="p">)</span>
- <span class="n">mask</span> <span class="o">=</span> <span class="n">ImageOps</span><span class="o">.</span><span class="n">expand</span><span class="p">(</span><span class="n">mask</span><span class="p">,</span> <span class="n">border</span><span class="o">=</span><span class="p">(</span><span class="n">pad_left</span><span class="p">,</span> <span class="n">pad_top</span><span class="p">,</span> <span class="n">pad_right</span><span class="p">,</span> <span class="n">pad_bottom</span><span class="p">),</span> <span class="n">fill</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"mask"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mask</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">check_valid_arguments</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">,</span> <span class="n">collections</span><span class="o">.</span><span class="n">abc</span><span class="o">.</span><span class="n">Iterable</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span> <span class="o">=</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">)</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o"><=</span> <span class="mi">0</span> <span class="ow">or</span> <span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o"><=</span> <span class="mi">0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Crop size must be positive numbers, found: </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">crop_size</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span> <span class="o">=</span> <span class="n">_validate_fill_values_arguments</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">fill_mask</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">fill_image</span><span class="p">)</span>
- <span class="k">class</span> <span class="nc">SegColorJitter</span><span class="p">(</span><span class="n">transforms</span><span class="o">.</span><span class="n">ColorJitter</span><span class="p">):</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">):</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="nb">super</span><span class="p">(</span><span class="n">SegColorJitter</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__call__</span><span class="p">(</span><span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">])</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">_validate_fill_values_arguments</span><span class="p">(</span><span class="n">fill_mask</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">int</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">,</span> <span class="n">List</span><span class="p">]):</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">fill_image</span><span class="p">,</span> <span class="n">collections</span><span class="o">.</span><span class="n">abc</span><span class="o">.</span><span class="n">Iterable</span><span class="p">):</span>
- <span class="c1"># If fill_image is single value, turn to grey color in RGB mode.</span>
- <span class="n">fill_image</span> <span class="o">=</span> <span class="p">(</span><span class="n">fill_image</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">,</span> <span class="n">fill_image</span><span class="p">)</span>
- <span class="k">elif</span> <span class="nb">len</span><span class="p">(</span><span class="n">fill_image</span><span class="p">)</span> <span class="o">!=</span> <span class="mi">3</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"fill_image must be an RGB tuple of size equal to 3, found: </span><span class="si">{</span><span class="n">fill_image</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="c1"># assert values are integers</span>
- <span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">fill_mask</span><span class="p">,</span> <span class="nb">int</span><span class="p">)</span> <span class="ow">or</span> <span class="ow">not</span> <span class="nb">all</span><span class="p">(</span><span class="nb">isinstance</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="nb">int</span><span class="p">)</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">fill_image</span><span class="p">):</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Fill value must be integers,"</span> <span class="sa">f</span><span class="s2">" found: fill_image = </span><span class="si">{</span><span class="n">fill_image</span><span class="si">}</span><span class="s2">, fill_mask = </span><span class="si">{</span><span class="n">fill_mask</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="c1"># assert values in range 0-255</span>
- <span class="k">if</span> <span class="nb">min</span><span class="p">(</span><span class="n">fill_image</span><span class="p">)</span> <span class="o"><</span> <span class="mi">0</span> <span class="ow">or</span> <span class="nb">max</span><span class="p">(</span><span class="n">fill_image</span><span class="p">)</span> <span class="o">></span> <span class="mi">255</span> <span class="ow">or</span> <span class="n">fill_mask</span> <span class="o"><</span> <span class="mi">0</span> <span class="ow">or</span> <span class="n">fill_mask</span> <span class="o">></span> <span class="mi">255</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Fill value must be a value from 0 to 255,"</span> <span class="sa">f</span><span class="s2">" found: fill_image = </span><span class="si">{</span><span class="n">fill_image</span><span class="si">}</span><span class="s2">, fill_mask = </span><span class="si">{</span><span class="n">fill_mask</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">fill_mask</span><span class="p">,</span> <span class="n">fill_image</span>
- <span class="k">class</span> <span class="nc">DetectionTransform</span><span class="p">:</span>
- <span class="sd">"""</span>
- <span class="sd"> Detection transform base class.</span>
- <span class="sd"> Complex transforms that require extra data loading can use the the additional_samples_count attribute in a</span>
- <span class="sd"> similar fashion to what's been done in COCODetectionDataset:</span>
- <span class="sd"> self._load_additional_inputs_for_transform(sample, transform)</span>
- <span class="sd"> # after the above call, sample["additional_samples"] holds a list of additional inputs and targets.</span>
- <span class="sd"> sample = transform(sample)</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> additional_samples_count: (int) additional samples to be loaded.</span>
- <span class="sd"> non_empty_targets: (bool) whether the additianl targets can have empty targets or not.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">additional_samples_count</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="n">non_empty_targets</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">additional_samples_count</span> <span class="o">=</span> <span class="n">additional_samples_count</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">non_empty_targets</span> <span class="o">=</span> <span class="n">non_empty_targets</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">dict</span><span class="p">,</span> <span class="nb">list</span><span class="p">]):</span>
- <span class="k">raise</span> <span class="ne">NotImplementedError</span>
- <span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="vm">__class__</span><span class="o">.</span><span class="vm">__name__</span> <span class="o">+</span> <span class="nb">str</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="vm">__dict__</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s2">"{"</span><span class="p">,</span> <span class="s2">"("</span><span class="p">)</span><span class="o">.</span><span class="n">replace</span><span class="p">(</span><span class="s2">"}"</span><span class="p">,</span> <span class="s2">")"</span><span class="p">)</span>
- <div class="viewcode-block" id="DetectionMosaic"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionMosaic">[docs]</a><span class="k">class</span> <span class="nc">DetectionMosaic</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> DetectionMosaic detection transform</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> input_dim: (tuple) input dimension.</span>
- <span class="sd"> prob: (float) probability of applying mosaic.</span>
- <span class="sd"> enable_mosaic: (bool) whether to apply mosaic at all (regardless of prob) (default=True).</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">:</span> <span class="nb">tuple</span><span class="p">,</span> <span class="n">prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">1.0</span><span class="p">,</span> <span class="n">enable_mosaic</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionMosaic</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">additional_samples_count</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span> <span class="o">=</span> <span class="n">input_dim</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable_mosaic</span> <span class="o">=</span> <span class="n">enable_mosaic</span>
- <div class="viewcode-block" id="DetectionMosaic.close"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionMosaic.close">[docs]</a> <span class="k">def</span> <span class="nf">close</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">additional_samples_count</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable_mosaic</span> <span class="o">=</span> <span class="kc">False</span></div>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">dict</span><span class="p">,</span> <span class="nb">list</span><span class="p">]):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">enable_mosaic</span> <span class="ow">and</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">:</span>
- <span class="n">mosaic_labels</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="n">mosaic_labels_seg</span> <span class="o">=</span> <span class="p">[]</span>
- <span class="n">input_h</span><span class="p">,</span> <span class="n">input_w</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="c1"># yc, xc = s, s # mosaic center x, y</span>
- <span class="n">yc</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="mf">0.5</span> <span class="o">*</span> <span class="n">input_h</span><span class="p">,</span> <span class="mf">1.5</span> <span class="o">*</span> <span class="n">input_h</span><span class="p">))</span>
- <span class="n">xc</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="mf">0.5</span> <span class="o">*</span> <span class="n">input_w</span><span class="p">,</span> <span class="mf">1.5</span> <span class="o">*</span> <span class="n">input_w</span><span class="p">))</span>
- <span class="c1"># 3 additional samples, total of 4</span>
- <span class="n">all_samples</span> <span class="o">=</span> <span class="p">[</span><span class="n">sample</span><span class="p">]</span> <span class="o">+</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"additional_samples"</span><span class="p">]</span>
- <span class="k">for</span> <span class="n">i_mosaic</span><span class="p">,</span> <span class="n">mosaic_sample</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">all_samples</span><span class="p">):</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">_labels</span> <span class="o">=</span> <span class="n">mosaic_sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">mosaic_sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span>
- <span class="n">_labels_seg</span> <span class="o">=</span> <span class="n">mosaic_sample</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="s2">"target_seg"</span><span class="p">)</span>
- <span class="n">h0</span><span class="p">,</span> <span class="n">w0</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[:</span><span class="mi">2</span><span class="p">]</span> <span class="c1"># orig hw</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="nb">min</span><span class="p">(</span><span class="mf">1.0</span> <span class="o">*</span> <span class="n">input_h</span> <span class="o">/</span> <span class="n">h0</span><span class="p">,</span> <span class="mf">1.0</span> <span class="o">*</span> <span class="n">input_w</span> <span class="o">/</span> <span class="n">w0</span><span class="p">)</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">w0</span> <span class="o">*</span> <span class="n">scale</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">h0</span> <span class="o">*</span> <span class="n">scale</span><span class="p">)),</span> <span class="n">interpolation</span><span class="o">=</span><span class="n">cv2</span><span class="o">.</span><span class="n">INTER_LINEAR</span><span class="p">)</span>
- <span class="c1"># generate output mosaic image</span>
- <span class="p">(</span><span class="n">h</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span> <span class="o">=</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[:</span><span class="mi">3</span><span class="p">]</span>
- <span class="k">if</span> <span class="n">i_mosaic</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
- <span class="n">mosaic_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">full</span><span class="p">((</span><span class="n">input_h</span> <span class="o">*</span> <span class="mi">2</span><span class="p">,</span> <span class="n">input_w</span> <span class="o">*</span> <span class="mi">2</span><span class="p">,</span> <span class="n">c</span><span class="p">),</span> <span class="mi">114</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span>
- <span class="c1"># suffix l means large image, while s means small image in mosaic aug.</span>
- <span class="p">(</span><span class="n">l_x1</span><span class="p">,</span> <span class="n">l_y1</span><span class="p">,</span> <span class="n">l_x2</span><span class="p">,</span> <span class="n">l_y2</span><span class="p">),</span> <span class="p">(</span><span class="n">s_x1</span><span class="p">,</span> <span class="n">s_y1</span><span class="p">,</span> <span class="n">s_x2</span><span class="p">,</span> <span class="n">s_y2</span><span class="p">)</span> <span class="o">=</span> <span class="n">get_mosaic_coordinate</span><span class="p">(</span><span class="n">i_mosaic</span><span class="p">,</span> <span class="n">xc</span><span class="p">,</span> <span class="n">yc</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">h</span><span class="p">,</span> <span class="n">input_h</span><span class="p">,</span> <span class="n">input_w</span><span class="p">)</span>
- <span class="n">mosaic_img</span><span class="p">[</span><span class="n">l_y1</span><span class="p">:</span><span class="n">l_y2</span><span class="p">,</span> <span class="n">l_x1</span><span class="p">:</span><span class="n">l_x2</span><span class="p">]</span> <span class="o">=</span> <span class="n">img</span><span class="p">[</span><span class="n">s_y1</span><span class="p">:</span><span class="n">s_y2</span><span class="p">,</span> <span class="n">s_x1</span><span class="p">:</span><span class="n">s_x2</span><span class="p">]</span>
- <span class="n">padw</span><span class="p">,</span> <span class="n">padh</span> <span class="o">=</span> <span class="n">l_x1</span> <span class="o">-</span> <span class="n">s_x1</span><span class="p">,</span> <span class="n">l_y1</span> <span class="o">-</span> <span class="n">s_y1</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">_labels</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="c1"># Normalized xywh to pixel xyxy format</span>
- <span class="k">if</span> <span class="n">_labels</span><span class="o">.</span><span class="n">size</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="n">labels</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">_labels</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">padw</span>
- <span class="n">labels</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">_labels</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">padh</span>
- <span class="n">labels</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">_labels</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">+</span> <span class="n">padw</span>
- <span class="n">labels</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">_labels</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">+</span> <span class="n">padh</span>
- <span class="n">mosaic_labels</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">_labels_seg</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">labels_seg</span> <span class="o">=</span> <span class="n">_labels_seg</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="k">if</span> <span class="n">_labels</span><span class="o">.</span><span class="n">size</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="n">labels_seg</span><span class="p">[:,</span> <span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">labels_seg</span><span class="p">[:,</span> <span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">+</span> <span class="n">padw</span>
- <span class="n">labels_seg</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">scale</span> <span class="o">*</span> <span class="n">labels_seg</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">+</span> <span class="n">padh</span>
- <span class="n">mosaic_labels_seg</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">labels_seg</span><span class="p">)</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">):</span>
- <span class="n">mosaic_labels</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_w</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">])</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_h</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">])</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_w</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">])</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_h</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">])</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">mosaic_labels_seg</span><span class="p">):</span>
- <span class="n">mosaic_labels_seg</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">(</span><span class="n">mosaic_labels_seg</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels_seg</span><span class="p">[:,</span> <span class="p">::</span><span class="mi">2</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_w</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels_seg</span><span class="p">[:,</span> <span class="p">::</span><span class="mi">2</span><span class="p">])</span>
- <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">mosaic_labels_seg</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">input_h</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mosaic_labels_seg</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">])</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mosaic_img</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mosaic_labels</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"info"</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="n">mosaic_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="n">mosaic_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">mosaic_labels_seg</span><span class="p">):</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target_seg"</span><span class="p">]</span> <span class="o">=</span> <span class="n">mosaic_labels_seg</span>
- <span class="k">return</span> <span class="n">sample</span></div>
- <div class="viewcode-block" id="DetectionRandomAffine"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionRandomAffine">[docs]</a><span class="k">class</span> <span class="nc">DetectionRandomAffine</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> DetectionRandomAffine detection transform</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> target_size: (tuple) desired output shape.</span>
- <span class="sd"> degrees: (Union[tuple, float]) degrees for random rotation, when float the random values are drawn uniformly</span>
- <span class="sd"> from (-degrees, degrees)</span>
- <span class="sd"> translate: (Union[tuple, float]) translate size (in pixels) for random translation, when float the random values</span>
- <span class="sd"> are drawn uniformly from (-translate, translate)</span>
- <span class="sd"> scales: (Union[tuple, float]) values for random rescale, when float the random values are drawn uniformly</span>
- <span class="sd"> from (0.1-scales, 0.1+scales)</span>
- <span class="sd"> shear: (Union[tuple, float]) degrees for random shear, when float the random values are drawn uniformly</span>
- <span class="sd"> from (shear, shear)</span>
- <span class="sd"> enable: (bool) whether to apply the below transform at all.</span>
- <span class="sd"> filter_box_candidates: (bool) whether to filter out transformed bboxes by edge size, area ratio, and aspect ratio (default=False).</span>
- <span class="sd"> wh_thr: (float) edge size threshold when filter_box_candidates = True. Bounding oxes with edges smaller</span>
- <span class="sd"> then this values will be filtered out. (default=2)</span>
- <span class="sd"> ar_thr: (float) aspect ratio threshold filter_box_candidates = True. Bounding boxes with aspect ratio larger</span>
- <span class="sd"> then this values will be filtered out. (default=20)</span>
- <span class="sd"> area_thr:(float) threshold for area ratio between original image and the transformed one, when when filter_box_candidates = True.</span>
- <span class="sd"> Bounding boxes with such ratio smaller then this value will be filtered out. (default=0.1)</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span>
- <span class="bp">self</span><span class="p">,</span> <span class="n">degrees</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">translate</span><span class="o">=</span><span class="mf">0.1</span><span class="p">,</span> <span class="n">scales</span><span class="o">=</span><span class="mf">0.1</span><span class="p">,</span> <span class="n">shear</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">target_size</span><span class="o">=</span><span class="p">(</span><span class="mi">640</span><span class="p">,</span> <span class="mi">640</span><span class="p">),</span> <span class="n">filter_box_candidates</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span><span class="p">,</span> <span class="n">wh_thr</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">ar_thr</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">area_thr</span><span class="o">=</span><span class="mf">0.1</span>
- <span class="p">):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionRandomAffine</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">degrees</span> <span class="o">=</span> <span class="n">degrees</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">translate</span> <span class="o">=</span> <span class="n">translate</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">scale</span> <span class="o">=</span> <span class="n">scales</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">shear</span> <span class="o">=</span> <span class="n">shear</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">target_size</span> <span class="o">=</span> <span class="n">target_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable</span> <span class="o">=</span> <span class="kc">True</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">filter_box_candidates</span> <span class="o">=</span> <span class="n">filter_box_candidates</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">wh_thr</span> <span class="o">=</span> <span class="n">wh_thr</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">ar_thr</span> <span class="o">=</span> <span class="n">ar_thr</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">area_thr</span> <span class="o">=</span> <span class="n">area_thr</span>
- <div class="viewcode-block" id="DetectionRandomAffine.close"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionRandomAffine.close">[docs]</a> <span class="k">def</span> <span class="nf">close</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable</span> <span class="o">=</span> <span class="kc">False</span></div>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">enable</span><span class="p">:</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">target</span> <span class="o">=</span> <span class="n">random_affine</span><span class="p">(</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">],</span>
- <span class="n">sample</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="s2">"target_seg"</span><span class="p">),</span>
- <span class="n">target_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">target_size</span><span class="p">,</span>
- <span class="n">degrees</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">degrees</span><span class="p">,</span>
- <span class="n">translate</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">translate</span><span class="p">,</span>
- <span class="n">scales</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">scale</span><span class="p">,</span>
- <span class="n">shear</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">shear</span><span class="p">,</span>
- <span class="n">filter_box_candidates</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">filter_box_candidates</span><span class="p">,</span>
- <span class="n">wh_thr</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">wh_thr</span><span class="p">,</span>
- <span class="n">area_thr</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">area_thr</span><span class="p">,</span>
- <span class="n">ar_thr</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">ar_thr</span><span class="p">,</span>
- <span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">img</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">target</span>
- <span class="k">return</span> <span class="n">sample</span></div>
- <span class="k">class</span> <span class="nc">DetectionMixup</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Mixup detection transform</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> input_dim: (tuple) input dimension.</span>
- <span class="sd"> mixup_scale: (tuple) scale range for the additional loaded image for mixup.</span>
- <span class="sd"> prob: (float) probability of applying mixup.</span>
- <span class="sd"> enable_mixup: (bool) whether to apply mixup at all (regardless of prob) (default=True).</span>
- <span class="sd"> flip_prob: (float) prbability to apply horizontal flip to the additional sample.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">,</span> <span class="n">mixup_scale</span><span class="p">,</span> <span class="n">prob</span><span class="o">=</span><span class="mf">1.0</span><span class="p">,</span> <span class="n">enable_mixup</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">flip_prob</span><span class="o">=</span><span class="mf">0.5</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionMixup</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span><span class="n">additional_samples_count</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">non_empty_targets</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span> <span class="o">=</span> <span class="n">input_dim</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">mixup_scale</span> <span class="o">=</span> <span class="n">mixup_scale</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable_mixup</span> <span class="o">=</span> <span class="n">enable_mixup</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">flip_prob</span> <span class="o">=</span> <span class="n">flip_prob</span>
- <span class="k">def</span> <span class="nf">close</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">additional_samples_count</span> <span class="o">=</span> <span class="mi">0</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">enable_mixup</span> <span class="o">=</span> <span class="kc">False</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">):</span>
- <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">enable_mixup</span> <span class="ow">and</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">:</span>
- <span class="n">origin_img</span><span class="p">,</span> <span class="n">origin_labels</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span>
- <span class="n">cp_sample</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"additional_samples"</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">cp_labels</span> <span class="o">=</span> <span class="n">cp_sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">cp_sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span>
- <span class="n">cp_boxes</span> <span class="o">=</span> <span class="n">cp_labels</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">cp_boxes</span> <span class="o">=</span> <span class="n">_mirror</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="n">cp_boxes</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">flip_prob</span><span class="p">)</span>
- <span class="c1"># PLUG IN TARGET THE FLIPPED BOXES</span>
- <span class="n">cp_labels</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span> <span class="o">=</span> <span class="n">cp_boxes</span>
- <span class="n">jit_factor</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="o">*</span><span class="bp">self</span><span class="o">.</span><span class="n">mixup_scale</span><span class="p">)</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span> <span class="o">==</span> <span class="mi">3</span><span class="p">:</span>
- <span class="n">cp_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="mi">3</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span> <span class="o">*</span> <span class="mi">114</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">cp_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span> <span class="o">*</span> <span class="mi">114</span>
- <span class="n">cp_scale_ratio</span> <span class="o">=</span> <span class="nb">min</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">/</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="n">resized_img</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span>
- <span class="n">img</span><span class="p">,</span>
- <span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">cp_scale_ratio</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">cp_scale_ratio</span><span class="p">)),</span>
- <span class="n">interpolation</span><span class="o">=</span><span class="n">cv2</span><span class="o">.</span><span class="n">INTER_LINEAR</span><span class="p">,</span>
- <span class="p">)</span>
- <span class="n">cp_img</span><span class="p">[:</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">cp_scale_ratio</span><span class="p">),</span> <span class="p">:</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">cp_scale_ratio</span><span class="p">)]</span> <span class="o">=</span> <span class="n">resized_img</span>
- <span class="n">cp_img</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span>
- <span class="n">cp_img</span><span class="p">,</span>
- <span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">cp_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">jit_factor</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">cp_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">jit_factor</span><span class="p">)),</span>
- <span class="p">)</span>
- <span class="n">cp_scale_ratio</span> <span class="o">*=</span> <span class="n">jit_factor</span>
- <span class="n">origin_h</span><span class="p">,</span> <span class="n">origin_w</span> <span class="o">=</span> <span class="n">cp_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[:</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">target_h</span><span class="p">,</span> <span class="n">target_w</span> <span class="o">=</span> <span class="n">origin_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[:</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">padded_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="nb">max</span><span class="p">(</span><span class="n">origin_h</span><span class="p">,</span> <span class="n">target_h</span><span class="p">),</span> <span class="nb">max</span><span class="p">(</span><span class="n">origin_w</span><span class="p">,</span> <span class="n">target_w</span><span class="p">),</span> <span class="mi">3</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span>
- <span class="n">padded_img</span><span class="p">[:</span><span class="n">origin_h</span><span class="p">,</span> <span class="p">:</span><span class="n">origin_w</span><span class="p">]</span> <span class="o">=</span> <span class="n">cp_img</span>
- <span class="n">x_offset</span><span class="p">,</span> <span class="n">y_offset</span> <span class="o">=</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span>
- <span class="k">if</span> <span class="n">padded_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">></span> <span class="n">target_h</span><span class="p">:</span>
- <span class="n">y_offset</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">padded_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">-</span> <span class="n">target_h</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">padded_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">></span> <span class="n">target_w</span><span class="p">:</span>
- <span class="n">x_offset</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">padded_img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="n">target_w</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)</span>
- <span class="n">padded_cropped_img</span> <span class="o">=</span> <span class="n">padded_img</span><span class="p">[</span><span class="n">y_offset</span> <span class="p">:</span> <span class="n">y_offset</span> <span class="o">+</span> <span class="n">target_h</span><span class="p">,</span> <span class="n">x_offset</span> <span class="p">:</span> <span class="n">x_offset</span> <span class="o">+</span> <span class="n">target_w</span><span class="p">]</span>
- <span class="n">cp_bboxes_origin_np</span> <span class="o">=</span> <span class="n">adjust_box_anns</span><span class="p">(</span><span class="n">cp_labels</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span><span class="o">.</span><span class="n">copy</span><span class="p">(),</span> <span class="n">cp_scale_ratio</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">origin_w</span><span class="p">,</span> <span class="n">origin_h</span><span class="p">)</span>
- <span class="n">cp_bboxes_transformed_np</span> <span class="o">=</span> <span class="n">cp_bboxes_origin_np</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="n">cp_bboxes_transformed_np</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cp_bboxes_transformed_np</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">-</span> <span class="n">x_offset</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">target_w</span><span class="p">)</span>
- <span class="n">cp_bboxes_transformed_np</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cp_bboxes_transformed_np</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">-</span> <span class="n">y_offset</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">target_h</span><span class="p">)</span>
- <span class="n">cls_labels</span> <span class="o">=</span> <span class="n">cp_labels</span><span class="p">[:,</span> <span class="mi">4</span><span class="p">:</span><span class="mi">5</span><span class="p">]</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="n">box_labels</span> <span class="o">=</span> <span class="n">cp_bboxes_transformed_np</span>
- <span class="n">labels</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">hstack</span><span class="p">((</span><span class="n">box_labels</span><span class="p">,</span> <span class="n">cls_labels</span><span class="p">))</span>
- <span class="n">origin_labels</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">vstack</span><span class="p">((</span><span class="n">origin_labels</span><span class="p">,</span> <span class="n">labels</span><span class="p">))</span>
- <span class="n">origin_img</span> <span class="o">=</span> <span class="n">origin_img</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="n">origin_img</span> <span class="o">=</span> <span class="mf">0.5</span> <span class="o">*</span> <span class="n">origin_img</span> <span class="o">+</span> <span class="mf">0.5</span> <span class="o">*</span> <span class="n">padded_cropped_img</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">origin_img</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">),</span> <span class="n">origin_labels</span>
- <span class="k">return</span> <span class="n">sample</span>
- <div class="viewcode-block" id="DetectionPaddedRescale"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionPaddedRescale">[docs]</a><span class="k">class</span> <span class="nc">DetectionPaddedRescale</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Preprocessing transform to be applied last of all transforms for validation.</span>
- <span class="sd"> Image- Rescales and pads to self.input_dim.</span>
- <span class="sd"> Targets- pads targets to max_targets, moves the class label to first index, converts boxes format- xyxy -> cxcywh.</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> input_dim: (tuple) final input dimension (default=(640,640))</span>
- <span class="sd"> swap: image axis's to be rearranged.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">,</span> <span class="n">swap</span><span class="o">=</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">max_targets</span><span class="o">=</span><span class="mi">50</span><span class="p">,</span> <span class="n">pad_value</span><span class="o">=</span><span class="mi">114</span><span class="p">):</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">swap</span> <span class="o">=</span> <span class="n">swap</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span> <span class="o">=</span> <span class="n">input_dim</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span> <span class="o">=</span> <span class="n">max_targets</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">pad_value</span> <span class="o">=</span> <span class="n">pad_value</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">]):</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">targets</span><span class="p">,</span> <span class="n">crowd_targets</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">],</span> <span class="n">sample</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="s2">"crowd_target"</span><span class="p">)</span>
- <span class="n">img</span><span class="p">,</span> <span class="n">r</span> <span class="o">=</span> <span class="n">rescale_and_pad_to_size</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_dim</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">swap</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">pad_value</span><span class="p">)</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">img</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_rescale_target</span><span class="p">(</span><span class="n">targets</span><span class="p">,</span> <span class="n">r</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">crowd_targets</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"crowd_target"</span><span class="p">]</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_rescale_target</span><span class="p">(</span><span class="n">crowd_targets</span><span class="p">,</span> <span class="n">r</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">sample</span>
- <span class="k">def</span> <span class="nf">_rescale_target</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">targets</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">,</span> <span class="n">r</span><span class="p">:</span> <span class="nb">float</span><span class="p">)</span> <span class="o">-></span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">:</span>
- <span class="sd">"""SegRescale the target according to a coefficient used to rescale the image.</span>
- <span class="sd"> This is done to have images and targets at the same scale.</span>
- <span class="sd"> :param targets: Targets to rescale, shape (batch_size, 6)</span>
- <span class="sd"> :param r: SegRescale coefficient that was applied to the image</span>
- <span class="sd"> :return: Rescaled targets, shape (batch_size, 6)</span>
- <span class="sd"> """</span>
- <span class="n">targets</span> <span class="o">=</span> <span class="n">targets</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span> <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">targets</span><span class="p">)</span> <span class="o">></span> <span class="mi">0</span> <span class="k">else</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span><span class="p">,</span> <span class="mi">5</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="n">boxes</span><span class="p">,</span> <span class="n">labels</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">],</span> <span class="n">targets</span><span class="p">[:,</span> <span class="mi">4</span><span class="p">]</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">xyxy2cxcywh</span><span class="p">(</span><span class="n">boxes</span><span class="p">)</span>
- <span class="n">boxes</span> <span class="o">*=</span> <span class="n">r</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">cxcywh2xyxy</span><span class="p">(</span><span class="n">boxes</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">((</span><span class="n">boxes</span><span class="p">,</span> <span class="n">labels</span><span class="p">[:,</span> <span class="n">np</span><span class="o">.</span><span class="n">newaxis</span><span class="p">]),</span> <span class="mi">1</span><span class="p">)</span></div>
- <span class="k">class</span> <span class="nc">DetectionHorizontalFlip</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Horizontal Flip for Detection</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> prob: float: probability of applying horizontal flip</span>
- <span class="sd"> max_targets: int: max objects in single image, padding target to this size in case of empty image.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">prob</span><span class="p">,</span> <span class="n">max_targets</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">120</span><span class="p">):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionHorizontalFlip</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span> <span class="o">=</span> <span class="n">max_targets</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">):</span>
- <span class="n">image</span><span class="p">,</span> <span class="n">targets</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">boxes</span><span class="p">)</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
- <span class="n">targets</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span><span class="p">,</span> <span class="mi">5</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span>
- <span class="n">image</span><span class="p">,</span> <span class="n">boxes</span> <span class="o">=</span> <span class="n">_mirror</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="n">boxes</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">)</span>
- <span class="n">targets</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">targets</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">]</span> <span class="o">=</span> <span class="n">image</span>
- <span class="k">return</span> <span class="n">sample</span>
- <div class="viewcode-block" id="DetectionHSV"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionHSV">[docs]</a><span class="k">class</span> <span class="nc">DetectionHSV</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Detection HSV transform.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">prob</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">hgain</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">,</span> <span class="n">sgain</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">,</span> <span class="n">vgain</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">,</span> <span class="n">bgr_channels</span><span class="o">=</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">)):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionHSV</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">prob</span> <span class="o">=</span> <span class="n">prob</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">hgain</span> <span class="o">=</span> <span class="n">hgain</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">sgain</span> <span class="o">=</span> <span class="n">sgain</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">vgain</span> <span class="o">=</span> <span class="n">vgain</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">bgr_channels</span> <span class="o">=</span> <span class="n">bgr_channels</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">:</span> <span class="nb">dict</span><span class="p">)</span> <span class="o">-></span> <span class="nb">dict</span><span class="p">:</span>
- <span class="k">if</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob</span><span class="p">:</span>
- <span class="n">augment_hsv</span><span class="p">(</span><span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="bp">self</span><span class="o">.</span><span class="n">hgain</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">sgain</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">vgain</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">bgr_channels</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">sample</span></div>
- <div class="viewcode-block" id="DetectionTargetsFormatTransform"><a class="viewcode-back" href="../../../../super_gradients.training.html#super_gradients.training.transforms.DetectionTargetsFormatTransform">[docs]</a><span class="k">class</span> <span class="nc">DetectionTargetsFormatTransform</span><span class="p">(</span><span class="n">DetectionTransform</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Detection targets format transform</span>
- <span class="sd"> Converts targets in input_format to output_format.</span>
- <span class="sd"> Attributes:</span>
- <span class="sd"> input_format: DetectionTargetsFormat: input target format</span>
- <span class="sd"> output_format: DetectionTargetsFormat: output target format</span>
- <span class="sd"> min_bbox_edge_size: int: bboxes with edge size lower then this values will be removed.</span>
- <span class="sd"> max_targets: int: max objects in single image, padding target to this size.</span>
- <span class="sd"> """</span>
- <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span>
- <span class="bp">self</span><span class="p">,</span>
- <span class="n">input_format</span><span class="p">:</span> <span class="n">DetectionTargetsFormat</span> <span class="o">=</span> <span class="n">DetectionTargetsFormat</span><span class="o">.</span><span class="n">XYXY_LABEL</span><span class="p">,</span>
- <span class="n">output_format</span><span class="p">:</span> <span class="n">DetectionTargetsFormat</span> <span class="o">=</span> <span class="n">DetectionTargetsFormat</span><span class="o">.</span><span class="n">LABEL_CXCYWH</span><span class="p">,</span>
- <span class="n">min_bbox_edge_size</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">1</span><span class="p">,</span>
- <span class="n">max_targets</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">120</span><span class="p">,</span>
- <span class="p">):</span>
- <span class="nb">super</span><span class="p">(</span><span class="n">DetectionTargetsFormatTransform</span><span class="p">,</span> <span class="bp">self</span><span class="p">)</span><span class="o">.</span><span class="fm">__init__</span><span class="p">()</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">input_format</span> <span class="o">=</span> <span class="n">input_format</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">output_format</span> <span class="o">=</span> <span class="n">output_format</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">min_bbox_edge_size</span> <span class="o">=</span> <span class="n">min_bbox_edge_size</span>
- <span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span> <span class="o">=</span> <span class="n">max_targets</span>
- <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">sample</span><span class="p">):</span>
- <span class="n">normalized_input</span> <span class="o">=</span> <span class="s2">"NORMALIZED"</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_format</span><span class="o">.</span><span class="n">value</span>
- <span class="n">normalized_output</span> <span class="o">=</span> <span class="s2">"NORMALIZED"</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">output_format</span><span class="o">.</span><span class="n">value</span>
- <span class="n">normalize</span> <span class="o">=</span> <span class="ow">not</span> <span class="n">normalized_input</span> <span class="ow">and</span> <span class="n">normalized_output</span>
- <span class="n">denormalize</span> <span class="o">=</span> <span class="n">normalized_input</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">normalized_output</span>
- <span class="n">label_first_in_input</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_format</span><span class="o">.</span><span class="n">value</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s2">"_"</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span> <span class="o">==</span> <span class="s2">"LABEL"</span>
- <span class="n">label_first_in_output</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">output_format</span><span class="o">.</span><span class="n">value</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s2">"_"</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span> <span class="o">==</span> <span class="s2">"LABEL"</span>
- <span class="n">input_xyxy_format</span> <span class="o">=</span> <span class="s2">"XYXY"</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">input_format</span><span class="o">.</span><span class="n">value</span>
- <span class="n">output_xyxy_format</span> <span class="o">=</span> <span class="s2">"XYXY"</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">output_format</span><span class="o">.</span><span class="n">value</span>
- <span class="n">convert2xyxy</span> <span class="o">=</span> <span class="ow">not</span> <span class="n">input_xyxy_format</span> <span class="ow">and</span> <span class="n">output_xyxy_format</span>
- <span class="n">convert2cxcy</span> <span class="o">=</span> <span class="n">input_xyxy_format</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">output_xyxy_format</span>
- <span class="n">image</span><span class="p">,</span> <span class="n">targets</span><span class="p">,</span> <span class="n">crowd_targets</span> <span class="o">=</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"image"</span><span class="p">],</span> <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">],</span> <span class="n">sample</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="s2">"crowd_target"</span><span class="p">)</span>
- <span class="n">_</span><span class="p">,</span> <span class="n">h</span><span class="p">,</span> <span class="n">w</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">shape</span>
- <span class="k">def</span> <span class="nf">_format_target</span><span class="p">(</span><span class="n">targets_in</span><span class="p">):</span>
- <span class="k">if</span> <span class="n">label_first_in_input</span><span class="p">:</span>
- <span class="n">labels</span><span class="p">,</span> <span class="n">boxes</span> <span class="o">=</span> <span class="n">targets_in</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">],</span> <span class="n">targets_in</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">:]</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">boxes</span><span class="p">,</span> <span class="n">labels</span> <span class="o">=</span> <span class="n">targets_in</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">],</span> <span class="n">targets_in</span><span class="p">[:,</span> <span class="mi">4</span><span class="p">]</span>
- <span class="k">if</span> <span class="n">convert2cxcy</span><span class="p">:</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">xyxy2cxcywh</span><span class="p">(</span><span class="n">boxes</span><span class="p">)</span>
- <span class="k">elif</span> <span class="n">convert2xyxy</span><span class="p">:</span>
- <span class="n">boxes</span> <span class="o">=</span> <span class="n">cxcywh2xyxy</span><span class="p">(</span><span class="n">boxes</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">normalize</span><span class="p">:</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">/</span> <span class="n">w</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">h</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">/</span> <span class="n">w</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">/</span> <span class="n">h</span>
- <span class="k">elif</span> <span class="n">denormalize</span><span class="p">:</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">w</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">h</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">*</span> <span class="n">w</span>
- <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">]</span> <span class="o">*</span> <span class="n">h</span>
- <span class="n">min_bbox_edge_size</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">min_bbox_edge_size</span> <span class="o">/</span> <span class="nb">max</span><span class="p">(</span><span class="n">w</span><span class="p">,</span> <span class="n">h</span><span class="p">)</span> <span class="k">if</span> <span class="n">normalized_output</span> <span class="k">else</span> <span class="bp">self</span><span class="o">.</span><span class="n">min_bbox_edge_size</span>
- <span class="n">cxcywh_boxes</span> <span class="o">=</span> <span class="n">boxes</span> <span class="k">if</span> <span class="ow">not</span> <span class="n">output_xyxy_format</span> <span class="k">else</span> <span class="n">xyxy2cxcywh</span><span class="p">(</span><span class="n">boxes</span><span class="o">.</span><span class="n">copy</span><span class="p">())</span>
- <span class="n">mask_b</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">minimum</span><span class="p">(</span><span class="n">cxcywh_boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">],</span> <span class="n">cxcywh_boxes</span><span class="p">[:,</span> <span class="mi">3</span><span class="p">])</span> <span class="o">></span> <span class="n">min_bbox_edge_size</span>
- <span class="n">boxes_t</span> <span class="o">=</span> <span class="n">boxes</span><span class="p">[</span><span class="n">mask_b</span><span class="p">]</span>
- <span class="n">labels_t</span> <span class="o">=</span> <span class="n">labels</span><span class="p">[</span><span class="n">mask_b</span><span class="p">]</span>
- <span class="n">labels_t</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">expand_dims</span><span class="p">(</span><span class="n">labels_t</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
- <span class="n">targets_t</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">hstack</span><span class="p">((</span><span class="n">labels_t</span><span class="p">,</span> <span class="n">boxes_t</span><span class="p">))</span> <span class="k">if</span> <span class="n">label_first_in_output</span> <span class="k">else</span> <span class="n">np</span><span class="o">.</span><span class="n">hstack</span><span class="p">((</span><span class="n">boxes_t</span><span class="p">,</span> <span class="n">labels_t</span><span class="p">))</span>
- <span class="n">padded_targets</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span><span class="p">,</span> <span class="mi">5</span><span class="p">))</span>
- <span class="n">padded_targets</span><span class="p">[</span><span class="nb">range</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">targets_t</span><span class="p">))[:</span> <span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span><span class="p">]]</span> <span class="o">=</span> <span class="n">targets_t</span><span class="p">[:</span> <span class="bp">self</span><span class="o">.</span><span class="n">max_targets</span><span class="p">]</span>
- <span class="n">padded_targets</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ascontiguousarray</span><span class="p">(</span><span class="n">padded_targets</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">padded_targets</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">_format_target</span><span class="p">(</span><span class="n">targets</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">crowd_targets</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
- <span class="n">sample</span><span class="p">[</span><span class="s2">"crowd_target"</span><span class="p">]</span> <span class="o">=</span> <span class="n">_format_target</span><span class="p">(</span><span class="n">crowd_targets</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">sample</span></div>
- <span class="k">def</span> <span class="nf">get_aug_params</span><span class="p">(</span><span class="n">value</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">tuple</span><span class="p">,</span> <span class="nb">float</span><span class="p">],</span> <span class="n">center</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mi">0</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Generates a random value for augmentations as described below</span>
- <span class="sd"> :param value: Union[tuple, float] defines the range of values for generation. Wen tuple-</span>
- <span class="sd"> drawn uniformly between (value[0], value[1]), and (center - value, center + value) when float</span>
- <span class="sd"> :param center: float, defines center to subtract when value is float.</span>
- <span class="sd"> :return: generated value</span>
- <span class="sd"> """</span>
- <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">value</span><span class="p">,</span> <span class="nb">float</span><span class="p">):</span>
- <span class="k">return</span> <span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">center</span> <span class="o">-</span> <span class="n">value</span><span class="p">,</span> <span class="n">center</span> <span class="o">+</span> <span class="n">value</span><span class="p">)</span>
- <span class="k">elif</span> <span class="nb">len</span><span class="p">(</span><span class="n">value</span><span class="p">)</span> <span class="o">==</span> <span class="mi">2</span><span class="p">:</span>
- <span class="k">return</span> <span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="n">value</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">value</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span>
- <span class="s2">"Affine params should be either a sequence containing two values</span><span class="se">\</span>
- <span class="s2"> or single float values. Got </span><span class="si">{}</span><span class="s2">"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
- <span class="n">value</span>
- <span class="p">)</span>
- <span class="p">)</span>
- <span class="k">def</span> <span class="nf">get_affine_matrix</span><span class="p">(</span>
- <span class="n">target_size</span><span class="p">,</span>
- <span class="n">degrees</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span>
- <span class="n">translate</span><span class="o">=</span><span class="mf">0.1</span><span class="p">,</span>
- <span class="n">scales</span><span class="o">=</span><span class="mf">0.1</span><span class="p">,</span>
- <span class="n">shear</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span>
- <span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Returns a random affine transform matrix.</span>
- <span class="sd"> :param target_size: (tuple) desired output shape.</span>
- <span class="sd"> :param degrees: (Union[tuple, float]) degrees for random rotation, when float the random values are drawn uniformly</span>
- <span class="sd"> from (-degrees, degrees)</span>
- <span class="sd"> :param translate: (Union[tuple, float]) translate size (in pixels) for random translation, when float the random values</span>
- <span class="sd"> are drawn uniformly from (-translate, translate)</span>
- <span class="sd"> :param scales: (Union[tuple, float]) values for random rescale, when float the random values are drawn uniformly</span>
- <span class="sd"> from (0.1-scales, 0.1+scales)</span>
- <span class="sd"> :param shear: (Union[tuple, float]) degrees for random shear, when float the random values are drawn uniformly</span>
- <span class="sd"> from (shear, shear)</span>
- <span class="sd"> :return: affine_transform_matrix, drawn_scale</span>
- <span class="sd"> """</span>
- <span class="n">twidth</span><span class="p">,</span> <span class="n">theight</span> <span class="o">=</span> <span class="n">target_size</span>
- <span class="c1"># Rotation and Scale</span>
- <span class="n">angle</span> <span class="o">=</span> <span class="n">get_aug_params</span><span class="p">(</span><span class="n">degrees</span><span class="p">)</span>
- <span class="n">scale</span> <span class="o">=</span> <span class="n">get_aug_params</span><span class="p">(</span><span class="n">scales</span><span class="p">,</span> <span class="n">center</span><span class="o">=</span><span class="mf">1.0</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">scale</span> <span class="o"><=</span> <span class="mf">0.0</span><span class="p">:</span>
- <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Argument scale should be positive"</span><span class="p">)</span>
- <span class="n">R</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">getRotationMatrix2D</span><span class="p">(</span><span class="n">angle</span><span class="o">=</span><span class="n">angle</span><span class="p">,</span> <span class="n">center</span><span class="o">=</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">),</span> <span class="n">scale</span><span class="o">=</span><span class="n">scale</span><span class="p">)</span>
- <span class="n">M</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">([</span><span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">])</span>
- <span class="c1"># Shear</span>
- <span class="n">shear_x</span> <span class="o">=</span> <span class="n">math</span><span class="o">.</span><span class="n">tan</span><span class="p">(</span><span class="n">get_aug_params</span><span class="p">(</span><span class="n">shear</span><span class="p">)</span> <span class="o">*</span> <span class="n">math</span><span class="o">.</span><span class="n">pi</span> <span class="o">/</span> <span class="mi">180</span><span class="p">)</span>
- <span class="n">shear_y</span> <span class="o">=</span> <span class="n">math</span><span class="o">.</span><span class="n">tan</span><span class="p">(</span><span class="n">get_aug_params</span><span class="p">(</span><span class="n">shear</span><span class="p">)</span> <span class="o">*</span> <span class="n">math</span><span class="o">.</span><span class="n">pi</span> <span class="o">/</span> <span class="mi">180</span><span class="p">)</span>
- <span class="n">M</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">R</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">shear_y</span> <span class="o">*</span> <span class="n">R</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">M</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">R</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">shear_x</span> <span class="o">*</span> <span class="n">R</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
- <span class="c1"># Translation</span>
- <span class="n">translation_x</span> <span class="o">=</span> <span class="n">get_aug_params</span><span class="p">(</span><span class="n">translate</span><span class="p">)</span> <span class="o">*</span> <span class="n">twidth</span> <span class="c1"># x translation (pixels)</span>
- <span class="n">translation_y</span> <span class="o">=</span> <span class="n">get_aug_params</span><span class="p">(</span><span class="n">translate</span><span class="p">)</span> <span class="o">*</span> <span class="n">theight</span> <span class="c1"># y translation (pixels)</span>
- <span class="n">M</span><span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">translation_x</span>
- <span class="n">M</span><span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">translation_y</span>
- <span class="k">return</span> <span class="n">M</span><span class="p">,</span> <span class="n">scale</span>
- <span class="k">def</span> <span class="nf">apply_affine_to_bboxes</span><span class="p">(</span><span class="n">targets</span><span class="p">,</span> <span class="n">targets_seg</span><span class="p">,</span> <span class="n">target_size</span><span class="p">,</span> <span class="n">M</span><span class="p">):</span>
- <span class="n">num_gts</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">targets</span><span class="p">)</span>
- <span class="n">twidth</span><span class="p">,</span> <span class="n">theight</span> <span class="o">=</span> <span class="n">target_size</span>
- <span class="c1"># targets_seg = [B x w x h]</span>
- <span class="c1"># if any is_not_nan in axis = 1</span>
- <span class="n">seg_is_present_mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">logical_or</span><span class="o">.</span><span class="n">reduce</span><span class="p">(</span><span class="o">~</span><span class="n">np</span><span class="o">.</span><span class="n">isnan</span><span class="p">(</span><span class="n">targets_seg</span><span class="p">),</span> <span class="n">axis</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
- <span class="n">num_gts_masks</span> <span class="o">=</span> <span class="n">seg_is_present_mask</span><span class="o">.</span><span class="n">sum</span><span class="p">()</span>
- <span class="n">num_gts_boxes</span> <span class="o">=</span> <span class="n">num_gts</span> <span class="o">-</span> <span class="n">num_gts_masks</span>
- <span class="k">if</span> <span class="n">num_gts_boxes</span><span class="p">:</span>
- <span class="c1"># warp corner points</span>
- <span class="n">corner_points</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">num_gts_boxes</span> <span class="o">*</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span>
- <span class="c1"># x1y1, x2y2, x1y2, x2y1</span>
- <span class="n">corner_points</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[</span><span class="o">~</span><span class="n">seg_is_present_mask</span><span class="p">][:,</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">1</span><span class="p">]]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">num_gts_boxes</span> <span class="o">*</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
- <span class="n">corner_points</span> <span class="o">=</span> <span class="n">corner_points</span> <span class="o">@</span> <span class="n">M</span><span class="o">.</span><span class="n">T</span> <span class="c1"># apply affine transform</span>
- <span class="n">corner_points</span> <span class="o">=</span> <span class="n">corner_points</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">num_gts_boxes</span><span class="p">,</span> <span class="mi">8</span><span class="p">)</span>
- <span class="c1"># create new boxes</span>
- <span class="n">corner_xs</span> <span class="o">=</span> <span class="n">corner_points</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">corner_ys</span> <span class="o">=</span> <span class="n">corner_points</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">new_bboxes</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">((</span><span class="n">np</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">corner_xs</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">corner_ys</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">max</span><span class="p">(</span><span class="n">corner_xs</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">max</span><span class="p">(</span><span class="n">corner_ys</span><span class="p">,</span> <span class="mi">1</span><span class="p">)))</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">.</span><span class="n">T</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">new_bboxes</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">0</span><span class="p">,</span> <span class="mi">4</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">num_gts_masks</span><span class="p">:</span>
- <span class="c1"># warp segmentation points</span>
- <span class="n">num_seg_points</span> <span class="o">=</span> <span class="n">targets_seg</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">//</span> <span class="mi">2</span>
- <span class="n">corner_points_seg</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">num_gts_masks</span> <span class="o">*</span> <span class="n">num_seg_points</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span>
- <span class="n">corner_points_seg</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">targets_seg</span><span class="p">[</span><span class="n">seg_is_present_mask</span><span class="p">]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">num_gts_masks</span> <span class="o">*</span> <span class="n">num_seg_points</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
- <span class="n">corner_points_seg</span> <span class="o">=</span> <span class="n">corner_points_seg</span> <span class="o">@</span> <span class="n">M</span><span class="o">.</span><span class="n">T</span>
- <span class="n">corner_points_seg</span> <span class="o">=</span> <span class="n">corner_points_seg</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">num_gts_masks</span><span class="p">,</span> <span class="n">num_seg_points</span> <span class="o">*</span> <span class="mi">2</span><span class="p">)</span>
- <span class="c1"># create new boxes</span>
- <span class="n">seg_points_xs</span> <span class="o">=</span> <span class="n">corner_points_seg</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">seg_points_ys</span> <span class="o">=</span> <span class="n">corner_points_seg</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span>
- <span class="n">new_tight_bboxes</span> <span class="o">=</span> <span class="p">(</span>
- <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">((</span><span class="n">np</span><span class="o">.</span><span class="n">nanmin</span><span class="p">(</span><span class="n">seg_points_xs</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">nanmin</span><span class="p">(</span><span class="n">seg_points_ys</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">nanmax</span><span class="p">(</span><span class="n">seg_points_xs</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">np</span><span class="o">.</span><span class="n">nanmax</span><span class="p">(</span><span class="n">seg_points_ys</span><span class="p">,</span> <span class="mi">1</span><span class="p">)))</span>
- <span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">)</span>
- <span class="o">.</span><span class="n">T</span>
- <span class="p">)</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">new_tight_bboxes</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">0</span><span class="p">,</span> <span class="mi">4</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float</span><span class="p">)</span>
- <span class="n">targets</span><span class="p">[</span><span class="o">~</span><span class="n">seg_is_present_mask</span><span class="p">,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span> <span class="o">=</span> <span class="n">new_bboxes</span>
- <span class="n">targets</span><span class="p">[</span><span class="n">seg_is_present_mask</span><span class="p">,</span> <span class="p">:</span><span class="mi">4</span><span class="p">]</span> <span class="o">=</span> <span class="n">new_tight_bboxes</span>
- <span class="c1"># clip boxes</span>
- <span class="n">targets</span><span class="p">[:,</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">]]</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">]]</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">twidth</span><span class="p">)</span>
- <span class="n">targets</span><span class="p">[:,</span> <span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">3</span><span class="p">]]</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">3</span><span class="p">]]</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">theight</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">targets</span>
- <span class="k">def</span> <span class="nf">random_affine</span><span class="p">(</span>
- <span class="n">img</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">,</span>
- <span class="n">targets</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span> <span class="o">=</span> <span class="p">(),</span>
- <span class="n">targets_seg</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span>
- <span class="n">target_size</span><span class="p">:</span> <span class="nb">tuple</span> <span class="o">=</span> <span class="p">(</span><span class="mi">640</span><span class="p">,</span> <span class="mi">640</span><span class="p">),</span>
- <span class="n">degrees</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="mi">10</span><span class="p">,</span>
- <span class="n">translate</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="mf">0.1</span><span class="p">,</span>
- <span class="n">scales</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="mf">0.1</span><span class="p">,</span>
- <span class="n">shear</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">float</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">]</span> <span class="o">=</span> <span class="mi">10</span><span class="p">,</span>
- <span class="n">filter_box_candidates</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span><span class="p">,</span>
- <span class="n">wh_thr</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span>
- <span class="n">ar_thr</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span>
- <span class="n">area_thr</span><span class="o">=</span><span class="mf">0.1</span><span class="p">,</span>
- <span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Performs random affine transform to img, targets</span>
- <span class="sd"> :param img: Input image</span>
- <span class="sd"> :param targets: Input target</span>
- <span class="sd"> :param targets_seg: Targets derived from segmentation masks</span>
- <span class="sd"> :param target_size: Desired output shape</span>
- <span class="sd"> :param degrees: Degrees for random rotation, when float the random values are drawn uniformly</span>
- <span class="sd"> from (-degrees, degrees).</span>
- <span class="sd"> :param translate: Translate size (in pixels) for random translation, when float the random values</span>
- <span class="sd"> are drawn uniformly from (-translate, translate)</span>
- <span class="sd"> :param scales: Values for random rescale, when float the random values are drawn uniformly</span>
- <span class="sd"> from (0.1-scales, 0.1+scales)</span>
- <span class="sd"> :param shear: Degrees for random shear, when float the random values are drawn uniformly</span>
- <span class="sd"> from (shear, shear)</span>
- <span class="sd"> :param filter_box_candidates: whether to filter out transformed bboxes by edge size, area ratio, and aspect ratio.</span>
- <span class="sd"> :param wh_thr: (float) edge size threshold when filter_box_candidates = True. Bounding oxes with edges smaller</span>
- <span class="sd"> then this values will be filtered out. (default=2)</span>
- <span class="sd"> :param ar_thr: (float) aspect ratio threshold filter_box_candidates = True. Bounding boxes with aspect ratio larger</span>
- <span class="sd"> then this values will be filtered out. (default=20)</span>
- <span class="sd"> :param area_thr:(float) threshold for area ratio between original image and the transformed one, when when filter_box_candidates = True.</span>
- <span class="sd"> Bounding boxes with such ratio smaller then this value will be filtered out. (default=0.1)</span>
- <span class="sd"> :return: Image and Target with applied random affine</span>
- <span class="sd"> """</span>
- <span class="n">targets_seg</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="n">targets</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="mi">0</span><span class="p">))</span> <span class="k">if</span> <span class="n">targets_seg</span> <span class="ow">is</span> <span class="kc">None</span> <span class="k">else</span> <span class="n">targets_seg</span>
- <span class="n">M</span><span class="p">,</span> <span class="n">scale</span> <span class="o">=</span> <span class="n">get_affine_matrix</span><span class="p">(</span><span class="n">target_size</span><span class="p">,</span> <span class="n">degrees</span><span class="p">,</span> <span class="n">translate</span><span class="p">,</span> <span class="n">scales</span><span class="p">,</span> <span class="n">shear</span><span class="p">)</span>
- <span class="n">img</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">warpAffine</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="n">M</span><span class="p">,</span> <span class="n">dsize</span><span class="o">=</span><span class="n">target_size</span><span class="p">,</span> <span class="n">borderValue</span><span class="o">=</span><span class="p">(</span><span class="mi">114</span><span class="p">,</span> <span class="mi">114</span><span class="p">,</span> <span class="mi">114</span><span class="p">))</span>
- <span class="c1"># Transform label coordinates</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">targets</span><span class="p">)</span> <span class="o">></span> <span class="mi">0</span><span class="p">:</span>
- <span class="n">targets_orig</span> <span class="o">=</span> <span class="n">targets</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="n">targets</span> <span class="o">=</span> <span class="n">apply_affine_to_bboxes</span><span class="p">(</span><span class="n">targets</span><span class="p">,</span> <span class="n">targets_seg</span><span class="p">,</span> <span class="n">target_size</span><span class="p">,</span> <span class="n">M</span><span class="p">)</span>
- <span class="k">if</span> <span class="n">filter_box_candidates</span><span class="p">:</span>
- <span class="n">box_candidates_ids</span> <span class="o">=</span> <span class="n">_filter_box_candidates</span><span class="p">(</span><span class="n">targets_orig</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">],</span> <span class="n">targets</span><span class="p">[:,</span> <span class="p">:</span><span class="mi">4</span><span class="p">],</span> <span class="n">wh_thr</span><span class="o">=</span><span class="n">wh_thr</span><span class="p">,</span> <span class="n">ar_thr</span><span class="o">=</span><span class="n">ar_thr</span><span class="p">,</span> <span class="n">area_thr</span><span class="o">=</span><span class="n">area_thr</span><span class="p">)</span>
- <span class="n">targets</span> <span class="o">=</span> <span class="n">targets</span><span class="p">[</span><span class="n">box_candidates_ids</span><span class="p">]</span>
- <span class="k">return</span> <span class="n">img</span><span class="p">,</span> <span class="n">targets</span>
- <span class="k">def</span> <span class="nf">_filter_box_candidates</span><span class="p">(</span><span class="n">box1</span><span class="p">,</span> <span class="n">box2</span><span class="p">,</span> <span class="n">wh_thr</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">ar_thr</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">area_thr</span><span class="o">=</span><span class="mf">0.1</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> compute candidate boxes</span>
- <span class="sd"> :param box1: before augment</span>
- <span class="sd"> :param box2: after augment</span>
- <span class="sd"> :param wh_thr: wh_thr (pixels)</span>
- <span class="sd"> :param ar_thr: aspect_ratio_thr</span>
- <span class="sd"> :param area_thr: area_ratio</span>
- <span class="sd"> :return:</span>
- <span class="sd"> """</span>
- <span class="n">box1</span> <span class="o">=</span> <span class="n">box1</span><span class="o">.</span><span class="n">T</span>
- <span class="n">box2</span> <span class="o">=</span> <span class="n">box2</span><span class="o">.</span><span class="n">T</span>
- <span class="n">w1</span><span class="p">,</span> <span class="n">h1</span> <span class="o">=</span> <span class="n">box1</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span> <span class="o">-</span> <span class="n">box1</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">box1</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span> <span class="o">-</span> <span class="n">box1</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">w2</span><span class="p">,</span> <span class="n">h2</span> <span class="o">=</span> <span class="n">box2</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span> <span class="o">-</span> <span class="n">box2</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">box2</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span> <span class="o">-</span> <span class="n">box2</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">ar</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">maximum</span><span class="p">(</span><span class="n">w2</span> <span class="o">/</span> <span class="p">(</span><span class="n">h2</span> <span class="o">+</span> <span class="mf">1e-16</span><span class="p">),</span> <span class="n">h2</span> <span class="o">/</span> <span class="p">(</span><span class="n">w2</span> <span class="o">+</span> <span class="mf">1e-16</span><span class="p">))</span> <span class="c1"># aspect ratio</span>
- <span class="k">return</span> <span class="p">(</span><span class="n">w2</span> <span class="o">></span> <span class="n">wh_thr</span><span class="p">)</span> <span class="o">&</span> <span class="p">(</span><span class="n">h2</span> <span class="o">></span> <span class="n">wh_thr</span><span class="p">)</span> <span class="o">&</span> <span class="p">(</span><span class="n">w2</span> <span class="o">*</span> <span class="n">h2</span> <span class="o">/</span> <span class="p">(</span><span class="n">w1</span> <span class="o">*</span> <span class="n">h1</span> <span class="o">+</span> <span class="mf">1e-16</span><span class="p">)</span> <span class="o">></span> <span class="n">area_thr</span><span class="p">)</span> <span class="o">&</span> <span class="p">(</span><span class="n">ar</span> <span class="o"><</span> <span class="n">ar_thr</span><span class="p">)</span> <span class="c1"># candidates</span>
- <span class="k">def</span> <span class="nf">_mirror</span><span class="p">(</span><span class="n">image</span><span class="p">,</span> <span class="n">boxes</span><span class="p">,</span> <span class="n">prob</span><span class="o">=</span><span class="mf">0.5</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Horizontal flips image and bboxes with probability prob.</span>
- <span class="sd"> :param image: (np.array) image to be flipped.</span>
- <span class="sd"> :param boxes: (np.array) bboxes to be modified.</span>
- <span class="sd"> :param prob: probability to perform flipping.</span>
- <span class="sd"> :return: flipped_image, flipped_bboxes</span>
- <span class="sd"> """</span>
- <span class="n">flipped_boxes</span> <span class="o">=</span> <span class="n">boxes</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
- <span class="n">_</span><span class="p">,</span> <span class="n">width</span><span class="p">,</span> <span class="n">_</span> <span class="o">=</span> <span class="n">image</span><span class="o">.</span><span class="n">shape</span>
- <span class="k">if</span> <span class="n">random</span><span class="o">.</span><span class="n">random</span><span class="p">()</span> <span class="o"><</span> <span class="n">prob</span><span class="p">:</span>
- <span class="n">image</span> <span class="o">=</span> <span class="n">image</span><span class="p">[:,</span> <span class="p">::</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
- <span class="n">flipped_boxes</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">::</span><span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">width</span> <span class="o">-</span> <span class="n">boxes</span><span class="p">[:,</span> <span class="mi">2</span><span class="p">::</span><span class="o">-</span><span class="mi">2</span><span class="p">]</span>
- <span class="k">return</span> <span class="n">image</span><span class="p">,</span> <span class="n">flipped_boxes</span>
- <span class="k">def</span> <span class="nf">augment_hsv</span><span class="p">(</span><span class="n">img</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">,</span> <span class="n">hgain</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">sgain</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">vgain</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">bgr_channels</span><span class="o">=</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">)):</span>
- <span class="n">hsv_augs</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">uniform</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span> <span class="o">*</span> <span class="p">[</span><span class="n">hgain</span><span class="p">,</span> <span class="n">sgain</span><span class="p">,</span> <span class="n">vgain</span><span class="p">]</span> <span class="c1"># random gains</span>
- <span class="n">hsv_augs</span> <span class="o">*=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span> <span class="c1"># random selection of h, s, v</span>
- <span class="n">hsv_augs</span> <span class="o">=</span> <span class="n">hsv_augs</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">int16</span><span class="p">)</span>
- <span class="n">img_hsv</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">cvtColor</span><span class="p">(</span><span class="n">img</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="n">bgr_channels</span><span class="p">],</span> <span class="n">cv2</span><span class="o">.</span><span class="n">COLOR_BGR2HSV</span><span class="p">)</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">int16</span><span class="p">)</span>
- <span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="n">hsv_augs</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span> <span class="o">%</span> <span class="mi">180</span>
- <span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="n">hsv_augs</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">255</span><span class="p">)</span>
- <span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">img_hsv</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="mi">2</span><span class="p">]</span> <span class="o">+</span> <span class="n">hsv_augs</span><span class="p">[</span><span class="mi">2</span><span class="p">],</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">255</span><span class="p">)</span>
- <span class="n">img</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="n">bgr_channels</span><span class="p">]</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">cvtColor</span><span class="p">(</span><span class="n">img_hsv</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">dtype</span><span class="p">),</span> <span class="n">cv2</span><span class="o">.</span><span class="n">COLOR_HSV2BGR</span><span class="p">)</span> <span class="c1"># no return needed</span>
- <span class="k">def</span> <span class="nf">rescale_and_pad_to_size</span><span class="p">(</span><span class="n">img</span><span class="p">,</span> <span class="n">input_size</span><span class="p">,</span> <span class="n">swap</span><span class="o">=</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">),</span> <span class="n">pad_val</span><span class="o">=</span><span class="mi">114</span><span class="p">):</span>
- <span class="sd">"""</span>
- <span class="sd"> Rescales image according to minimum ratio between the target height /image height, target width / image width,</span>
- <span class="sd"> and pads the image to the target size.</span>
- <span class="sd"> :param img: Image to be rescaled</span>
- <span class="sd"> :param input_size: Target size</span>
- <span class="sd"> :param swap: Axis's to be rearranged.</span>
- <span class="sd"> :return: rescaled image, ratio</span>
- <span class="sd"> """</span>
- <span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span> <span class="o">==</span> <span class="mi">3</span><span class="p">:</span>
- <span class="n">padded_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="n">input_size</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">input_size</span><span class="p">[</span><span class="mi">1</span><span class="p">],</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span> <span class="o">*</span> <span class="n">pad_val</span>
- <span class="k">else</span><span class="p">:</span>
- <span class="n">padded_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">input_size</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span> <span class="o">*</span> <span class="n">pad_val</span>
- <span class="n">r</span> <span class="o">=</span> <span class="nb">min</span><span class="p">(</span><span class="n">input_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">/</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">input_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
- <span class="n">resized_img</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span>
- <span class="n">img</span><span class="p">,</span>
- <span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">r</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">r</span><span class="p">)),</span>
- <span class="n">interpolation</span><span class="o">=</span><span class="n">cv2</span><span class="o">.</span><span class="n">INTER_LINEAR</span><span class="p">,</span>
- <span class="p">)</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span>
- <span class="n">padded_img</span><span class="p">[:</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="n">r</span><span class="p">),</span> <span class="p">:</span> <span class="nb">int</span><span class="p">(</span><span class="n">img</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="n">r</span><span class="p">)]</span> <span class="o">=</span> <span class="n">resized_img</span>
- <span class="n">padded_img</span> <span class="o">=</span> <span class="n">padded_img</span><span class="o">.</span><span class="n">transpose</span><span class="p">(</span><span class="n">swap</span><span class="p">)</span>
- <span class="n">padded_img</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ascontiguousarray</span><span class="p">(</span><span class="n">padded_img</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
- <span class="k">return</span> <span class="n">padded_img</span><span class="p">,</span> <span class="n">r</span>
- </pre></div>
- </div>
- </div>
- <footer>
- <hr/>
- <div role="contentinfo">
- <p>© Copyright 2021, SuperGradients team.</p>
- </div>
- Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
- <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
- provided by <a href="https://readthedocs.org">Read the Docs</a>.
-
- </footer>
- </div>
- </div>
- </section>
- </div>
- <script>
- jQuery(function () {
- SphinxRtdTheme.Navigation.enable(true);
- });
- </script>
- </body>
- </html>
|
Models
Annotations
Connect to our Discord channel