ChangeLog 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282
  1. 2019-07-07 - V4.1.0
  2. * Added new renders Alto, LSTMBox, WordStrBox.
  3. * Added character boxes in hOCR output.
  4. * Added python training scripts (experimental) as alternative shell scripts.
  5. * Better support AVX / AVX2 / SSE.
  6. * Disable OpenMP support by default (see e.g. #1171, #1081).
  7. * Fix for bounding box problem.
  8. * Implemented support for whitelist/blacklist in LSTM engine.
  9. * Improved cmake configuration.
  10. * Code modernization and improvements.
  11. * A lot of bug fixes...
  12. 2018-10-29 - V4.0.0
  13. * Added new neural network system based on LSTMs, with major accuracy gains.
  14. * Improvements to PDF rendering.
  15. * Fixes to trainingdata rendering.
  16. * Added LSTM models+lang models to 101 languages. (tessdata repository)
  17. * Improved multi-page TIFF handling.
  18. * Fixed damage to binary images when processing PDFs.
  19. * Fixes to training process to allow incremental training from a recognition model.
  20. * Made LSTM the default engine, pushed cube out.
  21. * Deleted cube code.
  22. * Changed OEModes --oem 0 for legacy tesseract engine, --oem 1 for LSTM, --oem 2 for both, --oem 3 for default.
  23. * Avoid use of Leptonica debug parameters or functions.
  24. * Fixed multi-language mode.
  25. * Removed support for VS2010.
  26. * Added Support for VS2015 and VS2017 with CPPAN.
  27. * Implemented invisible text only for PDF.
  28. * Added AVX / SSE support for Windows.
  29. * Enabled OpenMP support.
  30. * Parameter unlv_tilde_crunching change to false.
  31. * Miscellaneous Fixes.
  32. * Detailed Changelog can be found at https://tesseract-ocr.github.io/tessdoc/4.0x-Changelog.html and https://tesseract-ocr.github.io/tessdoc/ReleaseNotes.html#tesseract-release-notes-oct-29-2018---v400
  33. 2017-02-16 - V3.05.00
  34. * Made some fine tuning to the hOCR output.
  35. * Added TSV as another optional output format.
  36. * Fixed ABI break introduced in 3.04.00 with the AnalyseLayout() method.
  37. * text2image tool - Enable all OpenType ligatures available in a font. This feature requires Pango 1.38 or newer.
  38. * Training tools - Replaced asserts with tprintf() and exit(1).
  39. * Fixed Cygwin compatibility.
  40. * Improved multipage tiff processing.
  41. * Improved the embedded pdf font (pdf.ttf).
  42. * Enable selection of OCR engine mode from command line.
  43. * Changed tesseract command line parameter '-psm' to '--psm'.
  44. * Write output of tesseract --help, --version and --list-langs to stdout instead of stderr.
  45. * Added new C API for orientation and script detection, removed the old one.
  46. * Increased minimum autoconf version to 2.59.
  47. * Removed dead code.
  48. * Require Leptonica 1.74 or higher.
  49. * Fixed many compiler warning.
  50. * Fixed memory and resource leaks.
  51. * Fixed some issues with the 'Cube' OCR engine.
  52. * Fixed some openCL issues.
  53. * Added option to build Tesseract with CMake build system.
  54. * Implemented CPPAN support for easy Windows building.
  55. 2016-02-17 - V3.04.01
  56. * Added OSD renderer for psm 0. Works for single page and multi-page images.
  57. * Improve tesstrain.sh script.
  58. * Simplify build and run of ScrollView.
  59. * Improved PDF output for OS X Preview utility.
  60. * INCOMPATIBLE fix to hOCR line height information - commit 134ebc3.
  61. * Added option to build Tesseract without Cube OCR engine (-DNO_CUBE_BUILD).
  62. * Enable OpenMP support.
  63. * Many bug fixes.
  64. 2015-07-11 - V3.04.00
  65. * Tesseract development is now done with Git and hosted at github.com (Previously we used Subversion as a VCS and code.google.com for hosting).
  66. * Tesseract now requires leptonica 1.71 or a higher version.
  67. * Removed official support for VS 2008.
  68. * Added support for 39 additional scripts/languages, including: amh, asm, aze_cyrl, bod, bos, ceb, cym, dzo, fas, gle, guj, hat, iku, jav, kat, kat_old, kaz, khm, kir, kur, lao, lat, mar, mya, nep, ori, pan, pus, san, sin, srp_latn, syr, tgk, tir, uig, urd, uzb, uzb_cyrl, yid
  69. * Major updates to training system as a result of extensive testing on 100 languages.
  70. * New training data for over 100 languages
  71. * Improved performance with PIC compilation option.
  72. * Significant change to invisible font system in pdf output to improve correctness and compatibility with external programs, particularly ghostscript.
  73. * Improved font identification.
  74. * Major change to improve layout analysis for heavily diacritic languages: Thai, Vietnamese, Kannada, Telugu etc.
  75. * Fixed problems with shifted baselines so recognition can recover from layout analysis errors.
  76. * Major refactor to improve speed on difficult images, especially when running a heap checker.
  77. * Moved params from global in page layout to tesseractclass.
  78. * Improved single column layout analysis.
  79. * Allow ocr output to multiple formats using tesseract command line executable.
  80. * Fixed issues with mixed eng+ara scripts.
  81. * Improved script consistency in numbers.
  82. * Major refactor of control.cpp to enable line recognition.
  83. * Added tesstrain.sh - a master training script.
  84. * Added ability to text2image training tool to just list available fonts.
  85. * Added ability to text2image to underline words.
  86. * Improved efficiency of image processing for PDF output.
  87. * Added parameter description for each parameter listed with 'print-parameters' command line option.
  88. * Added font info to hOCR output.
  89. * Enabled streaming input and output of multi-page documents.
  90. * Many bug fixes.
  91. 2014-02-04 - V3.03(rc1)
  92. * Added new training tool text2image to generate box/tif file pairs from
  93. text and truetype fonts.
  94. * Added support for PDF output with searchable text.
  95. * Removed entire IMAGE class and all code in image directory.
  96. * Tesseract executable: support for output to stdout; limited support for one
  97. page images from stdin (especially on Windows)
  98. * Added Renderer to API to allow document-level processing and output
  99. of document formats, like hOCR, PDF.
  100. * Major refactor of word-level recognition, beam search, eliminating dead code.
  101. * Refactored classifier to make it easier to add new ones.
  102. * Generalized feature extractor to allow feature extraction from greyscale.
  103. * Improved sub/superscript treatment.
  104. * Improved baseline fit.
  105. * Added set_unicharset_properties to training tools.
  106. * Many bug fixes.
  107. * More training source data included.
  108. 2012-02-01 - V3.02
  109. * Moved ResultIterator/PageIterator to ccmain.
  110. * Added Right-to-left/Bidi capability in the output iterators for Hebrew/Arabic.
  111. * Added paragraph detection in layout analysis/post OCR.
  112. * Fixed inconsistent xheight during training and over-chopping.
  113. * Added simultaneous multi-language capability.
  114. * Refactored top-level word recognition module.
  115. * Added experimental equation detector.
  116. * Improved handling of resolution from input images.
  117. * Blamer module added for error analysis.
  118. * Cleaned up externally used namespace by removing includes from baseapi.h.
  119. * Removed dead memory mangagement code.
  120. * Tidied up constraints on control parameters.
  121. * Added support for ShapeTable in classifier and training.
  122. * Refactored class pruner.
  123. * Fixed training leaks and randomness.
  124. * Major improvements to layout analysis for better image detection, diacritic detection, better textline finding, better tabstop finding.
  125. * Improved line detection and removal.
  126. * Added fixed pitch chopper for CJK.
  127. * Added UNICHARSET to WERD_CHOICE to make mult-language handling easier.
  128. * Fixed problems with internally scaled images.
  129. * Added page and bbox to string in tr files to identify source of training data better.
  130. * Fixes to Hindi Shiroreka splitter.
  131. * Added word bigram correction.
  132. * Reduced stack memory consumption and eliminated some ugly typedefs.
  133. * Added new uniform classifier API.
  134. * Added new training error counter.
  135. * Fixed endian bug in dawg reader.
  136. * Many other fixes, including the way in which the chopper finds chops and messes with the outline while it does so.
  137. 2010-11-29 - V3.01
  138. * Removed old/dead serialise/deserialze methods on *LISTIZED classes.
  139. * Total rewrite of DENORM to better encapsulate operation and make
  140. for potential to extract features from images.
  141. * Thread-safety! Moved all critical globals and statics to members of the appropriate class. Tesseract is now thread-safe (multiple instances can be used in parallel in multiple threads.) with the minor exception that some control parameters are still global and affect all threads.
  142. * Added Cube, a new recognizer for Arabic. Cube can also be used in combination with normal Tesseract for other languages with an improvement in accuracy at the cost of (much) lower speed. *There is no training module for Cube yet.*
  143. * `OcrEngineMode` in `Init` replaces `AccuracyVSpeed` to control cube.
  144. * Greatly improved segmentation search with consequent accuracy and speed improvements, especially for Chinese.
  145. * Added `PageIterator` and `ResultIterator` as cleaner ways to get the full results out of Tesseract, that are not currently provided by any of the `TessBaseAPI::Get*` methods. All other methods, such as the `ETEXT_STRUCT` in particular are deprecated and will be deleted in the future.
  146. * ApplyBoxes totally rewritten to make training easier. It can now cope with touching/overlapping training characters, and a new boxfile format allows word boxes instead of character boxes, BUT to use that you have to have already boostrapped the language with character boxes. "Cyclic dependency" on traineddata.
  147. * Auto orientation and script detection added to page layout analysis.
  148. * Deleted *lots* of dead code.
  149. * Fixxht module replaced with scalable data-driven module.
  150. * Output font characteristics accuracy improved.
  151. * Removed the double conversion at each classification.
  152. * Upgraded oldest structs to be classes and deprecated PBLOB.
  153. * Removed non-deterministic baseline fit.
  154. * Added fixed length dawgs for Chinese.
  155. * Handling of vertical text improved.
  156. * Handling of leader dots improved.
  157. * Table detection greatly improved.
  158. * Fixed a couple of memory leaks.
  159. * Fixed font labels on output text. (Not perfect, but a lot better than before.)
  160. * Cleanup and more bug fixes
  161. * Special treatments for Hindi.
  162. * Support for build in VS2010 with Microsoft Windows SDK for Windows 7 (thanks to Michael Lutz)
  163. 2010-09-21 - V3.00
  164. * Preparations for thread safety:
  165. * Changed TessBaseAPI methods to be non-static
  166. * Created a class hierarchy for the directories to hold instance data,
  167. and began moving code into the classes.
  168. * Moved thresholding code to a separate class.
  169. * Added major new page layout analysis module.
  170. * Added HOCR output (issues 221, 263: thanks to amkryukov).
  171. * Added Leptonica as main image I/O and handling. Currently optional,
  172. but in future releases linking with Leptonica will be mandatory.
  173. * Ambiguity table rewritten to allow definite replacements in place
  174. of fix_quotes.
  175. * Added TessdataManager to combine data files into a single file.
  176. * Some dead code deleted.
  177. * VC++6 no longer supported. It can't cope with the use of templates.
  178. * Many more languages added.
  179. * Doxygenation of most of the function header comments.
  180. * Added man pages.
  181. * Added bash completion script (issue 247: thanks to neskiem)
  182. * Fix integer overview in thresholding (issue 366: thanks to Cyanide.Drake)
  183. * Add Danish Fraktur support (issues 300, 360: thanks to
  184. dsl602230@vip.cybercity.dk)
  185. * Fix file pointer leak (issue 359, thanks to yukihiro.nakadaira)
  186. * Fix an error using user-words (Issue 345: thanks to max.markin)
  187. * Fix a memory leak in tablefind.cpp (Issue 342, thanks to zdravco)
  188. * Fix a segfault due to double fclose (Issue 320, thanks to souther)
  189. * Fix an automake error (Issue 318, thanks to ichanjz)
  190. * Fix a Win32 crash on fileFormatIsTiff() (Issues 304, 316, 317, 330, 347,
  191. 349, 352: thanks to nguyenq87, max.markin, zdenop)
  192. * Fixed a number of errors in newer (stricter) versions of VC++ (Issues
  193. 301, among others)
  194. 2009-06-30 - V2.04
  195. * Integrated bug fixes and patches and misc changes for portability.
  196. * Integrated a patch to remove some of the "access" macros.
  197. * Removed dependence on lua from the viewer, speeding it up
  198. dramatically.
  199. * Fixed the viewer so it compiles and runs properly!
  200. * Specifically fixing issues: 1, 63, 67, 71, 76, 81, 82, 106, 111,
  201. 112, 128, 129, 130, 133, 135, 142, 143, 145, 147, 153, 154, 160,
  202. 165, 170, 175, 177, 187, 192, 195, 199, 201, 205, 209, 108, 169
  203. 2008-04-22 - V2.03
  204. * Fixed crash introduced in 2.02.
  205. * Fixed lack of tessembedded.cpp in distribution.
  206. * Added test for leptonica header files and conditional test for lib.
  207. 2008-04-21 - V2.02 (again)
  208. * Fixed namespace collisions with jpeg library (INT32).
  209. * Portability fixes for Windows for new code.
  210. * Updates to autoconf system for new code.
  211. 2008-01-23 - V2.02
  212. * Improvements to clustering, training and classifier.
  213. * Major internationalization improvements for large-character-set
  214. * languages, eg Kannada.
  215. * Removed some compiler warnings.
  216. * Added multipage tiff support for training and running.
  217. * Updated graphics output to talk to new java-based viewer.
  218. * Added ability to save n-best lists.
  219. * Added leptonica support for more file types.
  220. * Improved Init/End to make them safe.
  221. * Reduced memory use of dictionaries.
  222. * Added some new APIs to TessBaseAPI.
  223. 2007-08-27 - V2.01
  224. * Fixed UTF8 input problems with box file reader.
  225. * Fixed various infinite loops and crashes in dawg code.
  226. * Removed include of config_auto.h from host.h.
  227. * Added automatic wctype encoding to unicharset_extractor.
  228. * Fixed dawg table too full error.
  229. * Removed svn files from tarball.
  230. * Added new functions to tessdll.
  231. * Increased maximum utf8 string in a classification result to 8.
  232. 2007-07-02 - V2.00
  233. * Converted internal character handling to UTF8.
  234. * Trained with 6 languages.
  235. * Added unicharset_extractor, wordlist2dawg.
  236. * Added boxfile creation mode.
  237. * Added UNLV regression test capability.
  238. * Fixed problems with copyright and registered symbols.
  239. * Fixed extern "C" declarations problem.
  240. 2007-05-15 - V1.04
  241. * Added dll exports for Windows.
  242. * Fixed name collisions with stl etc.
  243. * Made some preliminary changes ready for unicodeization.
  244. * Several bug fixes discovered during unicodeization.
  245. 2007-02-02 - V1.03
  246. * Added mftraining and cntraining.
  247. * Added baseapi with adaptive thresholding for grey and color.
  248. * Fixed many memory leaks.
  249. * Fixed several bugs including lack of use of adaptive classifier.
  250. * Added ifdefs to eliminate graphics code and add embedded platform support.
  251. * Incorporated several patches, including 64-bit builds, Mac builds.
  252. * Minor accuracy improvements.
  253. 2006-10-04 - V1.02
  254. * Removed dependency on Aspirin.
  255. * Fixed a few missing Apache license headers.
  256. * Removed $log.
  257. 2006-09-07 - V1.01.
  258. * Added mfcpch.cpp and getopt.cpp for VC++.
  259. * Fixed problem with greyscale images and no libtiff.
  260. * Stopped debug window from being used for the usage output.
  261. * Fixed load of inttemp for big-endian architectures.
  262. * Fixed some Mac compilation issues.
  263. 2006-06-16 - V1.0 of open source Tesseract checked-in.
Tip!

Press p or to see the previous file or, n or to see the next file