Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel

fast-sam.md 10 KB

You have to be logged in to leave a comment. Sign In
comments description keywords
true FastSAM์€ ์ด๋ฏธ์ง€์—์„œ ์‹ค์‹œ๊ฐ„ ๊ฐ์ฒด ๋ถ„ํ• ์„ ์œ„ํ•œ CNN ๊ธฐ๋ฐ˜ ์†”๋ฃจ์…˜์œผ๋กœ, ํ–ฅ์ƒ๋œ ์‚ฌ์šฉ์ž ์ƒํ˜ธ์ž‘์šฉ, ๊ณ„์‚ฐ ํšจ์œจ์„ฑ, ๋‹ค์–‘ํ•œ ๋น„์ „ ์ž‘์—…์— ๋Œ€์‘ํ•  ์ˆ˜ ์žˆ๋Š” ํŠน์ง•์„ ๊ฐ–๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. FastSAM, ๋จธ์‹ ๋Ÿฌ๋‹, CNN ๊ธฐ๋ฐ˜ ์†”๋ฃจ์…˜, ๊ฐ์ฒด ๋ถ„ํ• , ์‹ค์‹œ๊ฐ„ ์†”๋ฃจ์…˜, Ultralytics, ๋น„์ „ ์ž‘์—…, ์ด๋ฏธ์ง€ ์ฒ˜๋ฆฌ, ์‚ฐ์—… ์‘์šฉ, ์‚ฌ์šฉ์ž ์ƒํ˜ธ์ž‘์šฉ

Fast Segment Anything Model (FastSAM)

Fast Segment Anything Model (FastSAM)์€ Segment Anything ์ž‘์—…์„ ์œ„ํ•œ ์ƒˆ๋กœ์šด ์‹ค์‹œ๊ฐ„ CNN ๊ธฐ๋ฐ˜ ์†”๋ฃจ์…˜์ž…๋‹ˆ๋‹ค. ์ด ์ž‘์—…์€ ๋‹ค์–‘ํ•œ ์‚ฌ์šฉ์ž ์ƒํ˜ธ์ž‘์šฉ ํ”„๋กฌํ”„ํŠธ์— ๋”ฐ๋ผ ์ด๋ฏธ์ง€ ๋‚ด์˜ ๋ชจ๋“  ๊ฐ์ฒด๋ฅผ ๋ถ„ํ• ํ•˜๋Š” ๊ฒƒ์„ ๋ชฉํ‘œ๋กœ ํ•ฉ๋‹ˆ๋‹ค. FastSAM์€ ๊ณ„์‚ฐ ์š”๊ตฌ ์‚ฌํ•ญ์„ ํฌ๊ฒŒ ์ค„์ด๋ฉด์„œ ๊ฒฝ์Ÿ๋ ฅ ์žˆ๋Š” ์„ฑ๋Šฅ์„ ์œ ์ง€ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ๋‹ค์–‘ํ•œ ๋น„์ „ ์ž‘์—…์— ์‹ค์šฉ์ ์ธ ์„ ํƒ์ง€๊ฐ€ ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Fast Segment Anything Model (FastSAM) ์•„ํ‚คํ…์ฒ˜ ๊ฐœ์š”

๊ฐœ์š”

FastSAM์€ ๊ณ„์‚ฐ ๋ฆฌ์†Œ์Šค ์š”๊ตฌ ์‚ฌํ•ญ์ด ํฐ Transformer ๋ชจ๋ธ์ธ Segment Anything Model (SAM)์˜ ํ•œ๊ณ„๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด ์„ค๊ณ„๋˜์—ˆ์Šต๋‹ˆ๋‹ค. FastSAM์€ Segment Anything ์ž‘์—…์„ ๋‘ ๋‹จ๊ณ„๋กœ ๋ถ„๋ฆฌํ•œ ๋ฐฉ์‹์„ ์ฑ„ํƒํ•ฉ๋‹ˆ๋‹ค: ๋ชจ๋“  ์ธ์Šคํ„ด์Šค ๋ถ„ํ• ๊ณผ ํ”„๋กฌํ”„ํŠธ๋กœ ์ธํ•œ ์˜์—ญ ์„ ํƒ. ์ฒซ ๋ฒˆ์งธ ๋‹จ๊ณ„์—์„œ๋Š” YOLOv8-seg๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ์ด๋ฏธ์ง€์˜ ๋ชจ๋“  ์ธ์Šคํ„ด์Šค์˜ ๋ถ„ํ•  ๋งˆ์Šคํฌ๋ฅผ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค. ๋‘ ๋ฒˆ์งธ ๋‹จ๊ณ„์—์„œ๋Š” ํ”„๋กฌํ”„ํŠธ์— ํ•ด๋‹นํ•˜๋Š” ๊ด€์‹ฌ ์˜์—ญ์„ ์ถœ๋ ฅํ•ฉ๋‹ˆ๋‹ค.

์ฃผ์š” ํŠน์ง•

  1. ์‹ค์‹œ๊ฐ„ ์†”๋ฃจ์…˜: CNN์˜ ๊ณ„์‚ฐ ํšจ์œจ์„ฑ์„ ํ™œ์šฉํ•˜์—ฌ FastSAM์€ Segment Anything ์ž‘์—…์— ๋Œ€ํ•œ ์‹ค์‹œ๊ฐ„ ์†”๋ฃจ์…˜์„ ์ œ๊ณตํ•˜๋ฉฐ, ๋น ๋ฅธ ๊ฒฐ๊ณผ๊ฐ€ ํ•„์š”ํ•œ ์‚ฐ์—… ์‘์šฉ์— ๊ฐ€์น˜๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.

  2. ํšจ์œจ์„ฑ๊ณผ ์„ฑ๋Šฅ: FastSAM์€ ์„ฑ๋Šฅ ํ’ˆ์งˆ์„ ํฌ์ƒํ•˜์ง€ ์•Š๊ณ  ๊ณ„์‚ฐ๊ณผ ๋ฆฌ์†Œ์Šค ์š”๊ตฌ ์‚ฌํ•ญ์„ ํฌ๊ฒŒ ์ค„์ž…๋‹ˆ๋‹ค. SAM๊ณผ ๋น„๊ตํ•ด ์œ ์‚ฌํ•œ ์„ฑ๋Šฅ์„ ๋‹ฌ์„ฑํ•˜๋ฉด์„œ ๊ณ„์‚ฐ ๋ฆฌ์†Œ์Šค๋ฅผ ํฌ๊ฒŒ ์ค„์—ฌ ์‹ค์‹œ๊ฐ„ ์‘์šฉ์ด ๊ฐ€๋Šฅํ•ด์ง‘๋‹ˆ๋‹ค.

  3. ํ”„๋กฌํ”„ํŠธ ์•ˆ๋‚ด ๋ถ„ํ• : FastSAM์€ ๋‹ค์–‘ํ•œ ์‚ฌ์šฉ์ž ์ƒํ˜ธ์ž‘์šฉ ํ”„๋กฌํ”„ํŠธ์— ๋”ฐ๋ผ ์ด๋ฏธ์ง€ ๋‚ด์˜ ๋ชจ๋“  ๊ฐ์ฒด๋ฅผ ๋ถ„ํ• ํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ ๋‹ค์–‘ํ•œ ์‹œ๋‚˜๋ฆฌ์˜ค์—์„œ ์œ ์—ฐ์„ฑ๊ณผ ์ ์‘์„ฑ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

  4. YOLOv8-seg ๊ธฐ๋ฐ˜: FastSAM์€ YOLOv8-seg๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•œ ๊ฒƒ์œผ๋กœ, ์ธ์Šคํ„ด์Šค ๋ถ„ํ•  ๋ธŒ๋žœ์น˜๊ฐ€ ์žฅ์ฐฉ๋œ ๊ฐ์ฒด ๊ฐ์ง€๊ธฐ์ž…๋‹ˆ๋‹ค. ์ด๋ฅผ ํ†ตํ•ด ์ด๋ฏธ์ง€์˜ ๋ชจ๋“  ์ธ์Šคํ„ด์Šค์˜ ๋ถ„ํ•  ๋งˆ์Šคํฌ๋ฅผ ํšจ๊ณผ์ ์œผ๋กœ ์ƒ์„ฑํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

  5. ๋ฒค์น˜๋งˆํฌ์—์„œ ๊ฒฝ์Ÿ ๊ฒฐ๊ณผ: MS COCO์—์„œ์˜ ๊ฐ์ฒด ์ œ์•ˆ ์ž‘์—…์—์„œ FastSAM์€ SAM์— ๋น„ํ•ด ๋‹จ์ผ NVIDIA RTX 3090์—์„œ ํ›จ์”ฌ ๋” ๋น ๋ฅธ ์†๋„๋กœ ๋†’์€ ์ ์ˆ˜๋ฅผ ๋‹ฌ์„ฑํ•˜์—ฌ ํšจ์œจ์„ฑ๊ณผ ๋Šฅ๋ ฅ์„ ์ž…์ฆํ–ˆ์Šต๋‹ˆ๋‹ค.

  6. ์‹ค์šฉ์ ์ธ ์‘์šฉ: FastSAM์€ ํ˜„์žฌ ๋ฐฉ๋ฒ•๋ณด๋‹ค ์ˆ˜์‹ญ ๋ฐฐ ๋˜๋Š” ์ˆ˜๋ฐฑ ๋ฐฐ ๋” ๋น ๋ฅธ ์†๋„๋กœ ์—ฌ๋Ÿฌ ๋น„์ „ ์ž‘์—…์˜ ์‹ ์†ํ•œ ์†”๋ฃจ์…˜์„ ์ œ๊ณตํ•˜์—ฌ ์‹ค์งˆ์ ์ธ ์ ์šฉ ๊ฐ€๋Šฅ์„ฑ์„ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค.

  7. ๋ชจ๋ธ ์••์ถ• ๊ฐ€๋Šฅ์„ฑ: FastSAM์€ ๊ตฌ์กฐ์— ์ธ๊ณต ์‚ฌ์ „์„ ๋„์ž…ํ•˜์—ฌ ๊ณ„์‚ฐ ๋น„์šฉ์„ ํฌ๊ฒŒ ์ค„์ผ ์ˆ˜ ์žˆ๋Š” ๊ฒฝ๋กœ๋ฅผ ๋ณด์—ฌ์ฃผ์–ด ์ผ๋ฐ˜ ๋น„์ „ ์ž‘์—…์— ๋Œ€ํ•œ ๋Œ€ํ˜• ๋ชจ๋ธ ์•„ํ‚คํ…์ฒ˜์— ๋Œ€ํ•œ ์ƒˆ๋กœ์šด ๊ฐ€๋Šฅ์„ฑ์„ ์—ด์–ด์ค๋‹ˆ๋‹ค.

์‚ฌ์šฉ ๊ฐ€๋Šฅํ•œ ๋ชจ๋ธ, ์ง€์›๋˜๋Š” ์ž‘์—… ๋ฐ ์šด์˜ ๋ชจ๋“œ

์ด ํ‘œ๋Š” ์‚ฌ์šฉ ๊ฐ€๋Šฅํ•œ ๋ชจ๋ธ๊ณผ ํ•ด๋‹นํ•˜๋Š” ์‚ฌ์ „ ํ›ˆ๋ จ ๊ฐ€์ค‘์น˜, ์ง€์›ํ•˜๋Š” ์ž‘์—… ๋ฐ Inference, Validation, Training, Export์™€ ๊ฐ™์€ ๋‹ค๋ฅธ ์šด์˜ ๋ชจ๋“œ์— ๋Œ€ํ•œ ํ˜ธํ™˜์„ฑ์„ ๋‚˜ํƒ€๋‚ด๋ฉฐ, ์ง€์›๋˜๋Š” ๋ชจ๋“œ๋Š” โœ… ์ด๋ชจ์ง€๋กœ, ์ง€์›๋˜์ง€ ์•Š๋Š” ๋ชจ๋“œ๋Š” โŒ ์ด๋ชจ์ง€๋กœ ํ‘œ์‹œ๋ฉ๋‹ˆ๋‹ค.

๋ชจ๋ธ ์œ ํ˜• ์‚ฌ์ „ ํ›ˆ๋ จ ๊ฐ€์ค‘์น˜ ์ง€์›๋˜๋Š” ์ž‘์—… Inference Validation Training Export
FastSAM-s FastSAM-s.pt ์ธ์Šคํ„ด์Šค ๋ถ„ํ•  โœ… โŒ โŒ โœ…
FastSAM-x FastSAM-x.pt ์ธ์Šคํ„ด์Šค ๋ถ„ํ•  โœ… โŒ โŒ โœ…

์‚ฌ์šฉ ์˜ˆ์‹œ

FastSAM ๋ชจ๋ธ์„ Python ์• ํ”Œ๋ฆฌ์ผ€์ด์…˜์— ์‰ฝ๊ฒŒ ํ†ตํ•ฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. Ultralytics๋Š” ๊ฐœ๋ฐœ์„ ๊ฐ„์†Œํ™”ํ•˜๊ธฐ ์œ„ํ•ด ์‚ฌ์šฉ์ž ์นœํ™”์ ์ธ Python API ๋ฐ CLI ๋ช…๋ น์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

์˜ˆ์ธก ์‚ฌ์šฉ๋ฒ•

์ด๋ฏธ์ง€์—์„œ ๊ฐ์ฒด ๊ฒ€์ถœ์„ ์ˆ˜ํ–‰ํ•˜๋ ค๋ฉด ๋‹ค์Œ๊ณผ ๊ฐ™์ด predict ๋ฉ”์„œ๋“œ๋ฅผ ์‚ฌ์šฉํ•ฉ๋‹ˆ๋‹ค:

!!! Example "์˜ˆ์ œ"

=== "Python"
    ```python
    from ultralytics import FastSAM
    from ultralytics.models.fastsam import FastSAMPrompt

    # ์ถ”๋ก  ์†Œ์Šค ์ •์˜
    source = 'path/to/bus.jpg'

    # FastSAM ๋ชจ๋ธ ์ƒ์„ฑ
    model = FastSAM('FastSAM-s.pt')  # ๋˜๋Š” FastSAM-x.pt

    # ์ด๋ฏธ์ง€์— ๋Œ€ํ•œ ์ถ”๋ก  ์‹คํ–‰
    everything_results = model(source, device='cpu', retina_masks=True, imgsz=1024, conf=0.4, iou=0.9)

    # Prompt Process ๊ฐ์ฒด ์ค€๋น„
    prompt_process = FastSAMPrompt(source, everything_results, device='cpu')

    # ๋ชจ๋“  ํ”„๋กฌํ”„ํŠธ
    ann = prompt_process.everything_prompt()

    # ๋ฐ”์šด๋”ฉ ๋ฐ•์Šค์˜ ๊ธฐ๋ณธ ๋ชจ์–‘์€ [0,0,0,0]์—์„œ [x1,y1,x2,y2]๋กœ ๋ณ€๊ฒฝ
    ann = prompt_process.box_prompt(bbox=[200, 200, 300, 300])

    # ํ…์ŠคํŠธ ํ”„๋กฌํ”„ํŠธ
    ann = prompt_process.text_prompt(text='a photo of a dog')

    # ํฌ์ธํŠธ ํ”„๋กฌํ”„ํŠธ
    # ๊ธฐ๋ณธ ํฌ์ธํŠธ [[0,0]] [[x1,y1],[x2,y2]]
    # ๊ธฐ๋ณธ ํฌ์ธํŠธ ๋ ˆ์ด๋ธ” [0] [1,0] 0:๋ฐฐ๊ฒฝ, 1:์ „๊ฒฝ
    ann = prompt_process.point_prompt(points=[[200, 200]], pointlabel=[1])
    prompt_process.plot(annotations=ann, output='./')
    ```

=== "CLI"
    ```bash
    # FastSAM ๋ชจ๋ธ ๋กœ๋“œ ๋ฐ ๋ชจ๋“  ๊ฒƒ์„ ์„ธ๋ถ„ํ™”ํ•˜์—ฌ ์ถ”์ถœ
    yolo segment predict model=FastSAM-s.pt source=path/to/bus.jpg imgsz=640
    ```

์ด ์ฝ”๋“œ ์กฐ๊ฐ์€ ์‚ฌ์ „ ํ›ˆ๋ จ๋œ ๋ชจ๋ธ์„ ๋กœ๋“œํ•˜๊ณ  ์ด๋ฏธ์ง€์— ๋Œ€ํ•œ ์˜ˆ์ธก์„ ์‹คํ–‰ํ•˜๋Š” ๊ฐ„ํŽธํ•จ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค.

๊ฒ€์ฆ ์‚ฌ์šฉ๋ฒ•

๋ฐ์ดํ„ฐ์…‹์—์„œ ๋ชจ๋ธ์„ ๊ฒ€์ฆํ•˜๋Š” ๋ฐฉ๋ฒ•์€ ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค:

!!! Example "์˜ˆ์ œ"

=== "Python"
    ```python
    from ultralytics import FastSAM

    # FastSAM ๋ชจ๋ธ ์ƒ์„ฑ
    model = FastSAM('FastSAM-s.pt')  # ๋˜๋Š” FastSAM-x.pt

    # ๋ชจ๋ธ ๊ฒ€์ฆ
    results = model.val(data='coco8-seg.yaml')
    ```

=== "CLI"
    ```bash
    # FastSAM ๋ชจ๋ธ ๋กœ๋“œ ๋ฐ ์ด๋ฏธ์ง€ ํฌ๊ธฐ 640์—์„œ COCO8 ์˜ˆ์ œ ๋ฐ์ดํ„ฐ์…‹์— ๋Œ€ํ•ด ์œ ํšจ์„ฑ ๊ฒ€์‚ฌ
    yolo segment val model=FastSAM-s.pt data=coco8.yaml imgsz=640
    ```

FastSAM์€ ๋‹จ์ผ ํด๋ž˜์Šค ๊ฐ์ฒด์˜ ๊ฐ์ง€์™€ ๋ถ„ํ• ๋งŒ ์ง€์›ํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ๋ชจ๋“  ๊ฐ์ฒด๋ฅผ ๋™์ผํ•œ ํด๋ž˜์Šค๋กœ ์ธ์‹ํ•˜๊ณ  ๋ถ„ํ• ํ•œ๋‹ค๋Š” ์˜๋ฏธ์ž…๋‹ˆ๋‹ค. ๋”ฐ๋ผ์„œ ๋ฐ์ดํ„ฐ์…‹์„ ์ค€๋น„ํ•  ๋•Œ ๋ชจ๋“  ๊ฐ์ฒด ์นดํ…Œ๊ณ ๋ฆฌ ID๋ฅผ 0์œผ๋กœ ๋ณ€ํ™˜ํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.

FastSAM ๊ณต์‹ ์‚ฌ์šฉ๋ฒ•

FastSAM์€ https://github.com/CASIA-IVA-Lab/FastSAM ์ €์žฅ์†Œ์—์„œ ์ง์ ‘ ์‚ฌ์šฉํ•  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค. FastSAM์„ ์‚ฌ์šฉํ•˜๊ธฐ ์œ„ํ•ด ์ˆ˜ํ–‰ํ•  ์ผ๋ฐ˜์ ์ธ ๋‹จ๊ณ„๋ฅผ ๊ฐ„๋‹จํžˆ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค:

์„ค์น˜

  1. FastSAM ์ €์žฅ์†Œ๋ฅผ ๋ณต์ œํ•ฉ๋‹ˆ๋‹ค:

    git clone https://github.com/CASIA-IVA-Lab/FastSAM.git
    
  2. Python 3.9๋กœ Conda ํ™˜๊ฒฝ์„ ์ƒ์„ฑํ•˜๊ณ  ํ™œ์„ฑํ™”ํ•ฉ๋‹ˆ๋‹ค:

    conda create -n FastSAM python=3.9
    conda activate FastSAM
    
  3. ๋ณต์ œํ•œ ์ €์žฅ์†Œ๋กœ ์ด๋™ํ•˜์—ฌ ํ•„์š”ํ•œ ํŒจํ‚ค์ง€๋ฅผ ์„ค์น˜ํ•ฉ๋‹ˆ๋‹ค:

    cd FastSAM
    pip install -r requirements.txt
    
  4. CLIP ๋ชจ๋ธ์„ ์„ค์น˜ํ•ฉ๋‹ˆ๋‹ค:

    pip install git+https://github.com/openai/CLIP.git
    

์˜ˆ์‹œ ์‚ฌ์šฉ๋ฒ•

  1. ๋ชจ๋ธ ์ฒดํฌํฌ์ธํŠธ๋ฅผ ๋‹ค์šด๋กœ๋“œํ•ฉ๋‹ˆ๋‹ค.

  2. FastSAM์„ ์ถ”๋ก ํ•˜๊ธฐ ์œ„ํ•ด ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์‚ฌ์šฉํ•ฉ๋‹ˆ๋‹ค. ์˜ˆ์‹œ ๋ช…๋ น์–ด:

    • ์ด๋ฏธ์ง€์—์„œ ๋ชจ๋“  ๊ฒƒ์„ ์„ธ๋ถ„ํ™”:

      python Inference.py --model_path ./weights/FastSAM.pt --img_path ./images/dogs.jpg
      
    • ํ…์ŠคํŠธ ํ”„๋กฌํ”„ํŠธ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ํŠน์ • ๊ฐ์ฒด๋ฅผ ์„ธ๋ถ„ํ™”:

      python Inference.py --model_path ./weights/FastSAM.pt --img_path ./images/dogs.jpg --text_prompt "the yellow dog"
      
    • ๋ฐ”์šด๋”ฉ ๋ฐ•์Šค ๋‚ด์˜ ๊ฐ์ฒด๋ฅผ ์„ธ๋ถ„ํ™” (xywh ํ˜•์‹์œผ๋กœ ์ƒ์ž ์ขŒํ‘œ ์ œ๊ณต):

      python Inference.py --model_path ./weights/FastSAM.pt --img_path ./images/dogs.jpg --box_prompt "[570,200,230,400]"
      
    • ํŠน์ • ์ง€์  ๊ทผ์ฒ˜์˜ ๊ฐ์ฒด๋ฅผ ์„ธ๋ถ„ํ™”:

      python Inference.py --model_path ./weights/FastSAM.pt --img_path ./images/dogs.jpg --point_prompt "[[520,360],[620,300]]" --point_label "[1,0]"
      

๋˜ํ•œ, FastSAM์„ Colab ๋ฐ๋ชจ ๋˜๋Š” HuggingFace ์›น ๋ฐ๋ชจ์—์„œ ์‹œ๊ฐ์ ์ธ ๊ฒฝํ—˜์œผ๋กœ ์‹œ๋„ํ•ด ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์ธ์šฉ ๋ฐ ๊ฐ์‚ฌ์˜ ๋ง์”€

FastSAM์˜ ์‹ค์‹œ๊ฐ„ ์ธ์Šคํ„ด์Šค ๋ถ„ํ•  ๋ถ„์•ผ์— ๋Œ€ํ•œ ํ˜์‹ ์ ์ธ ๊ธฐ์—ฌ๋ฅผ ์œ„ํ•ด FastSAM ์ €์ž๋“ค์—๊ฒŒ ๊ฐ์‚ฌ์˜ ๋ง์”€์„ ์ „ํ•ฉ๋‹ˆ๋‹ค:

!!! Quote ""

=== "BibTeX"

  ```bibtex
  @misc{zhao2023fast,
        title={Fast Segment Anything},
        author={Xu Zhao and Wenchao Ding and Yongqi An and Yinglong Du and Tao Yu and Min Li and Ming Tang and Jinqiao Wang},
        year={2023},
        eprint={2306.12156},
        archivePrefix={arXiv},
        primaryClass={cs.CV}
  }
  ```

FastSAM ์›๋ณธ ๋…ผ๋ฌธ์€ arXiv์—์„œ ์ฐพ์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ €์ž๋“ค์€ ์ž์‹ ๋“ค์˜ ์ž‘์—…์„ ๊ณต๊ฐœ์ ์œผ๋กœ ์ œ๊ณตํ•˜์˜€์œผ๋ฉฐ, ์ฝ”๋“œ๋ฒ ์ด์Šค๋Š” GitHub์—์„œ ์ด์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ €์ž๋“ค์˜ ๋…ธ๋ ฅ์— ๊ฐ์‚ฌ๋“œ๋ฆฌ๋ฉฐ ์ €์ž‘๋ฌผ์„ ๋” ํญ๋„“์€ ์ปค๋ฎค๋‹ˆํ‹ฐ์— ์•Œ๋ฆฌ๊ธฐ ์œ„ํ•œ ๊ธฐ์—ฌ๋ฅผ ๊ธฐ๋Œ€ํ•ฉ๋‹ˆ๋‹ค.

Tip!

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

Comments

Loading...