์ฝ˜ํ…์ธ ๋กœ ๊ฑด๋„ˆ๋›ฐ๊ธฐ

Roboflow

Roboflow has everything you need to build and deploy computer vision models. Connect Roboflow at any step in your pipeline with APIs and SDKs, or use the end-to-end interface to automate the entire process from image to inference. Whether you're in need of data labeling, model training, or model deployment, Roboflow gives you building blocks to bring custom computer vision solutions to your project.

๋ผ์ด์„ ์Šค

Ultralytics ๋Š” ๋‘ ๊ฐ€์ง€ ๋ผ์ด์„ ์Šค ์˜ต์…˜์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค:

์ž์„ธํ•œ ๋‚ด์šฉ์€ Ultralytics ๋ผ์ด์„ ์Šค๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”.

In this guide, we are going to showcase how to find, label, and organize data for use in training a custom Ultralytics YOLO11 model. Use the table of contents below to jump directly to a specific section:

  • Gather data for training a custom YOLO11 model
  • Upload, convert and label data for YOLO11 format
  • ๋ชจ๋ธ ๊ฒฌ๊ณ ์„ฑ์„ ์œ„ํ•œ ๋ฐ์ดํ„ฐ ์‚ฌ์ „ ์ฒ˜๋ฆฌ ๋ฐ ๋ณด๊ฐ•
  • Dataset management for YOLO11
  • ๋ชจ๋ธ ํ•™์Šต์„ ์œ„ํ•ด 40๊ฐœ ์ด์ƒ์˜ ํ˜•์‹์œผ๋กœ ๋ฐ์ดํ„ฐ ๋‚ด๋ณด๋‚ด๊ธฐ
  • Upload custom YOLO11 model weights for testing and deployment
  • Gather Data for Training a Custom YOLO11 Model

Roboflow provides two services that can help you collect data for YOLO11 models: Universe and Collect.

Universe๋Š” ์ด 1์–ต ๊ฐœ๊ฐ€ ๋„˜๋Š” ์ด๋ฏธ์ง€๊ฐ€ ํฌํ•จ๋œ 25๋งŒ ๊ฐœ ์ด์ƒ์˜ ๋น„์ „ ๋ฐ์ดํ„ฐ ์„ธํŠธ๊ฐ€ ์žˆ๋Š” ์˜จ๋ผ์ธ ์ €์žฅ์†Œ์ž…๋‹ˆ๋‹ค.

Roboflow ์šฐ์ฃผ

๋ฌด๋ฃŒ ๊ณ„์ •( Roboflow )์„ ์‚ฌ์šฉํ•˜๋ฉด ์œ ๋‹ˆ๋ฒ„์Šค์—์„œ ์‚ฌ์šฉ ๊ฐ€๋Šฅํ•œ ๋ชจ๋“  ๋ฐ์ดํ„ฐ์„ธํŠธ๋ฅผ ๋‚ด๋ณด๋‚ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฐ์ดํ„ฐ์„ธํŠธ๋ฅผ ๋‚ด๋ณด๋‚ด๋ ค๋ฉด ๋ฐ์ดํ„ฐ์„ธํŠธ์—์„œ "์ด ๋ฐ์ดํ„ฐ์„ธํŠธ ๋‹ค์šด๋กœ๋“œ" ๋ฒ„ํŠผ์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค.

Roboflow ์œ ๋‹ˆ๋ฒ„์Šค ๋ฐ์ดํ„ฐ ์„ธํŠธ ๋‚ด๋ณด๋‚ด๊ธฐ

For YOLO11, select "YOLO11" as the export format:

Roboflow ์œ ๋‹ˆ๋ฒ„์Šค ๋ฐ์ดํ„ฐ ์„ธํŠธ ๋‚ด๋ณด๋‚ด๊ธฐ

Universe also has a page that aggregates all public fine-tuned YOLO11 models uploaded to Roboflow. You can use this page to explore pre-trained models you can use for testing or for automated data labeling or to prototype with Roboflow inference.

์ด๋ฏธ์ง€๋ฅผ ์ง์ ‘ ์ˆ˜์ง‘ํ•˜๋ ค๋ฉด ์—ฃ์ง€์—์„œ ์›น์บ ์„ ์‚ฌ์šฉํ•˜์—ฌ ์ด๋ฏธ์ง€๋ฅผ ์ž๋™์œผ๋กœ ์ˆ˜์ง‘ํ•  ์ˆ˜ ์žˆ๋Š” ์˜คํ”ˆ ์†Œ์Šค ํ”„๋กœ์ ํŠธ์ธ Collect๋ฅผ ์‚ฌ์šฉํ•ด ๋ณด์„ธ์š”. Collect์—์„œ ํ…์ŠคํŠธ ๋˜๋Š” ์ด๋ฏธ์ง€ ํ”„๋กฌํ”„ํŠธ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ์ˆ˜์ง‘ํ•  ๋ฐ์ดํ„ฐ๋ฅผ ์ง€์‹œํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ ๋น„์ „ ๋ชจ๋ธ์„ ๊ตฌ์ถ•ํ•˜๋Š” ๋ฐ ํ•„์š”ํ•œ ์œ ์šฉํ•œ ๋ฐ์ดํ„ฐ๋งŒ ์บก์ฒ˜ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Upload, Convert and Label Data for YOLO11 Format

Roboflow Annotate is an online annotation tool for use in labeling images for object detection, classification, and segmentation.

To label data for a YOLO11 object detection, instance segmentation, or classification model, first create a project in Roboflow.

Roboflow ํ”„๋กœ์ ํŠธ ๋งŒ๋“ค๊ธฐ

๊ทธ๋Ÿฐ ๋‹ค์Œ, ์ด๋ฏธ์ง€์™€ ๋‹ค๋ฅธ ๋„๊ตฌ์—์„œ๊ฐ€์ ธ์˜จ ๊ธฐ์กด ์ฃผ์„(์ง€์›๋˜๋Š” 40๊ฐœ ์ด์ƒ์˜ ๊ฐ€์ ธ์˜ค๊ธฐ ํ˜•์‹ ์ค‘ ํ•˜๋‚˜ ์‚ฌ์šฉ)์„ Roboflow ์— ์—…๋กœ๋“œํ•ฉ๋‹ˆ๋‹ค.

๋‹ค์Œ ์ฃผ์†Œ๋กœ ์ด๋ฏธ์ง€ ์—…๋กœ๋“œ Roboflow

์ด๋ฏธ์ง€๋ฅผ ์—…๋กœ๋“œํ•œ ํ›„ ์ฃผ์„ ๋‹ฌ๊ธฐ ํŽ˜์ด์ง€์—์„œ ์—…๋กœ๋“œํ•œ ์ด๋ฏธ์ง€์˜ ์ผ๊ด„ ์ฒ˜๋ฆฌ๋ฅผ ์„ ํƒํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฐ ๋‹ค์Œ '์ฃผ์„ ๋‹ฌ๊ธฐ ์‹œ์ž‘'์„ ํด๋ฆญํ•˜์—ฌ ์ด๋ฏธ์ง€์— ๋ผ๋ฒจ์„ ์ง€์ •ํ•ฉ๋‹ˆ๋‹ค.

๊ฒฝ๊ณ„ ์ƒ์ž๋กœ ๋ ˆ์ด๋ธ”์„ ์ง€์ •ํ•˜๋ ค๋ฉด B key on your keyboard or click the box icon in the sidebar. Click on a point where you want to start your bounding box, then drag to create the box:

์—์„œ ์ด๋ฏธ์ง€์— ์ฃผ์„์„ ๋‹ฌ๊ธฐ Roboflow

์ฃผ์„์„ ์ƒ์„ฑํ•˜๋ฉด ์ฃผ์„์— ์‚ฌ์šฉํ•  ํด๋ž˜์Šค๋ฅผ ์„ ํƒํ•˜๋ผ๋Š” ํŒ์—…์ด ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค.

๋‹ค๊ฐํ˜•์œผ๋กœ ๋ ˆ์ด๋ธ”์„ ์ง€์ •ํ•˜๋ ค๋ฉด P ํ‚ค๋ฅผ ๋ˆ„๋ฅด๊ฑฐ๋‚˜ ์‚ฌ์ด๋“œ๋ฐ”์˜ ๋‹ค๊ฐํ˜• ์•„์ด์ฝ˜์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๋‹ค๊ฐํ˜• ์ฃผ์„ ๋„๊ตฌ๋ฅผ ํ™œ์„ฑํ™”ํ•œ ์ƒํƒœ์—์„œ ์ด๋ฏธ์ง€์˜ ๊ฐœ๋ณ„ ์ ์„ ํด๋ฆญํ•˜์—ฌ ๋‹ค๊ฐํ˜•์„ ๊ทธ๋ฆฝ๋‹ˆ๋‹ค.

Roboflow ๋Š” ์ด๋ฏธ์ง€์— ๋ผ๋ฒจ์„ ๋” ๋น ๋ฅด๊ฒŒ ๋ถ™์ผ ์ˆ˜ ์žˆ๋Š” SAM ๊ธฐ๋ฐ˜ ๋ผ๋ฒจ ์–ด์‹œ์Šคํ„ดํŠธ๋ฅผ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค. SAM (Segment Anything Model)์€ ์ด๋ฏธ์ง€์— ์ •๋ฐ€ํ•˜๊ฒŒ ๋ผ๋ฒจ์„ ๋ถ™์ผ ์ˆ˜ ์žˆ๋Š” ์ตœ์ฒจ๋‹จ ์ปดํ“จํ„ฐ ๋น„์ „ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. SAM ์„ ์‚ฌ์šฉํ•˜๋ฉด ์ด๋ฏธ์ง€ ๋ผ๋ฒจ๋ง ํ”„๋กœ์„ธ์Šค์˜ ์†๋„๋ฅผ ํฌ๊ฒŒ ๋†’์ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฌผ์ฒด ์ฃผ๋ณ€์˜ ์ ์„ ์ •ํ™•ํ•˜๊ฒŒ ํด๋ฆญํ•˜๋Š” ์ง€๋ฃจํ•œ ๊ณผ์ • ๋Œ€์‹  ๋ช‡ ๋ฒˆ์˜ ํด๋ฆญ๋งŒ์œผ๋กœ ๋‹ค๊ฐํ˜•์œผ๋กœ ์ด๋ฏธ์ง€์— ์ฃผ์„์„ ๋‹ฌ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋ผ๋ฒจ ๋„์šฐ๋ฏธ๋ฅผ ์‚ฌ์šฉํ•˜๋ ค๋ฉด ์‚ฌ์ด๋“œ๋ฐ”์—์„œ ์ปค์„œ ์•„์ด์ฝ˜์„ ํด๋ฆญํ•˜๋ฉด ํ”„๋กœ์ ํŠธ์—์„œ ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋„๋ก SAM ์ด ๋กœ๋“œ๋ฉ๋‹ˆ๋‹ค.

SAM- ๊ธฐ๋ฐ˜ ๋ ˆ์ด๋ธ” ์ง€์›์œผ๋กœ Roboflow ์—์„œ ์ด๋ฏธ์ง€์— ์ฃผ์„ ๋‹ฌ๊ธฐ

์ด๋ฏธ์ง€์˜ ๊ฐœ์ฒด ์œ„๋กœ ๋งˆ์šฐ์Šค๋ฅผ ๊ฐ€์ ธ๊ฐ€๋ฉด SAM ์—์„œ ์ฃผ์„์„ ์ถ”์ฒœํ•ฉ๋‹ˆ๋‹ค. ๋งˆ์šฐ์Šค๋ฅผ ๊ฐ€์ ธ๊ฐ€์„œ ์ฃผ์„์„ ๋‹ฌ ์ ์ ˆํ•œ ์œ„์น˜๋ฅผ ์ฐพ์€ ๋‹ค์Œ ํด๋ฆญํ•˜์—ฌ ์ฃผ์„์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ฃผ์„์„ ๋” ๊ตฌ์ฒด์ ์œผ๋กœ ์ˆ˜์ •ํ•˜๋ ค๋ฉด SAM ์ด ๋ฌธ์„œ์— ๋งŒ๋“  ์ฃผ์„์˜ ์•ˆ์ชฝ ๋˜๋Š” ๋ฐ”๊นฅ์ชฝ์„ ํด๋ฆญํ•˜๋ฉด ๋ฉ๋‹ˆ๋‹ค.

์‚ฌ์ด๋“œ๋ฐ”์˜ ํƒœ๊ทธ ํŒจ๋„์—์„œ ์ด๋ฏธ์ง€์— ํƒœ๊ทธ๋ฅผ ์ถ”๊ฐ€ํ•  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค. ํŠน์ • ์ง€์—ญ์˜ ๋ฐ์ดํ„ฐ, ํŠน์ • ์นด๋ฉ”๋ผ์—์„œ ์ดฌ์˜ํ•œ ๋ฐ์ดํ„ฐ ๋“ฑ์— ํƒœ๊ทธ๋ฅผ ์ ์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฐ ๋‹ค์Œ ์ด๋Ÿฌํ•œ ํƒœ๊ทธ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋ฐ์ดํ„ฐ์—์„œ ํƒœ๊ทธ์™€ ์ผ์น˜ํ•˜๋Š” ์ด๋ฏธ์ง€๋ฅผ ๊ฒ€์ƒ‰ํ•˜๊ณ  ํŠน์ • ํƒœ๊ทธ ๋˜๋Š” ํƒœ๊ทธ ์ง‘ํ•ฉ์ด ํฌํ•จ๋œ ์ด๋ฏธ์ง€๋กœ ๋ฐ์ดํ„ฐ ์„ธํŠธ์˜ ๋ฒ„์ „์„ ์ƒ์„ฑํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์—์„œ ์ด๋ฏธ์ง€์— ํƒœ๊ทธ ์ถ”๊ฐ€ํ•˜๊ธฐ Roboflow

Models hosted on Roboflow can be used with Label Assist, an automated annotation tool that uses your YOLO11 model to recommend annotations. To use Label Assist, first upload a YOLO11 model to Roboflow (see instructions later in the guide). Then, click the magic wand icon in the left sidebar and select your model for use in Label Assist.

๋ชจ๋ธ์„ ์„ ํƒํ•œ ๋‹ค์Œ '๊ณ„์†'์„ ํด๋ฆญํ•˜์—ฌ ๋ผ๋ฒจ ์ง€์›์„ ํ™œ์„ฑํ™”ํ•ฉ๋‹ˆ๋‹ค:

๋ผ๋ฒจ ์ง€์› ํ™œ์„ฑํ™”

์ฃผ์„์„ ๋‹ฌ ์ƒˆ ์ด๋ฏธ์ง€๋ฅผ ์—ด๋ฉด ๋ผ๋ฒจ ์–ด์‹œ์ŠคํŠธ๊ฐ€ ์ฃผ์„์„ ํŠธ๋ฆฌ๊ฑฐํ•˜๊ณ  ์ถ”์ฒœํ•ฉ๋‹ˆ๋‹ค.

์–ด๋…ธํ…Œ์ด์…˜์„ ์ถ”์ฒœํ•˜๋Š” ALabel Assist

Dataset Management for YOLO11

Roboflow ๋Š” ์ปดํ“จํ„ฐ ๋น„์ „ ๋ฐ์ดํ„ฐ ์„ธํŠธ๋ฅผ ์ดํ•ดํ•˜๊ธฐ ์œ„ํ•œ ๋„๊ตฌ ๋ชจ์Œ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

๋จผ์ € ๋ฐ์ดํ„ฐ ์„ธํŠธ ๊ฒ€์ƒ‰์„ ์‚ฌ์šฉํ•˜์—ฌ ์‹œ๋งจํ‹ฑ ํ…์ŠคํŠธ ์„ค๋ช…์„ ์ถฉ์กฑํ•˜๋Š” ์ด๋ฏธ์ง€(์ฆ‰, ์‚ฌ๋žŒ์ด ํฌํ•จ๋œ ๋ชจ๋“  ์ด๋ฏธ์ง€ ์ฐพ๊ธฐ)๋‚˜ ์ง€์ •๋œ ๋ผ๋ฒจ์„ ์ถฉ์กฑํ•˜๋Š” ์ด๋ฏธ์ง€(์ฆ‰, ์ด๋ฏธ์ง€๊ฐ€ ํŠน์ • ํƒœ๊ทธ์™€ ์—ฐ๊ด€๋˜์–ด ์žˆ์Œ)๋ฅผ ์ฐพ์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฐ์ดํ„ฐ ์„ธํŠธ ๊ฒ€์ƒ‰์„ ์‚ฌ์šฉํ•˜๋ ค๋ฉด ์‚ฌ์ด๋“œ๋ฐ”์—์„œ '๋ฐ์ดํ„ฐ ์„ธํŠธ'๋ฅผ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฐ ๋‹ค์Œ ํŽ˜์ด์ง€ ์ƒ๋‹จ์˜ ๊ฒ€์ƒ‰์ฐฝ๊ณผ ๊ด€๋ จ ํ•„ํ„ฐ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๊ฒ€์ƒ‰์–ด๋ฅผ ์ž…๋ ฅํ•ฉ๋‹ˆ๋‹ค.

์˜ˆ๋ฅผ ๋“ค์–ด, ๋‹ค์Œ ํ…์ŠคํŠธ ์ฟผ๋ฆฌ๋Š” ๋ฐ์ดํ„ฐ ์„ธํŠธ์— ์‚ฌ๋žŒ์ด ํฌํ•จ๋œ ์ด๋ฏธ์ง€๋ฅผ ์ฐพ์Šต๋‹ˆ๋‹ค:

์ด๋ฏธ์ง€ ๊ฒ€์ƒ‰

'ํƒœ๊ทธ' ์„ ํƒ๊ธฐ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ํŠน์ • ํƒœ๊ทธ๊ฐ€ ์žˆ๋Š” ์ด๋ฏธ์ง€๋กœ ๊ฒ€์ƒ‰ ๋ฒ”์œ„๋ฅผ ์ขํž ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค:

ํƒœ๊ทธ๋ณ„๋กœ ์ด๋ฏธ์ง€ ํ•„ํ„ฐ๋ง

๋ฐ์ดํ„ฐ ์„ธํŠธ๋กœ ๋ชจ๋ธ ํ•™์Šต์„ ์‹œ์ž‘ํ•˜๊ธฐ ์ „์— ๋น„์ „ ๋ชจ๋ธ์„ ํ•™์Šตํ•˜๊ธฐ ์ „์— ๋ฐ์ดํ„ฐ ์„ธํŠธ์— ๋Œ€ํ•œ ์ธ์‚ฌ์ดํŠธ์™€ ๋ฐ์ดํ„ฐ ์„ธํŠธ๋ฅผ ๊ฐœ์„ ํ•  ์ˆ˜ ์žˆ๋Š” ๋ฐฉ๋ฒ•์„ ์ œ๊ณตํ•˜๋Š” ์›น ๋„๊ตฌ์ธ Roboflow ์ƒํƒœ ํ™•์ธ์„ ์‚ฌ์šฉํ•˜๋Š” ๊ฒƒ์ด ์ข‹์Šต๋‹ˆ๋‹ค.

์ƒํƒœ ํ™•์ธ์„ ์‚ฌ์šฉํ•˜๋ ค๋ฉด ์‚ฌ์ด๋“œ๋ฐ”์˜ '์ƒํƒœ ํ™•์ธ' ๋งํฌ๋ฅผ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์— ์žˆ๋Š” ์ด๋ฏธ์ง€์˜ ํ‰๊ท  ํฌ๊ธฐ, ํด๋ž˜์Šค ์ž”์•ก, ์ด๋ฏธ์ง€์— ์ฃผ์„์ด ์žˆ๋Š” ์œ„์น˜์˜ ํžˆํŠธ๋งต ๋“ฑ์„ ๋ณด์—ฌ์ฃผ๋Š” ํ†ต๊ณ„ ๋ชฉ๋ก์ด ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค.

Roboflow ์ƒํƒœ ํ™•์ธ ๋ถ„์„

์ƒํƒœ ์ ๊ฒ€์€ ๋ฐ์ดํ„ฐ ์„ธํŠธ ์„ฑ๋Šฅ ํ–ฅ์ƒ์— ๋„์›€์ด ๋˜๋Š” ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ๊ถŒ์žฅํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ํด๋ž˜์Šค ๊ท ํ˜• ๊ธฐ๋Šฅ์€ ๋ ˆ์ด๋ธ”์— ๋ถˆ๊ท ํ˜•์ด ์žˆ์Œ์„ ๋ณด์—ฌ์ค„ ์ˆ˜ ์žˆ์œผ๋ฉฐ, ์ด๋ฅผ ํ•ด๊ฒฐํ•˜๋ฉด ์„ฑ๋Šฅ์ด๋‚˜ ๋ชจ๋ธ์„ ํ–ฅ์ƒ์‹œํ‚ฌ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋ชจ๋ธ ํ•™์Šต์„ ์œ„ํ•ด 40๊ฐœ ์ด์ƒ์˜ ํ˜•์‹์œผ๋กœ ๋ฐ์ดํ„ฐ ๋‚ด๋ณด๋‚ด๊ธฐ

๋ฐ์ดํ„ฐ๋ฅผ ๋‚ด๋ณด๋‚ด๋ ค๋ฉด ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ ๋ฒ„์ „์ด ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. ๋ฒ„์ „์€ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์˜ ์‹œ๊ฐ„์ด ๊ณ ์ •๋œ ์ƒํƒœ์ž…๋‹ˆ๋‹ค. ๋ฒ„์ „์„ ๋งŒ๋“ค๋ ค๋ฉด ๋จผ์ € ์‚ฌ์ด๋“œ๋ฐ”์—์„œ '๋ฒ„์ „'์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฐ ๋‹ค์Œ "์ƒˆ ๋ฒ„์ „ ๋งŒ๋“ค๊ธฐ" ๋ฒ„ํŠผ์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ์ด ํŽ˜์ด์ง€์—์„œ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์— ์ ์šฉํ•  ์ฆ๊ฐ• ๋ฐ ์ „์ฒ˜๋ฆฌ ๋‹จ๊ณ„๋ฅผ ์„ ํƒํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค:

๋‹ค์Œ์—์„œ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ ๋ฒ„์ „ ๋งŒ๋“ค๊ธฐ Roboflow

์„ ํƒํ•œ ๊ฐ ํ™•๋Œ€ ๊ธฐ๋Šฅ์— ๋Œ€ํ•ด ํ•„์š”์— ๋”ฐ๋ผ ํ™•๋Œ€ ๊ธฐ๋Šฅ์„ ์กฐ์ •ํ•  ์ˆ˜ ์žˆ๋Š” ํŒ์—…์ด ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค. ๋‹ค์Œ์€ ์ง€์ •๋œ ๋งค๊ฐœ๋ณ€์ˆ˜ ๋‚ด์—์„œ ๋ฐ๊ธฐ ์ฆ๊ฐ•์„ ์กฐ์ •ํ•˜๋Š” ์˜ˆ์‹œ์ž…๋‹ˆ๋‹ค:

๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์— ์ฆ๊ฐ• ๊ธฐ๋Šฅ ์ ์šฉํ•˜๊ธฐ

๋ฐ์ดํ„ฐ ์„ธํŠธ ๋ฒ„์ „์ด ์ƒ์„ฑ๋˜๋ฉด ๋ฐ์ดํ„ฐ๋ฅผ ๋‹ค์–‘ํ•œ ํ˜•์‹์œผ๋กœ ๋‚ด๋ณด๋‚ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฐ์ดํ„ฐ ์„ธํŠธ ๋ฒ„์ „ ํŽ˜์ด์ง€์—์„œ '๋ฐ์ดํ„ฐ ์„ธํŠธ ๋‚ด๋ณด๋‚ด๊ธฐ' ๋ฒ„ํŠผ์„ ํด๋ฆญํ•˜์—ฌ ๋ฐ์ดํ„ฐ๋ฅผ ๋‚ด๋ณด๋‚ด์„ธ์š”:

๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ ๋‚ด๋ณด๋‚ด๊ธฐ

You are now ready to train YOLO11 on a custom dataset. Follow this written guide and YouTube video for step-by-step instructions or refer to the Ultralytics documentation.

Upload Custom YOLO11 Model Weights for Testing and Deployment

Roboflow ๋Š” ๋ฐฐํฌ๋œ ๋ชจ๋ธ์— ๋Œ€ํ•ด ๋ฌดํ•œํžˆ ํ™•์žฅ ๊ฐ€๋Šฅํ•œ API๋ฅผ ์ œ๊ณตํ•˜๋ฉฐ NVIDIA Jetsons, Luxonis OAK, Raspberry Pis, GPU-๊ธฐ๋ฐ˜ ๋””๋ฐ”์ด์Šค ๋“ฑ์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋Š” SDK๋ฅผ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

You can deploy YOLO11 models by uploading YOLO11 weights to Roboflow. You can do this in a few lines of Python code. Create a new Python file and add the following code:

import roboflow  # install with 'pip install roboflow'

roboflow.login()

rf = roboflow.Roboflow()

project = rf.workspace(WORKSPACE_ID).project("football-players-detection-3zvbc")
dataset = project.version(VERSION).download("yolov8")

project.version(dataset.version).deploy(model_type="yolov8", model_path=f"{HOME}/runs/detect/train/")

์ด ์ฝ”๋“œ์—์„œ ํ”„๋กœ์ ํŠธ ID์™€ ๋ฒ„์ „ ID๋ฅผ ๊ณ„์ • ๋ฐ ํ”„๋กœ์ ํŠธ์˜ ๊ฐ’์œผ๋กœ ๋ฐ”๊ฟ‰๋‹ˆ๋‹ค. Roboflow API ํ‚ค๋ฅผ ๊ฒ€์ƒ‰ํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์•Œ์•„๋ณด์„ธ์š”.

์œ„์˜ ์ฝ”๋“œ๋ฅผ ์‹คํ–‰ํ•˜๋ฉด ์ธ์ฆ์„ ์š”์ฒญํ•˜๋Š” ๋ฉ”์‹œ์ง€๊ฐ€ ํ‘œ์‹œ๋ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฐ ๋‹ค์Œ ๋ชจ๋ธ์ด ์—…๋กœ๋“œ๋˜๊ณ  ํ”„๋กœ์ ํŠธ์— ๋Œ€ํ•œ API๊ฐ€ ์ƒ์„ฑ๋ฉ๋‹ˆ๋‹ค. ์ด ๊ณผ์ •์€ ์™„๋ฃŒํ•˜๋Š” ๋ฐ ์ตœ๋Œ€ 30๋ถ„์ด ๊ฑธ๋ฆด ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋ชจ๋ธ์„ ํ…Œ์ŠคํŠธํ•˜๊ณ  ์ง€์›๋˜๋Š” SDK์— ๋Œ€ํ•œ ๋ฐฐํฌ ์ง€์นจ์„ ์ฐพ์œผ๋ ค๋ฉด Roboflow ์‚ฌ์ด๋“œ๋ฐ”์˜ '๋ฐฐํฌ' ํƒญ์œผ๋กœ ์ด๋™ํ•˜์„ธ์š”. ์ด ํŽ˜์ด์ง€ ์ƒ๋‹จ์— ๋ชจ๋ธ์„ ํ…Œ์ŠคํŠธํ•  ์ˆ˜ ์žˆ๋Š” ์œ„์ ฏ์ด ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค. ์›น์บ ์„ ์‚ฌ์šฉํ•˜์—ฌ ์‹ค์‹œ๊ฐ„ ํ…Œ์ŠคํŠธ๋ฅผ ํ•˜๊ฑฐ๋‚˜ ์ด๋ฏธ์ง€ ๋˜๋Š” ๋™์˜์ƒ์„ ์—…๋กœ๋“œํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์˜ˆ์ œ ์ด๋ฏธ์ง€์—์„œ ์ถ”๋ก  ์‹คํ–‰

์—…๋กœ๋“œํ•œ ๋ชจ๋ธ์„ ๋ผ๋ฒจ๋ง ๋„์šฐ๋ฏธ๋กœ ์‚ฌ์šฉํ•  ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ๊ธฐ๋Šฅ์€ ํ•™์Šต๋œ ๋ชจ๋ธ์„ ์‚ฌ์šฉํ•˜์—ฌ Roboflow ์— ์—…๋กœ๋“œ๋œ ์ด๋ฏธ์ง€์— ์ฃผ์„์„ ์ถ”์ฒœํ•ฉ๋‹ˆ๋‹ค.

How to Evaluate YOLO11 Models

Roboflow ๋Š” ๋ชจ๋ธ ํ‰๊ฐ€์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ๋‹ค์–‘ํ•œ ๊ธฐ๋Šฅ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

Once you have uploaded a model to Roboflow, you can access our model evaluation tool, which provides a confusion matrix showing the performance of your model as well as an interactive vector analysis plot. These features can help you find opportunities to improve your model.

ํ˜ผ๋™ ๋งคํŠธ๋ฆญ์Šค์— ์•ก์„ธ์Šคํ•˜๋ ค๋ฉด Roboflow ๋Œ€์‹œ๋ณด๋“œ์˜ ๋ชจ๋ธ ํŽ˜์ด์ง€๋กœ ์ด๋™ํ•œ ๋‹ค์Œ '์ƒ์„ธ ํ‰๊ฐ€ ๋ณด๊ธฐ'๋ฅผ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค:

Roboflow ๋ชจ๋ธ ํ‰๊ฐ€ ์‹œ์ž‘ํ•˜๊ธฐ

ํ˜ผ๋™ ๋งคํŠธ๋ฆญ์Šค๊ฐ€ ํ‘œ์‹œ๋œ ํŒ์—…์ด ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค:

ํ˜ผ๋™ ๋งคํŠธ๋ฆญ์Šค

ํ˜ผ๋™ ๋งคํŠธ๋ฆญ์Šค์˜ ์ƒ์ž ์œ„๋กœ ๋งˆ์šฐ์Šค๋ฅผ ๊ฐ€์ ธ๊ฐ€๋ฉด ํ•ด๋‹น ์ƒ์ž์™€ ๊ด€๋ จ๋œ ๊ฐ’์„ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ƒ์ž๋ฅผ ํด๋ฆญํ•˜๋ฉด ํ•ด๋‹น ์นดํ…Œ๊ณ ๋ฆฌ์˜ ์ด๋ฏธ์ง€๋ฅผ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋ฏธ์ง€๋ฅผ ํด๋ฆญํ•˜๋ฉด ํ•ด๋‹น ์ด๋ฏธ์ง€์™€ ์—ฐ๊ด€๋œ ๋ชจ๋ธ ์˜ˆ์ธก ๋ฐ ์‹ค์ธก ๋ฐ์ดํ„ฐ๋ฅผ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋” ๋งŽ์€ ์ธ์‚ฌ์ดํŠธ๋ฅผ ์–ป์œผ๋ ค๋ฉด ๋ฒกํ„ฐ ๋ถ„์„์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋ฉด CLIP์„ ์‚ฌ์šฉํ•˜์—ฌ ๊ณ„์‚ฐ๋œ ๋ฐ์ดํ„ฐ ์„ธํŠธ์˜ ์ด๋ฏธ์ง€ ๋ถ„์‚ฐํ˜• ์ฐจํŠธ๊ฐ€ ํ‘œ์‹œ๋ฉ๋‹ˆ๋‹ค. ํ”Œ๋กฏ์—์„œ ์ด๋ฏธ์ง€๊ฐ€ ๊ฐ€๊นŒ์šธ์ˆ˜๋ก ์˜๋ฏธ์ ์œผ๋กœ ๋” ์œ ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๊ฐ ์ด๋ฏธ์ง€๋Š” ํฐ์ƒ‰๊ณผ ๋นจ๊ฐ„์ƒ‰ ์‚ฌ์ด์˜ ์ƒ‰์ƒ์„ ๊ฐ€์ง„ ์ ์œผ๋กœ ํ‘œ์‹œ๋ฉ๋‹ˆ๋‹ค. ์ ์ด ๋นจ๊ฐ„์ƒ‰์ผ์ˆ˜๋ก ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์ด ๋–จ์–ด์ง‘๋‹ˆ๋‹ค.

๋ฒกํ„ฐ ๋ถ„์„ ํ”Œ๋กฏ

๋ฒกํ„ฐ ๋ถ„์„์„ ์‚ฌ์šฉํ•˜์—ฌ ๋‹ค์Œ์„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค:

  • ์ด๋ฏธ์ง€ ํด๋Ÿฌ์Šคํ„ฐ๋ฅผ ์ฐพ์Šต๋‹ˆ๋‹ค;
  • ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์ด ์ €์กฐํ•œ ํด๋Ÿฌ์Šคํ„ฐ๋ฅผ ์‹๋ณ„ํ•˜๊ณ ;
  • ๋ชจ๋ธ์ด ์ œ๋Œ€๋กœ ์ž‘๋™ํ•˜์ง€ ์•Š๋Š” ์ด๋ฏธ์ง€ ๊ฐ„์˜ ๊ณตํ†ต์ ์„ ์‹œ๊ฐํ™”ํ•ฉ๋‹ˆ๋‹ค.

ํ•™์Šต ๋ฆฌ์†Œ์Šค

Want to learn more about using Roboflow for creating YOLO11 models? The following resources may be helpful in your work.

  • Train YOLO11 on a Custom Dataset: Follow our interactive notebook that shows you how to train a YOLO11 model on a custom dataset.
  • Autodistill: Use large foundation vision models to label data for specific models. You can label images for use in training YOLO11 classification, detection, and segmentation models with Autodistill.
  • ๊ฐ๋…: ์ปดํ“จํ„ฐ ๋น„์ „ ๋ชจ๋ธ ์ž‘์—…์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ์œ ์šฉํ•œ ์œ ํ‹ธ๋ฆฌํ‹ฐ๊ฐ€ ํฌํ•จ๋œ Python ํŒจํ‚ค์ง€์ž…๋‹ˆ๋‹ค. ์ˆ˜ํผ๋น„์ „์„ ์‚ฌ์šฉํ•˜์—ฌ ํƒ์ง€ ํ•„ํ„ฐ๋ง, ํ˜ผ๋™ ํ–‰๋ ฌ ๊ณ„์‚ฐ ๋“ฑ์„ Python ์ฝ”๋“œ ๋ช‡ ์ค„๋กœ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • Roboflow Blog: The Roboflow Blog features over 500 articles on computer vision, covering topics from how to train a YOLO11 model to annotation best practices.
  • Roboflow YouTube channel: Browse dozens of in-depth computer vision guides on our YouTube channel, covering topics from training YOLO11 models to automated image labeling.

ํ”„๋กœ์ ํŠธ ์‡ผ์ผ€์ด์Šค

Below are a few of the many pieces of feedback we have received for using YOLO11 and Roboflow together to create computer vision models.

์‡ผ์ผ€์ด์Šค ์ด๋ฏธ์ง€ ์‡ผ์ผ€์ด์Šค ์ด๋ฏธ์ง€ ์‡ผ์ผ€์ด์Šค ์ด๋ฏธ์ง€

์ž์ฃผ ๋ฌป๋Š” ์งˆ๋ฌธ

How do I label data for YOLO11 models using Roboflow?

Labeling data for YOLO11 models using Roboflow is straightforward with Roboflow Annotate. First, create a project on Roboflow and upload your images. After uploading, select the batch of images and click "Start Annotating." You can use the B ํ‚ค๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ์ƒ์ž์˜ ๊ฒฝ๊ณ„๋ฅผ ์ง€์ •ํ•˜๊ฑฐ๋‚˜ P ํ‚ค๋ฅผ ๋ˆ„๋ฅด์„ธ์š”. ๋” ๋น ๋ฅด๊ฒŒ ์ฃผ์„์„ ๋‹ฌ๋ ค๋ฉด ์‚ฌ์ด๋“œ๋ฐ”์—์„œ ์ปค์„œ ์•„์ด์ฝ˜์„ ํด๋ฆญํ•˜์—ฌ SAM-๊ธฐ๋ฐ˜ ๋ผ๋ฒจ ๋„์šฐ๋ฏธ๋ฅผ ์‚ฌ์šฉํ•˜์„ธ์š”. ์ž์„ธํ•œ ๋‹จ๊ณ„๋Š” ๋‹ค์Œ์—์„œ ํ™•์ธํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์—ฌ๊ธฐ.

What services does Roboflow offer for collecting YOLO11 training data?

Roboflow provides two key services for collecting YOLO11 training data: Universe and Collect. Universe offers access to over 250,000 vision datasets, while Collect helps you gather images using a webcam and automated prompts.

How can I manage and analyze my YOLO11 dataset using Roboflow?

Roboflow offers robust dataset management tools, including dataset search, tagging, and Health Check. Use the search feature to find images based on text descriptions or tags. Health Check provides insights into dataset quality, showing class balance, image sizes, and annotation heatmaps. This helps optimize dataset performance before training YOLO11 models. Detailed information can be found here.

How do I export my YOLO11 dataset from Roboflow?

To export your YOLO11 dataset from Roboflow, you need to create a dataset version. Click "Versions" in the sidebar, then "Create New Version" and apply any desired augmentations. Once the version is generated, click "Export Dataset" and choose the YOLO11 format. Follow this process here.

How can I integrate and deploy YOLO11 models with Roboflow?

Integrate and deploy YOLO11 models on Roboflow by uploading your YOLO11 weights through a few lines of Python code. Use the provided script to authenticate and upload your model, which will create an API for deployment. For details on the script and further instructions, see this section.

What tools does Roboflow provide for evaluating YOLO11 models?

Roboflow ์—์„œ๋Š” ํ˜ผ๋™ ํ–‰๋ ฌ ๋ฐ ๋ฒกํ„ฐ ๋ถ„์„ ํ”Œ๋กฏ์„ ํฌํ•จํ•œ ๋ชจ๋ธ ํ‰๊ฐ€ ๋„๊ตฌ๋ฅผ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค. ๋ชจ๋ธ ํŽ˜์ด์ง€์˜ '์ƒ์„ธ ํ‰๊ฐ€ ๋ณด๊ธฐ' ๋ฒ„ํŠผ์—์„œ ์ด๋Ÿฌํ•œ ๋„๊ตฌ์— ์•ก์„ธ์Šคํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ ๊ธฐ๋Šฅ์€ ๋ชจ๋ธ ์„ฑ๋Šฅ ๋ฌธ์ œ๋ฅผ ํŒŒ์•…ํ•˜๊ณ  ๊ฐœ์„ ํ•  ์˜์—ญ์„ ์ฐพ๋Š” ๋ฐ ๋„์›€์ด ๋ฉ๋‹ˆ๋‹ค. ์ž์„ธํ•œ ๋‚ด์šฉ์€ ์ด ์„น์…˜์„ ์ฐธ์กฐํ•˜์„ธ์š”.


๐Ÿ“… Created 11 months ago โœ๏ธ Updated 12 days ago

๋Œ“๊ธ€