Computer Vision Annotation Tool (CVAT)

CVAT is free, online, interactive video and image annotation tool for computer vision. It is being used by our team to annotate million of objects with different properties. Many UI and UX decisions are based on feedbacks from professional data annotation team. Try it online cvat.org.

Documentation

• Installation guide
• User's guide
• Django REST API documentation
• Datumaro dataset framework
• Command line interface
• XML annotation format
• AWS Deployment Guide
• Frequently asked questions
• Questions

Screencasts

• Introduction
• Annotation mode
• Interpolation of bounding boxes
• Interpolation of polygons
• Tag annotation video
• Attribute mode
• Segmentation mode
• Tutorial for polygons
• Semi-automatic segmentation

Supported annotation formats

Format selection is possible after clicking on the Upload annotation and Dump annotation buttons. Datumaro dataset framework allows additional dataset transformations via its command line tool and Python library.

For more information about supported formats look at the documentation.

Annotation format	Import	Export
CVAT for images	X	X
CVAT for a video	X	X
Datumaro		X
PASCAL VOC	X	X
Segmentation masks from PASCAL VOC	X	X
YOLO	X	X
MS COCO Object Detection	X	X
TFrecord	X	X
MOT	X	X
LabelMe 3.0	X	X
ImageNet	X	X
CamVid	X	X

Deep learning serverless functions for automatic labeling

Name	Type	Framework	CPU	GPU
Deep Extreme Cut	interactor	OpenVINO	X
Faster RCNN	detector	OpenVINO	X
Mask RCNN	detector	OpenVINO	X
YOLO v3	detector	OpenVINO	X
Object reidentification	reid	OpenVINO	X
Semantic segmentation for ADAS	detector	OpenVINO	X
Text detection v4	detector	OpenVINO	X
SiamMask	tracker	PyTorch	X
f-BRS	interactor	PyTorch	X
Inside-Outside Guidance	interactor	PyTorch	X
Faster RCNN	detector	TensorFlow	X	X
Mask RCNN	detector	TensorFlow	X	X

Online demo: cvat.org

This is an online demo with the latest version of the annotation tool. Try it online without local installation. Only own or assigned tasks are visible to users.

Disabled features:

• Analytics: management and monitoring of data annotation team

Limitations:

• No more than 10 tasks per user
• Uploaded data is limited to 500Mb

Prebuilt Docker images

Prebuilt docker images for CVAT releases are available on Docker Hub:

• cvat_server
• cvat_ui

REST API

Automatically generated Swagger documentation for Django REST API is available on <cvat_origin>/api/swagger(default: localhost:8080/api/swagger).

Swagger documentation is visiable on allowed hostes, Update environement variable in docker-compose.yml file with cvat hosted machine IP or domain name. Example - ALLOWED_HOSTS: 'localhost, 127.0.0.1'.

LICENSE

Code released under the MIT License.

This software uses LGPL licensed libraries from the FFmpeg project. The exact steps on how FFmpeg was configured and compiled can be found in the Dockerfile.

FFmpeg is an open source framework licensed under LGPL and GPL. See https://www.ffmpeg.org/legal.html. You are solely responsible for determining if your use of FFmpeg requires any additional licenses. Intel is not responsible for obtaining any such licenses, nor liable for any licensing fees due in connection with your use of FFmpeg.

Questions

CVAT usage related questions or unclear concepts can be posted in our Gitter chat for quick replies from contributors and other users.

However, if you have a feature request or a bug report that can reproduced, feel free to open an issue (with steps to reproduce the bug if it's a bug report) on GitHub* issues.

If you are not sure or just want to browse other users common questions, Gitter chat is the way to go.

Other ways to ask questions and get our support:

• #cvat tag on StackOverflow*
• Forum on Intel Developer Zone

Links

• Intel AI blog: New Computer Vision Tool Accelerates Annotation of Digital Images and Video
• Intel Software: Computer Vision Annotation Tool: A Universal Approach to Data Annotation
• VentureBeat: Intel open-sources CVAT, a toolkit for data labeling

Projects using CVAT

• Onepanel - Onepanel is an open source vision AI platform that fully integrates CVAT with scalable data processing and parallelized training pipelines.