Overview
CVAT (Computer Vision Annotation Tool) is a professional platform for annotating visual data, supporting tasks such as object detection, semantic segmentation, and keypoint detection. It provides end-to-end workflows from task creation, assignment, annotation to review and export, suitable for research and industry-scale data preparation.
Key Features
- Rich annotation tools: bounding boxes, segmentation, polygons, and keypoints.
- Collaboration: task assignment, review, and quality control.
- Scalable deployment: Docker and Kubernetes support for large-scale annotation.
Use Cases
- Building datasets for object detection, segmentation, and keypoint tasks.
- Data quality assurance via review and consistency checks.
- Large-scale annotation projects requiring multi-team coordination.
Technical Details
- Stack: Python backend with modern frontend technologies, supporting various storage backends.
- Extensibility: modular architecture and plugin support for different scales.
- License: MIT.