Quantify vision system performance in human terms

Fork me on GitHub
  • Support various types of annotations on both images and videos
  • Build innovative features and user-friendly interface
  • Improve speed by using semi-automated annotations
  • Support concurrent annotation sessions and progress monitoring
  • Accessible through a web browser without installation
  • Supported Annotation Types

    A configurable system that can support various types of annotations and can be easily adapted to new tasks.

    Bounding Boxes

    Support simple "click and drag" actions and options to add multiple attributes.

    Semantic Segmentation

    Feature functions to fit boundaries with Bezier curves and copy shared boundaries.

    Video Tracking

    Predict annotations between frames using object tracking and interpolation algorithm.

    Drivable Area

    Annotate the area that the driver is currently driving on.

    Lane Marking

    Annotate lane marking for vision-based vehicle localization and trajectory planing.

    3D bouding boxes in Point Cloud

    Coming up soon.

    How to get started?

    Download our latest code from github

    Team Info

    Core Contributors

    Haofeng Chen
    Haoping Bai
    Sean Foley
    Yingying Chen

    Gary Chen
    Xuanyu Zhou
    Teer Ba

    Lead

    Advisors