Track-Anything is a flexible and interactive tool for video object tracking and segmentation, based on Segment Anything, XMem, and E2FGVI.

Out-of-the-box code and models for CMU's object detection and tracking system for multi-camera surveillance videos. Speed optimized Faster-RCNN model. Tensorflow based. Also supports EfficientDet. WACVW'20

[NeurIPS'21] Unified tracking framework with a single appearance model. It supports Single Object Tracking (SOT), Video Object Segmentation (VOS), Multi-Object Tracking (MOT), Multi-Object Tracking and Segmentation (MOTS), Pose Tracking, Video Instance Segmentation (VIS), and class-agnostic MOT (e.g. TAO dataset).

[ECCV-20] Official PyTorch implementation of HoughNet, a voting-based object detector.

Implementation of Lucas Kanade Tracking system using six parameter affine model and recursive Gauss-Newton process.

ROS compatible package for object tracking based on SAM, Cutie, GroundingDINO, YOLO-World, VLPart and DEVA

TrackGPT: Track What You Need in Videos via Text Prompts

Characterize Anything: A Wondrous Chemical Reaction between vision models and AI Characters

[TPAMI-22] Bottom-up, voting based video object detection method

This project explores Lucas-Kanade algorithm to track an object in a video sequence