xinshuoweng/AB3DMOT

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1845 stars · 415 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official Python implementation of AB3DMOT, a simple real-time baseline for 3D multi-object tracking using oriented 3D bounding boxes from LiDAR point clouds, with new evaluation metrics (e.g., sAMOTA). Supports KITTI and nuScenes datasets.

Use cases

• track 3d objects from lidar point clouds
• evaluate 3d multi-object tracking on kitti
• baseline 3d mot system for autonomous driving
• compute samota and mot evaluation metrics
• run multi-object tracking on nuscenes
• compare 3d tracking algorithms against a baseline

When to choose

• you need a simple, real-time 3D MOT baseline for research comparison
• you want standardized 3D/2D MOT evaluation metrics on KITTI or nuScenes
• you work on autonomous driving perception with LiDAR detections

When to avoid

• you need a production-ready tracking system with learned association models
• you track from camera images only without 3D detections
• you need active development or commercial support

Facets

library · maturity maintenance

computer-vision machine-learning benchmarking simulation computer-vision autonomous-vehicles robotics machine-learning python cross-platform 3d-multi-object-tracking kalman-filter lidar kitti nuscenes evaluation-metrics research-code point-cloud linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="xinshuoweng/AB3DMOT")