boxmot 19.0.0

BoxMOT: pluggable SOTA tracking modules for segmentation, object detection and pose estimation models

BoxMOT gives you one CLI and one Python API for running modern multi-object tracking workflows. It covers direct tracking, cached benchmark evaluation, tuning, research loops, and ReID export without forcing you to rebuild the detector and tracker stack for each experiment.

Docs • Installation • Modes • API Reference • Trackers • Contributing

Why BoxMOT

• One interface for track, generate, eval, tune, research, and export.
• Swappable trackers with shared detector and ReID plumbing.
• Benchmark-oriented workflows with reusable detections and embeddings.
• Support for both AABB and OBB tracking paths.
• Optional production-ready native C++ tracker implementations with the same metrics as the Python path, opted into via --tracker-backend cpp and embeddable in standalone C++ projects via CMake (see Native C++ Integration).
• Public Python API for embedding the same workflows in applications and notebooks.

Installation

BoxMOT supports Python 3.9 through 3.12.

pip install boxmot
boxmot --help

For mode-specific extras such as yolo, evolve, research, onnx, openvino, and tflite, see the installation guide.

Benchmark Results (MOT17 ablation split)

Tracker	Python	C++
HOTA	MOTA	IDF1	OBB	HOTA	MOTA	IDF1	OBB
occluboost	70.47	78.32	84.14	✅	70.48	78.31	84.14	✅
botsort	69.44	78.24	81.94	✅	69.43	78.26	82.00	✅
boosttrack	69.25	75.91	83.20	❌	—	—	—	—
strongsort	68.05	76.19	80.76	❌	—	—	—	—
deepocsort	67.95	75.83	80.54	❌	—	—	—	—
bytetrack	67.68	78.04	79.16	✅	67.75	78.03	79.38	✅
hybridsort	67.31	74.09	78.87	❌	—	—	—	—
ocsort	66.44	74.55	77.90	✅	66.44	74.55	77.90	✅
sfsort	62.65	76.87	69.18	✅	62.66	76.74	69.18	✅

_{Evaluation was run on the second half of the MOT17 training set because the validation split is not public and the ablation detector was trained on the first half. Results used pre-generated detections and embeddings with each tracker configured from its default repository settings. Native cpp cells use --tracker-backend cpp for the implemented replay backends.}

Related guides:

• Evaluation and Postprocessing
• Benchmark Workflows
• Native C++ Integration

Minimal Usage

CLI:

boxmot track --detector yolov8n --reid osnet_x0_25_msmt17 --tracker botsort --source video.mp4 --save

Python:

from boxmot import Boxmot

run = Boxmot(detector="yolov8n", reid="osnet_x0_25_msmt17", tracker="botsort").track(
    source="video.mp4",
    save=True,
)
print(run)

Contributing

Start with CONTRIBUTING.md and the contributor docs.

Contributors

Support and Citation

• Bugs and feature requests: GitHub Issues
• Questions and discussion: GitHub Discussions or Discord
• Citation metadata: CITATION.cff
• Commercial support: [e-mail]