trackers 2.2.0

A unified library for object tracking featuring clean room re-implementations of leading multi-object tracking algorithms

trackers

Plug-and-play multi-object tracking for any detection model.

Try It

No install needed. Try trackers in your browser with our Hugging Face Playground.

Install

pip install trackers

install from source

pip install git+https://github.com/roboflow/trackers.git

https://github.com/user-attachments/assets/eef9b00a-cfe4-40f7-a495-954550e3ef1f

Track from CLI

Point at a video, webcam, RTSP stream, or image directory. Get tracked output.

Use our interactive command builder to configure your tracking pipeline.

trackers track \
    --source video.mp4 \
    --output output.mp4 \
    --model rfdetr-medium \
    --tracker bytetrack \
    --show-labels \
    --show-trajectories

Track from Python

Plug trackers into your existing detection pipeline. Works with any detector.

import cv2
import supervision as sv
from inference import get_model
from trackers import ByteTrackTracker

model = get_model(model_id="rfdetr-medium")
tracker = ByteTrackTracker()

label_annotator = sv.LabelAnnotator()
trajectory_annotator = sv.TrajectoryAnnotator()

cap = cv2.VideoCapture("video.mp4")
while cap.isOpened():
    ret, frame = cap.read()
    if not ret:
        break

    result = model.infer(frame)[0]
    detections = sv.Detections.from_inference(result)
    tracked = tracker.update(detections)

    frame = label_annotator.annotate(frame, tracked)
    frame = trajectory_annotator.annotate(frame, tracked)

Evaluate

Benchmark your tracker against ground truth with standard MOT metrics.

trackers eval \
    --gt-dir data/gt \
    --tracker-dir data/trackers \
    --metrics CLEAR HOTA Identity

Sequence                        MOTA    HOTA    IDF1  IDSW
----------------------------------------------------------
MOT17-02-FRCNN                75.600  62.300  72.100    42
MOT17-04-FRCNN                78.200  65.100  74.800    31
----------------------------------------------------------
COMBINED                      75.033  62.400  72.033    73

Algorithms

Clean, modular implementations of leading trackers. See the tracker comparison for detailed benchmarks.

Algorithm	MOT17	SportsMOT	SoccerNet
SORT	58.4	70.9	81.6
ByteTrack	60.1	73.0	84.0
OC-SORT	—	—	—
BoT-SORT	—	—	—
McByte	—	—	—

Contributing

We welcome contributions. Read our contributor guidelines to get started.

License

The code is released under the Apache 2.0 license.