BoT-SORT

Robust multi-object tracking combining motion and appearance cues.

Open SourceSelf HostedOffline CapableGPU Required (4GB+ VRAM)
0.0 (0)

About

BoT-SORT is a multi-object tracking method that associates detections across video frames by combining motion and appearance cues. It adds camera-motion compensation, an improved Kalman filter state vector, and an optional re-identification branch (BoT-SORT-ReID) for appearance matching. It reached state-of-the-art results on the MOT17 and MOT20 benchmarks and is commonly paired with YOLO detectors for tracking pipelines.

Reviews (0)

Leave a Review

No reviews yet. Be the first to review!

Details

Price
Free
Platform
Local/Desktop
Difficulty
Intermediate (3/5)
Minimum VRAM
4 GB
Added
Apr 3, 2026

Related Tools

Monocular depth estimation model producing detailed depth maps from single images.

Open SourceSelf HostedOfflineGPU 4GB+
Easy
0.0 (0)

End-to-end object detection with transformers by Meta, eliminating hand-designed components.

Open SourceSelf HostedOfflineGPU 8GB+
Advanced
0.0 (0)

Self-supervised vision transformer by Meta producing universal visual features.

Open SourceSelf HostedOfflineGPU 6GB+
Intermediate
0.0 (0)

Unified vision foundation model by Microsoft for captioning, detection, and segmentation.

Open SourceSelf HostedOfflineGPU 6GB+
Intermediate
0.0 (0)

Monocular depth estimation model by Intel ISL supporting multiple backbones.

Open SourceSelf HostedOffline
Easy
0.0 (0)

Simple and effective multi-object tracking using every detection box.

Open SourceSelf HostedOfflineGPU 4GB+
Intermediate
0.0 (0)
Browse all Computer Vision & Object Detection tools