Adding Object Tracking and Re-Identification Using YOLOV9 and Strong SORT and OSNET

[TheNobody-12]

Description

This Release contains a highly configurable two-stage tracker that adjusts to different deployment scenarios. The detections generated by YOLOv9, a family of object detection architectures and models pre-trained on the COCO dataset, are passed to StrongSORT which combines motion and appearance information based on OSNet to track the objects. It can track any object that your Yolov9 model was trained to detect.

Example of Tracking Algorithm

Key features:

🎯 Accurate object detection powered by YOLOv9 🔄 Seamless object tracking using StrongSORT and OSNet 🔧 Highly configurable to adapt to different deployment scenarios

Repository Link

The YOLOv9_StrongSORT_OSNET

[TheNobody-12]

@WongKinYiu Do check this out

[JFMeyer2k]

Do you have comparisons of performance to other trackers like (older) trackers like ByteTrack or DeepSORT?

[TheNobody-12]

Yes, the STRONG sort is the advancement to the Deep SORT framework. https://github.com/dyhBUPT/StrongSORT here is the original code repository of the StrongSORT which performs better than ByteTrack and DeepSort.

Results:

[WongKinYiu]

Added to readme.