Skip to content

This repository is intended to focus on analyzing the instances movements in a scenario using multiple object tracking with ByteTrack and YOLOX approach.

License

Notifications You must be signed in to change notification settings

MB-MuratBayraktar/ByteTrack-MOT-with-UI-YOLOX-

Repository files navigation

ByteTrack-MOT-with-UI-YOLOX-

This repository is intended to focus on analyzing the instances movements in a scenario using multiple object tracking with ByteTrack and YOLOX approach.

Below you can see the ByteTrack algorithm in action with YOLOX as the detection backbone.

You can see how it performs well in tracking instances with uniqe IDs even when occulsion is happening. ezgif com-gif-maker

ezgif com-gif-maker (1)

The model scored 80.3 on MOTA , 77.3 on IDF1 at MOTA17 dataset and performs at 30 FPS!

I am adding some analytical features with UI using streamlit.

The original repo is found here

About

This repository is intended to focus on analyzing the instances movements in a scenario using multiple object tracking with ByteTrack and YOLOX approach.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages