search

Zhongdao / UniTrack

[NeurIPS'21] Unified tracking framework with a single appearance model. It supports Single Object Tracking (SOT), Video Object Segmentation (VOS), Multi-Object Tracking (MOT), Multi-Object Tracking and Segmentation (MOTS), Pose Tracking, Video Instance Segmentation (VIS), and class-agnostic MOT (e.g. TAO dataset).
MIT License
334 stars 35 forks source link

mot demo not producing valid tracking results #17

Closed kentaroy47 closed 2 years ago

kentaroy47 commented 2 years ago

Thanks for releasing this repo!

I tried to run mot on a sample video. The script ran fine and was able to produce the video, but it did not contain any tracking results. I checked up on yolox demo.py, but the detector was working fine.

Do I need to set the pretrained resnet18 to get the demo working?

kentaroy47 commented 2 years ago

I solved this issue with the demo.py, the person label is 0 in yolo-x but that label was excluded with the default settings.

Changing args.classes = [x for x in args.classes] solves the issue.