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

Bad result on YOLOX + UniTrack demo #19

Closed kumi123 closed 2 years ago

kumi123 commented 2 years ago

Hi, Zhangdao thank you for your wonderful UniTrack framwork. but when i tried use YOLOX + UniTrack demo mot_demo.py on a video sequence,the result was not that good. i put the yolox_x(the better version) pre-traind model on yolox/weights and i used the config imagenet_resnet18_s3.yaml . it seems like the framwork could not detect the person object but works pretty well on other object . by the way i use the default classed arg list(range(80)) what should i do?

img

kentaroy47 commented 2 years ago

You may want to see #17.

The person label is excluded in the demo script, so you should change args.classes = [x for x in args.classes] in https://github.com/Zhongdao/UniTrack/blob/main/demo/mot_demo.py#L192

Zhongdao commented 2 years ago

Hi @kumi123，and thanks @kentaroy47！ Already solved (#20)

kumi123 commented 2 years ago

You may want to see #17.

The person label is excluded in the demo script, so you should change args.classes = [x for x in args.classes] in https://github.com/Zhongdao/UniTrack/blob/main/demo/mot_demo.py#L192

appreciat that! man that works very well! I always thought that the person class index was 1, but it turned out to be 0

kumi123 commented 2 years ago

Hi @kumi123，and thanks @kentaroy47！ Already solved (#20)

thanks a lot

liuyang77712 commented 2 years ago

hi @kumi123 , I have a problem. I also had a bad result like this pic. 00162 I only donwload the yolox_x.pth and run the python demo. what should i do?