Is possible multi-class tracking ?

[ifzhang]

Is possible multi-class tracking ?

[deep-practice]

ByteTrack only use iou matching,you can simply replace the person detector with a multi object detector.

[betterhalfwzm]

ByteTrack only use iou matching,you can simply replace the person detector with a multi object detector.

那我直接用yolox多分类检测器替换这个人体的单分类是不是就可以了，谢谢了

[deep-practice]

Yes

[betterhalfwzm]

Yes
未检测到目标会挂掉 @deep-practice
Traceback (most recent call last):
File "tools/demo_track.py", line 350, in
main(exp, args)
File "tools/demo_track.py", line 343, in main
imageflow_demo(predictor, vis_folder, current_time, args)
File "tools/demo_track.py", line 250, in imageflow_demo
online_targets = tracker.update(outputs[0], [img_info['height'], img_info['width']], exp.test_size)
File "/mnt/e/ubuntu1804/workspace/ByteTrack/yolox/tracker/byte_tracker.py", line 166, in update
if output_results.shape[1] == 5:
AttributeError: 'NoneType' object has no attribute 'shape'

[deep-practice]

你加个判断就好了吧，如果output_results为空，dets和dets_second也置为空

[CaptainEven]

@betterhalfwzm @deep-practice Hey! You can refer to my repo:
https://github.com/CaptainEven/MCMOT-ByteTrack