search

ifzhang / ByteTrack

[ECCV 2022] ByteTrack: Multi-Object Tracking by Associating Every Detection Box
MIT License
4.63k stars 878 forks source link

Question: How to connect the input and output of ByteTrack? #71

Open shu1106 opened 2 years ago

shu1106 commented 2 years ago 
tracker = BYTETracker(args)
for image in images:
   dets = detector(image)
   online_targets = tracker.update(dets, info_imgs, img_size)

For the case where you consider the above code.

Is there a way to know which of the inputs (dets) to the tracker are being tracked and which are not? I want to tie the input (dets) to the output (online_targets) and get all the information about what was tracked and what was not.

Thank you very much for your answer.

iamrajee commented 2 years ago

@shu1106 I'm not sure about the tracked and untracked dets but this is how you can use it in your code. Thanks!

AydinAlptug commented 2 years ago

I also wonder how to understand the object is lost since there is not a field like "time_since_update" the one in the DeepSort. Any advice would be helpful..

iamrajee commented 2 years ago

In ByteTrack we have --track_buffer, which does the same thing. https://github.com/ifzhang/ByteTrack/blob/main/tools/demo_track.py#L81