search

ifzhang / ByteTrack

[ECCV 2022] ByteTrack: Multi-Object Tracking by Associating Every Detection Box
MIT License
4.79k stars 904 forks source link

New position estimation of object without detection #192

Open rose-jinyang opened 2 years ago

rose-jinyang commented 2 years ago

Hello How are you? Thanks for contributing to this project. I found that this method does NOT estimate new position of object without detection of object. In general, there are cases when there is NOT a detection result of an object at certain moment. I think that new position estimation of an object should be done in even such case.

LamnouarMohamed commented 2 years ago

Hi, yeah, I agree with you, I think just we need to add some modification like increment number of lost of this person of course if Stracks not empty, and remove all person that are there number of lost are higher than threshold.

rose-jinyang commented 2 years ago

Thanks for your reply. Excuse me, are u a developer of this project? If so, I have another question. This project shows a simple explanation in combining this tracker with other detector. But I think that the detailed detector specification is needed to combine with this tracker. That's because a general detector filters objects with confidences lower than a certain threshold in post-processing like NMS. I found that this tracker uses a high threshold and a low threshold for detection confidence. I think that the logic using confidences in a detector and this tracker should be consistent. I hope that this confidence problem will be discussed in the issue: https://github.com/ifzhang/ByteTrack/issues/193.

rose-jinyang commented 2 years ago

Hi @ifzhang I found that this method does NOT estimate new position of object without detection of object. In general, there are cases when there is NOT a detection result of an object at certain moment. I think that new position estimation of an object should be done in even such case.