I used a three-level network, which is vehicle detection, tracking, vehicle color recognition and I set it to detect every frame.
As follows, I have two questions:
1，How do I decide if the target is gone？
2，After vehicle detection and vehicle color recognition for a certain target, will vehicle detection and vehicle color recognition be carried out for the same target in subsequent tracking?
Do you use a customized model?
If yes, have you checked these questions with the model developer?
You should get a better suggestion from the author directly.