Are the NVIDIA Trackers Adaptive?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Howdy y’all,

I’ve got a questions about the various trackers that NVIDIA has implemented (klt, iou, and DCF). Do the tracker algorithms output adaptive bounding boxes automatically or does it rely on the inference to send in a ObjToTrack which then the NVIDIA tracker NvMOT_Process() will take in and correlate to a existing tracked object and resize to the new inferred bounding box.

Thanks

2

It rely on the inference to send in a ObjToTrack