ifzhang/ByteTrack

Entire Detection of occluded/partially_visible object

aimaicai opened this issue · 0 comments

Hi, I'm using this amazing project and I notice that yolox is able to detect the entire object even if it is occluded or partially visible. It seems that the detection can go outside the image size estimating the entire person shape.
I never see this before. Is it a peculiarity of yolox or can I find it also in other detector?
Thanks