Hi sir, is there any way to know that, for example, a feet key-point is occluded by the body part or not? i.e. to get the "visibility" label from the 2d/3d kpts information? thanks.
Not directly: there's no label saying if a joint is occluded.
However, that being said, we provide semantic segmentation and that can be used to detect what you are asking for.
Hi sir, is there any way to know that, for example, a feet key-point is occluded by the body part or not? i.e. to get the "visibility" label from the 2d/3d kpts information? thanks.