BBox from the detection model #34
Description
I wanted to know why not return the Bbox with the key points and score in call from the Wholebody or PoseTracker classes? Wouldn't this be more accurate than calculating them from the keypoints?
{{ message }}