MotionCtrl
MotionCtrl copied to clipboard
How does it address scale ambiguity issue?
From my understanding, CMCM digests over a set of poses represented by extrinsic matrix. But then how does it deal with scale ambiguity issue on SVD case?