A fully differentiable end-to-end model for arbitrary-skeleton motion capture from video that jointly learns pose estimation and rotation recovery using reference pose-rotation pairs to eliminate IK ambiguities.
InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MoCapAnything V2: End-to-End Motion Capture for Arbitrary Skeletons
A fully differentiable end-to-end model for arbitrary-skeleton motion capture from video that jointly learns pose estimation and rotation recovery using reference pose-rotation pairs to eliminate IK ambiguities.