motion-transfer

SCAIL-2 is a framework that achieves end-to-end controlled character animation by directly transferring motion from driving videos without intermediate representations, using unified task decomposition, synthetic data (MotionPair-60K), and novel conditioning techniques like in-context mask conditioning and Bias-Aware DPO.

OmniHumanoid is a framework that enables scalable cross-embodiment video generation by factorizing motion transfer and embodiment-specific adaptation, using unpaired data and branch-isolated attention to reduce interference.