Multimodal geometry modeling for scenes and hands
Predict temporally consistent scene geometry to improve reconstruction efficiency in dynamic scenes.
Fuse multi-sensor inputs to model hands and objects, addressing pose estimation challenges.
