Reconstructing Animatable 3D Categories from Videos

Without modeling the soft deformation, our method fails to handle motion not explained by the skeletal movements, such as the shaking ear of a dog.

Ours

without soft deformation field