affiliation not provided to SSRN
talking face generation, Cross-modality generation, Teeth motion estimation, Multiple conditions constraints, Motion consistency