EmbedTalk: Triplane-Free Talking Head Synthesis using Embedding-Driven Gaussian Deformation

Arpita Saggar, Jonathan C. Darling, Duygu Sarikaya +1

Real-time talking head synthesis increasingly relies on deformable 3D Gaussian Splatting (3DGS) due to its low latency. Tri-planes are the standard choice for encoding Gaussians pr…