CVPR2023
LipFormer: High-fidelity and Generalizable Talking Face Generation with A Pre-learned Facial Codebook
Jiayu Wang, Kang Zhao, Shiwei Zhang, Yingya Zhang, Yujun Shen, Deli Zhao, Jingren Zhou
Abstract
Figure 1. High-fidelity talking face generation with LipFormer. Top: Five target face pairs. Middle: LipFormer-generated results, driven by target face's own audio. Bottom: LipFormer-generated results, after exchanging the audio of each target pair. It is clear that LipFormer successfully captures the relationship between voice and mouth shape.