1 paper
Xiangyue Zhang, Yiyi Cai, Kunhang Li +6
We propose PersonaGesture, a diffusion-based pipeline for single-reference co-speech gesture personalization of unseen speakers. Given target speech and one motion clip from a new…