The paper introduces DREAM-Talk, a two-stage diffusion-based framework for generating emotional talking faces. It achieves both expressive emotional talking and accurate lip-sync by using a novel diffusion module and a video-to-video rendering module. DREAM-Talk outperforms state-of-the-art methods in terms of expressiveness, lip-sync accuracy, and perceptual quality.

https://arxiv.org/abs//2312.13578

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers