Enregistré dans:
Détails bibliographiques
Auteurs principaux: Hogue, Steven, Zhang, Chenxu, Daruger, Hamza, Tian, Yapeng, Guo, Xiaohu
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2409.07649
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866929498395508736
author Hogue, Steven
Zhang, Chenxu
Daruger, Hamza
Tian, Yapeng
Guo, Xiaohu
author_facet Hogue, Steven
Zhang, Chenxu
Daruger, Hamza
Tian, Yapeng
Guo, Xiaohu
contents Audio-driven talking video generation has advanced significantly, but existing methods often depend on video-to-video translation techniques and traditional generative networks like GANs and they typically generate taking heads and co-speech gestures separately, leading to less coherent outputs. Furthermore, the gestures produced by these methods often appear overly smooth or subdued, lacking in diversity, and many gesture-centric approaches do not integrate talking head generation. To address these limitations, we introduce DiffTED, a new approach for one-shot audio-driven TED-style talking video generation from a single image. Specifically, we leverage a diffusion model to generate sequences of keypoints for a Thin-Plate Spline motion model, precisely controlling the avatar's animation while ensuring temporally coherent and diverse gestures. This innovative approach utilizes classifier-free guidance, empowering the gestures to flow naturally with the audio input without relying on pre-trained classifiers. Experiments demonstrate that DiffTED generates temporally coherent talking videos with diverse co-speech gestures.
format Preprint
id arxiv_https___arxiv_org_abs_2409_07649
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DiffTED: One-shot Audio-driven TED Talk Video Generation with Diffusion-based Co-speech Gestures
Hogue, Steven
Zhang, Chenxu
Daruger, Hamza
Tian, Yapeng
Guo, Xiaohu
Computer Vision and Pattern Recognition
Audio-driven talking video generation has advanced significantly, but existing methods often depend on video-to-video translation techniques and traditional generative networks like GANs and they typically generate taking heads and co-speech gestures separately, leading to less coherent outputs. Furthermore, the gestures produced by these methods often appear overly smooth or subdued, lacking in diversity, and many gesture-centric approaches do not integrate talking head generation. To address these limitations, we introduce DiffTED, a new approach for one-shot audio-driven TED-style talking video generation from a single image. Specifically, we leverage a diffusion model to generate sequences of keypoints for a Thin-Plate Spline motion model, precisely controlling the avatar's animation while ensuring temporally coherent and diverse gestures. This innovative approach utilizes classifier-free guidance, empowering the gestures to flow naturally with the audio input without relying on pre-trained classifiers. Experiments demonstrate that DiffTED generates temporally coherent talking videos with diverse co-speech gestures.
title DiffTED: One-shot Audio-driven TED Talk Video Generation with Diffusion-based Co-speech Gestures
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.07649