SMooDi: Stylized Motion Diffusion Model

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhong, Lei, Xie, Yiming, Jampani, Varun, Sun, Deqing, Jiang, Huaizu
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866907982728527872
author Zhong, Lei
Xie, Yiming
Jampani, Varun
Sun, Deqing
Jiang, Huaizu
author_facet Zhong, Lei
Xie, Yiming
Jampani, Varun
Sun, Deqing
Jiang, Huaizu
contents We introduce a novel Stylized Motion Diffusion model, dubbed SMooDi, to generate stylized motion driven by content texts and style motion sequences. Unlike existing methods that either generate motion of various content or transfer style from one sequence to another, SMooDi can rapidly generate motion across a broad range of content and diverse styles. To this end, we tailor a pre-trained text-to-motion model for stylization. Specifically, we propose style guidance to ensure that the generated motion closely matches the reference style, alongside a lightweight style adaptor that directs the motion towards the desired style while ensuring realism. Experiments across various applications demonstrate that our proposed framework outperforms existing methods in stylized motion generation.
format Preprint
id arxiv_https___arxiv_org_abs_2407_12783
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SMooDi: Stylized Motion Diffusion Model
Zhong, Lei
Xie, Yiming
Jampani, Varun
Sun, Deqing
Jiang, Huaizu
Computer Vision and Pattern Recognition
Graphics
We introduce a novel Stylized Motion Diffusion model, dubbed SMooDi, to generate stylized motion driven by content texts and style motion sequences. Unlike existing methods that either generate motion of various content or transfer style from one sequence to another, SMooDi can rapidly generate motion across a broad range of content and diverse styles. To this end, we tailor a pre-trained text-to-motion model for stylization. Specifically, we propose style guidance to ensure that the generated motion closely matches the reference style, alongside a lightweight style adaptor that directs the motion towards the desired style while ensuring realism. Experiments across various applications demonstrate that our proposed framework outperforms existing methods in stylized motion generation.
title SMooDi: Stylized Motion Diffusion Model
topic Computer Vision and Pattern Recognition
Graphics
url https://arxiv.org/abs/2407.12783