Denoising Diffusion Probabilistic Models for Styled Walking Synthesis

Findlay, Edmund; Zhang, Haozheng; Chang, Ziyi; Shum, Hubert P.H.

Abstract

Generating realistic motions for digital humans is time-consuming for many graphics applications. Data-driven motion synthesis approaches have seen solid progress in recent years through deep generative models. These results offer high-quality motions but typically suffer in motion style diversity. For the first time, we propose a framework using the denoising diffusion probabilistic model (DDPM) to synthesize styled human motions, integrating two tasks into one pipeline with increased style diversity compared with traditional motion synthesis methods. Experimental results show that our system can generate high-quality and diverse walking motions.

Citation

Findlay, E., Zhang, H., Chang, Z., & Shum, H. P. (2022, November). Denoising Diffusion Probabilistic Models for Styled Walking Synthesis. Presented at MIG 2022: The 15th Annual ACM SIGGRAPH Conference on Motion, Interaction and Games, Guanajuato, Mexico

• Presentation Conference Type: Conference Paper (published)
• Conference Name: MIG 2022: The 15th Annual ACM SIGGRAPH Conference on Motion, Interaction and Games
• Start Date: Nov 3, 2022
• End Date: Nov 5, 2022
• Acceptance Date: Sep 16, 2022
• Publication Date: 2022
• Deposit Date: Oct 4, 2022
• Publicly Available Date: Oct 5, 2022
• Publisher: Association for Computing Machinery (ACM)
• ISBN: 9781450398886
• DOI: https://doi.org/10.1145/3561975
• Public URL: https://durham-repository.worktribe.com/output/1135582
• Publisher URL: https://dl.acm.org/conference/mig