Modiff: Action-Conditioned 3D Motion Generation with Denoising Diffusion Probabilistic Models

TL;DR

Modiff introduces a novel diffusion probabilistic model for generating diverse, realistic 3D skeleton-based motions conditioned on actions, demonstrating superior performance on a large-scale dataset.

Contribution

This work pioneers the use of DDPM for action-conditioned 3D motion synthesis, enabling variable-length sequence generation conditioned on categorical actions.

Findings

Outperforms state-of-the-art motion generation methods

Generates diverse and realistic 3D skeleton motions

Effective on large-scale NTU RGB+D dataset

Abstract

Diffusion-based generative models have recently emerged as powerful solutions for high-quality synthesis in multiple domains. Leveraging the bidirectional Markov chains, diffusion probabilistic models generate samples by inferring the reversed Markov chain based on the learned distribution mapping at the forward diffusion process. In this work, we propose Modiff, a conditional paradigm that benefits from the denoising diffusion probabilistic model (DDPM) to tackle the problem of realistic and diverse action-conditioned 3D skeleton-based motion generation. We are a pioneering attempt that uses DDPM to synthesize a variable number of motion sequences conditioned on a categorical action. We evaluate our approach on the large-scale NTU RGB+D dataset and show improvements over state-of-the-art motion generation methods.