- CVPR 2023
- Sigal Raab, Inbal Leibovitch, Peizhuo Li, Kfir Aberman, Olga Sorkine-Hornung, Daniel Cohen-Or
The emergence of neural networks has revolutionized the field of motion synthesis. Yet, learning to unconditionally synthesize motions from a given distribution remains chal- lenging, especially when the motions are highly diverse. In this work, we present MoDi – a generative model trained in an unsupervised setting from an extremely diverse, un- structured and unlabeled dataset. During inference, MoDi can synthesize high-quality, diverse motions. Despite the lack of any structure in the dataset, our model yields a well-behaved and highly structured latent space, which can be semantically clustered, constituting a strong motion prior that facilitates various applications including seman- tic editing and crowd simulation. In addition, we present an encoder that inverts real motions into MoDi’s natural mo- tion manifold, issuing solutions to various ill-posed chal- lenges such as completion from prefix and spatial editing. Our qualitative and quantitative experiments achieve state- of-the-art results that outperform recent SOTA techniques. Code and trained models are available at https://sigal- raab.github.io/MoDi.
This work was supported in part by the Israel Science Foundation (grants no. 2492/20 and 3441/21).