NeMF: Neural Motion Fields for Kinematic Animation
This work addresses the challenge of continuous motion representation for animation and robotics, offering a versatile prior for task-agnostic problems, though it is incremental as it builds on existing neural implicit representations.
The authors tackled the problem of representing kinematic motions as discrete sequential samples by proposing Neural Motion Fields (NeMF), an implicit neural representation that models motion as a continuous function over time, resulting in a generative model that outperforms existing methods in tasks like motion interpolation and editing.
We present an implicit neural representation to learn the spatio-temporal space of kinematic motions. Unlike previous work that represents motion as discrete sequential samples, we propose to express the vast motion space as a continuous function over time, hence the name Neural Motion Fields (NeMF). Specifically, we use a neural network to learn this function for miscellaneous sets of motions, which is designed to be a generative model conditioned on a temporal coordinate $t$ and a random vector $z$ for controlling the style. The model is then trained as a Variational Autoencoder (VAE) with motion encoders to sample the latent space. We train our model with a diverse human motion dataset and quadruped dataset to prove its versatility, and finally deploy it as a generic motion prior to solve task-agnostic problems and show its superiority in different motion generation and editing applications, such as motion interpolation, in-betweening, and re-navigating. More details can be found on our project page: https://cs.yale.edu/homes/che/projects/nemf/.