We introduce a framework for learning latent representations of 4D objects which are descriptive, faithfully capturing object geometry and appearance; compressive, aiding in downstream efficiency; and accessible, requiring minimal input, i.e., an unstructured dynamic point cloud, to construct. Specifically, Velox trains an encoder to compress spatiotemporal color point clouds into a set of dynamic shape tokens. These tokens are supervised using two complementary decoders: a 4D surface decoder, which models the time-varying surface distribution capturing the geometry; and a Gaussian decoder, which maps the tokens to 3D Gaussians, helping learn appearance. To demonstrate the utility of our representation, we evaluate it across three downstream tasks -- video-to-4D generation, 3D tracking, and cloth simulation via image-to-4D generation -- and observe strong performances in all settings.
@article{arxiv.2605.04527,
title = {Velox: Learning Representations of 4D Geometry and Appearance},
author = {Anagh Malik and Dorian Chan and Xiaoming Zhao and David B. Lindell and Oncel Tuzel and Jen-Hao Rick Chang},
journal= {arXiv preprint arXiv:2605.04527},
year = {2026}
}