Dynamic Surface Function Networks for Clothed Human Bodies

TL;DR

This paper introduces a novel dynamic surface function network that enables temporally coherent reconstruction and animation of clothed human bodies from monocular RGB-D sequences, leveraging self-supervised learning and differentiable rasterization.

Contribution

It proposes a person-specific, pose-conditioned surface model using a multi-layer perceptron embedded in SMPL space, learned in a self-supervised manner for coherent 3D reconstruction and animation.

Findings

Achieves temporally coherent mesh sequences from monocular RGB-D data.

Enables synthesis of new animations with pose-dependent deformations.

Uses self-supervised learning with differentiable rasterization for surface modeling.

Abstract

We present a novel method for temporal coherent reconstruction and tracking of clothed humans. Given a monocular RGB-D sequence, we learn a person-specific body model which is based on a dynamic surface function network. To this end, we explicitly model the surface of the person using a multi-layer perceptron (MLP) which is embedded into the canonical space of the SMPL body model. With classical forward rendering, the represented surface can be rasterized using the topology of a template mesh. For each surface point of the template mesh, the MLP is evaluated to predict the actual surface location. To handle pose-dependent deformations, the MLP is conditioned on the SMPL pose parameters. We show that this surface representation as well as the pose parameters can be learned in a self-supervised fashion using the principle of analysis-by-synthesis and differentiable rasterization. As a result, we are able to reconstruct a temporally coherent mesh sequence from the input data. The underlying surface representation can be used to synthesize new animations of the reconstructed person including pose-dependent deformations.