A Deep Emulator for Secondary Motion of 3D Characters

TL;DR

This paper introduces a neural network-based emulator that efficiently generates realistic secondary motion in 3D character animations, offering a fast and adaptable alternative to traditional physics-based simulations.

Contribution

A novel local neural network model that encodes internal forces for 3D character animation, generalizes across mesh topologies, and allows easy adjustment of material properties.

Findings

Over 30 times faster than traditional physics simulation

Outperforms existing fast approximation methods

Generalizes to arbitrary mesh shapes

Abstract

Fast and light-weight methods for animating 3D characters are desirable in various applications such as computer games. We present a learning-based approach to enhance skinning-based animations of 3D characters with vivid secondary motion effects. We design a neural network that encodes each local patch of a character simulation mesh where the edges implicitly encode the internal forces between the neighboring vertices. The network emulates the ordinary differential equations of the character dynamics, predicting new vertex positions from the current accelerations, velocities and positions. Being a local method, our network is independent of the mesh topology and generalizes to arbitrarily shaped 3D character meshes at test time. We further represent per-vertex constraints and material properties such as stiffness, enabling us to easily adjust the dynamics in different parts of the mesh. We evaluate our method on various character meshes and complex motion sequences. Our method can be over 30 times more efficient than ground-truth physically based simulation, and outperforms alternative solutions that provide fast approximations.