DeePSD: Automatic Deep Skinning And Pose Space Deformation For 3D Garment Animation

TL;DR

DeePSD introduces a deep learning approach for automatically animating 3D garments of arbitrary topology, improving scalability and compatibility with existing graphics engines by mapping outfits to standard animation formats.

Contribution

The paper presents a simple, topology-agnostic deep learning model for garment animation, combining supervised and unsupervised physically based learning to enhance cloth quality and collision handling.

Findings

Able to animate any outfit regardless of topology or vertex order

Automatically generates blend weights and shapes compatible with standard engines

Improves cloth realism through physically based unsupervised learning

Abstract

We present a novel solution to the garment animation problem through deep learning. Our contribution allows animating any template outfit with arbitrary topology and geometric complexity. Recent works develop models for garment edition, resizing and animation at the same time by leveraging the support body model (encoding garments as body homotopies). This leads to complex engineering solutions that suffer from scalability, applicability and compatibility. By limiting our scope to garment animation only, we are able to propose a simple model that can animate any outfit, independently of its topology, vertex order or connectivity. Our proposed architecture maps outfits to animated 3D models into the standard format for 3D animation (blend weights and blend shapes matrices), automatically providing of compatibility with any graphics engine. We also propose a methodology to complement supervised learning with an unsupervised physically based learning that implicitly solves collisions and enhances cloth quality.