Deformable Shape Completion with Graph Convolutional Autoencoders

TL;DR

This paper introduces a graph convolutional autoencoder-based method for completing partial 3D shapes, especially non-rigid objects like human bodies and faces, handling occlusions and deformations effectively.

Contribution

It presents a novel variational autoencoder with graph convolutions that models realistic complete shapes from partial scans, focusing on non-rigid deformations.

Findings

Effective completion of synthetic and real scans

Handles non-rigid deformations like articulation

Produces realistic shape reconstructions

Abstract

The availability of affordable and portable depth sensors has made scanning objects and people simpler than ever. However, dealing with occlusions and missing parts is still a significant challenge. The problem of reconstructing a (possibly non-rigidly moving) 3D object from a single or multiple partial scans has received increasing attention in recent years. In this work, we propose a novel learning-based method for the completion of partial shapes. Unlike the majority of existing approaches, our method focuses on objects that can undergo non-rigid deformations. The core of our method is a variational autoencoder with graph convolutional operations that learns a latent space for complete realistic shapes. At inference, we optimize to find the representation in this latent space that best fits the generated shape to the known partial input. The completed shape exhibits a realistic appearance on the unknown part. We show promising results towards the completion of synthetic and real scans of human body and face meshes exhibiting different styles of articulation and partiality.