URAvatar: Universal Relightable Gaussian Codec Avatars

TL;DR

This paper introduces URAvatar, a method for creating photorealistic, relightable head avatars from phone scans that can be animated and relit in real time across diverse environments, using a universal Gaussian-based model.

Contribution

It proposes a universal 3D Gaussian-based avatar model trained on multi-view scans, enabling real-time relighting and personalization from a single phone scan.

Findings

Outperforms existing relightable avatar methods.

Achieves real-time rendering with high realism.

Successfully personalizes avatars from phone scans.

Abstract

We present a new approach to creating photorealistic and relightable head avatars from a phone scan with unknown illumination. The reconstructed avatars can be animated and relit in real time with the global illumination of diverse environments. Unlike existing approaches that estimate parametric reflectance parameters via inverse rendering, our approach directly models learnable radiance transfer that incorporates global light transport in an efficient manner for real-time rendering. However, learning such a complex light transport that can generalize across identities is non-trivial. A phone scan in a single environment lacks sufficient information to infer how the head would appear in general environments. To address this, we build a universal relightable avatar model represented by 3D Gaussians. We train on hundreds of high-quality multi-view human scans with controllable point lights. High-resolution geometric guidance further enhances the reconstruction accuracy and generalization. Once trained, we finetune the pretrained model on a phone scan using inverse rendering to obtain a personalized relightable avatar. Our experiments establish the efficacy of our design, outperforming existing approaches while retaining real-time rendering capability.