RANA: Relightable Articulated Neural Avatars

TL;DR

RANA is a neural avatar framework that enables photorealistic human rendering under arbitrary poses and lighting from minimal input, disentangling geometry, texture, and lighting for realistic synthesis.

Contribution

It introduces a novel method for creating relightable articulated neural avatars from short videos without prior lighting knowledge.

Findings

Pretraining on synthetic data improves disentanglement and robustness.

RANA achieves photorealistic rendering under diverse poses and lighting.

The Relighting Humans dataset enables quantitative evaluation of relighting methods.

Abstract

We propose RANA, a relightable and articulated neural avatar for the photorealistic synthesis of humans under arbitrary viewpoints, body poses, and lighting. We only require a short video clip of the person to create the avatar and assume no knowledge about the lighting environment. We present a novel framework to model humans while disentangling their geometry, texture, and also lighting environment from monocular RGB videos. To simplify this otherwise ill-posed task we first estimate the coarse geometry and texture of the person via SMPL+D model fitting and then learn an articulated neural representation for photorealistic image generation. RANA first generates the normal and albedo maps of the person in any given target body pose and then uses spherical harmonics lighting to generate the shaded image in the target lighting environment. We also propose to pretrain RANA using synthetic images and demonstrate that it leads to better disentanglement between geometry and texture while also improving robustness to novel body poses. Finally, we also present a new photorealistic synthetic dataset, Relighting Humans, to quantitatively evaluate the performance of the proposed approach.