StyleRig: Rigging StyleGAN for 3D Control over Portrait Images

TL;DR

This paper introduces StyleRig, a method that combines the photorealism of StyleGAN with 3D semantic control over portrait images by rigging StyleGAN with a 3D morphable face model using a self-supervised network.

Contribution

The paper presents RigNet, a novel self-supervised network that enables explicit 3D semantic control over StyleGAN-generated portraits by integrating a 3D morphable face model.

Findings

Enables control over face pose, expressions, and illumination in StyleGAN images.

Produces photorealistic portraits with adjustable 3D semantic parameters.

Operates without manual annotations through self-supervised training.

Abstract

StyleGAN generates photorealistic portrait images of faces with eyes, teeth, hair and context (neck, shoulders, background), but lacks a rig-like control over semantic face parameters that are interpretable in 3D, such as face pose, expressions, and scene illumination. Three-dimensional morphable face models (3DMMs) on the other hand offer control over the semantic parameters, but lack photorealism when rendered and only model the face interior, not other parts of a portrait image (hair, mouth interior, background). We present the first method to provide a face rig-like control over a pretrained and fixed StyleGAN via a 3DMM. A new rigging network, RigNet is trained between the 3DMM's semantic parameters and StyleGAN's input. The network is trained in a self-supervised manner, without the need for manual annotations. At test time, our method generates portrait images with the photorealism of StyleGAN and provides explicit control over the 3D semantic parameters of the face.