VariTex: Variational Neural Face Textures

TL;DR

VariTex introduces a variational neural face texture model that enables independent control over identity, pose, shape, and expressions, facilitating realistic and customizable human head image generation.

Contribution

It is the first to learn a variational latent space of neural face textures combined with explicit pose and expression control using a novel additive decoder.

Findings

Enables sampling of novel identities with independent control.

Produces geometrically consistent head images under various poses.

Supports applications like identity sampling, pose adjustment, and expression transfer.

Abstract

Deep generative models can synthesize photorealistic images of human faces with novel identities. However, a key challenge to the wide applicability of such techniques is to provide independent control over semantically meaningful parameters: appearance, head pose, face shape, and facial expressions. In this paper, we propose VariTex - to the best of our knowledge the first method that learns a variational latent feature space of neural face textures, which allows sampling of novel identities. We combine this generative model with a parametric face model and gain explicit control over head pose and facial expressions. To generate complete images of human heads, we propose an additive decoder that adds plausible details such as hair. A novel training scheme enforces a pose-independent latent space and in consequence, allows learning a one-to-many mapping between latent codes and pose-conditioned exterior regions. The resulting method can generate geometrically consistent images of novel identities under fine-grained control over head pose, face shape, and facial expressions. This facilitates a broad range of downstream tasks, like sampling novel identities, changing the head pose, expression transfer, and more. Code and models are available for research on https://mcbuehler.github.io/VariTex.