Neural Face Editing with Intrinsic Image Disentangling

TL;DR

This paper introduces an end-to-end GAN that disentangles intrinsic face properties like shape, albedo, and lighting from in-the-wild images, enabling flexible and semantically meaningful face editing.

Contribution

It presents a novel GAN framework with an integrated physically-based image formation model for intrinsic face property disentanglement from unconstrained images.

Findings

Effective disentangling of face attributes in wild images

Facilitates semantically relevant facial edits

Supports various facial editing applications

Abstract

Traditional face editing methods often require a number of sophisticated and task specific algorithms to be applied one after the other --- a process that is tedious, fragile, and computationally intensive. In this paper, we propose an end-to-end generative adversarial network that infers a face-specific disentangled representation of intrinsic face properties, including shape (i.e. normals), albedo, and lighting, and an alpha matte. We show that this network can be trained on "in-the-wild" images by incorporating an in-network physically-based image formation module and appropriate loss functions. Our disentangling latent representation allows for semantically relevant edits, where one aspect of facial appearance can be manipulated while keeping orthogonal properties fixed, and we demonstrate its use for a number of facial editing applications.