Unsupervised Disentanglement of Linear-Encoded Facial Semantics

TL;DR

This paper introduces an unsupervised method to disentangle and interpret facial semantics encoded in StyleGAN's latent space, enabling meaningful manipulation and data augmentation without external labels.

Contribution

It presents a novel unsupervised approach combining linear regression and sparse learning to extract interpretable facial semantics from StyleGAN without supervision.

Findings

Disentangled facial semantics are interpretable and align with human intuition.

Guided extrapolation improves data augmentation for unbalanced datasets.

The method works effectively in the wild without external labels.

Abstract

We propose a method to disentangle linear-encoded facial semantics from StyleGAN without external supervision. The method derives from linear regression and sparse representation learning concepts to make the disentangled latent representations easily interpreted as well. We start by coupling StyleGAN with a stabilized 3D deformable facial reconstruction method to decompose single-view GAN generations into multiple semantics. Latent representations are then extracted to capture interpretable facial semantics. In this work, we make it possible to get rid of labels for disentangling meaningful facial semantics. Also, we demonstrate that the guided extrapolation along the disentangled representations can help with data augmentation, which sheds light on handling unbalanced data. Finally, we provide an analysis of our learned localized facial representations and illustrate that the semantic information is encoded, which surprisingly complies with human intuition. The overall unsupervised design brings more flexibility to representation learning in the wild.