An Adversarial Neuro-Tensorial Approach For Learning Disentangled Representations

TL;DR

This paper introduces an unsupervised deep learning method that uses a neuro-tensorial approach to learn disentangled representations of facial factors like expression and pose from in-the-wild images, enabling applications such as face editing and 3D reconstruction.

Contribution

It presents the first unsupervised deep learning framework employing a neuro-tensorial model to disentangle multiple facial factors in uncontrolled conditions.

Findings

Successfully disentangles facial expression and pose in wild images

Enables face editing, 3D reconstruction, and classification tasks

Demonstrates effectiveness of multilinear tensor modeling in deep learning

Abstract

Several factors contribute to the appearance of an object in a visual scene, including pose, illumination, and deformation, among others. Each factor accounts for a source of variability in the data, while the multiplicative interactions of these factors emulate the entangled variability, giving rise to the rich structure of visual object appearance. Disentangling such unobserved factors from visual data is a challenging task, especially when the data have been captured in uncontrolled recording conditions (also referred to as "in-the-wild") and label information is not available. In this paper, we propose the first unsupervised deep learning method (with pseudo-supervision) for disentangling multiple latent factors of variation in face images captured in-the-wild. To this end, we propose a deep latent variable model, where the multiplicative interactions of multiple latent factors of variation are explicitly modelled by means of multilinear (tensor) structure. We demonstrate that the proposed approach indeed learns disentangled representations of facial expressions and pose, which can be used in various applications, including face editing, as well as 3D face reconstruction and classification of facial expression, identity and pose.