Batch Face Alignment using a Low-rank GAN

TL;DR

This paper introduces a novel unsupervised method for face image alignment that leverages a low-rank GAN to handle occlusions, expressions, and illumination variations, achieving superior accuracy and efficiency.

Contribution

The paper proposes a low-rank GAN framework for face alignment that does not require ground-truth data and effectively separates aligned images from noise and outliers.

Findings

Higher accuracy than existing methods

Efficient alignment of faces with occlusions and expressions

Effective unsupervised learning approach

Abstract

This paper studies the problem of aligning a set of face images of the same individual into a normalized image while removing the outliers like partial occlusion, extreme facial expression as well as significant illumination variation. Our model seeks an optimal image domain transformation such that the matrix of misaligned images can be decomposed as the sum of a sparse matrix of noise and a rank-one matrix of aligned images. The image transformation is learned in an unsupervised manner, which means that ground-truth aligned images are unnecessary for our model. Specifically, we make use of the remarkable non-linear transforming ability of generative adversarial network(GAN) and guide it with low-rank generation as well as sparse noise constraint to achieve the face alignment. We verify the efficacy of the proposed model with extensive experiments on real-world face databases, demonstrating higher accuracy and efficiency than existing methods.