Pixel-Level Alignment of Facial Images for High Accuracy Recognition Using Ensemble of Patches

TL;DR

This paper introduces a pixel-level face alignment method that improves recognition accuracy by mapping facial images to a reference face, effectively handling pose and expression variations, and combining intensity and geometry information.

Contribution

The novel pixel-wise geometry alignment method enhances face recognition accuracy over traditional eye-alignment techniques by effectively removing pose and expression variations.

Findings

Recognition rates improved by 24-33% on Yale and AT&T datasets.

20% improvement in LFW dataset at FAR of 0.1.

Significant reduction in pose and expression effects.

Abstract

The variation of pose, illumination and expression makes face recognition still a challenging problem. As a pre-processing in holistic approaches, faces are usually aligned by eyes. The proposed method tries to perform a pixel alignment rather than eye-alignment by mapping the geometry of faces to a reference face while keeping their own textures. The proposed geometry alignment not only creates a meaningful correspondence among every pixel of all faces, but also removes expression and pose variations effectively. The geometry alignment is performed pixel-wise, i.e., every pixel of the face is corresponded to a pixel of the reference face. In the proposed method, the information of intensity and geometry of faces are separated properly, trained by separate classifiers, and finally fused together to recognize human faces. Experimental results show a great improvement using the proposed method in comparison to eye-aligned recognition. For instance, at the false acceptance rate of 0.001, the recognition rates are respectively improved by 24% and 33% in Yale and AT&T datasets. In LFW dataset, which is a challenging big dataset, improvement is 20% at FAR of 0.1.