Unconstrained Face Identification Based on 3D Face Frontalization and Support Vector Guided Dictionary Learning

Zhi Zhang,Xin Xu,Bingyu Sun,Jiuzhen Liang,Jerzy Baranowski

doi:10.1155/2020/8817571

Abstract

Face identification aims at putting a label on an unknown face with respect to some training set. Unconstrained face identification is a challenging problem because of the possible variations in face pose, illumination, occlusion, and facial expression. This paper presents an unconstrained face identification method based on face frontalization and learning-based data representation. Firstly, the frontal views of unconstrained face images are automatically generated by using a single, unchanged 3D face model. Then, we crop the face relevant regions of the frontal views to segment faces from the backgrounds. At last, to enhance the discriminative capability of the coding vectors, a support vector-guided dictionary learning (SVGDL) model is applied to adaptively assign different weights to different pairs of coding vectors. The performance of the proposed method FSVGDL (frontalization-based support vector guided dictionary learning) is evaluated on the Labeled Faces in the wild (LFW) database. After decision fusion, the identification accuracy yields 97.17% when using 7 images per individual for training and 3 images per individual for testing with 158 classes in total.

Highlights

As one of the biometric technologies, face recognition has developed rapidly in recent decades
Our contributions can be summarized as follows: (1) We propose a frontalization-based support vector guided dictionary learning (FSVGDL) method to cope with unconstrained face identification
(2) In this experiment, to show the effectiveness of our improved face frontalization method, we perform experiments on the LFW3D-hassner database for all the competing algorithms, and our frontalization-based support vector-guided dictionary learning (FSVGDL) method is still evaluated on the LFWsubdatabase: (1) In this experiment, we evaluate all the algorithms on the Labeled Faces in the wild (LFW)-subdatabase

Summary

Introduction

As one of the biometric technologies, face recognition has developed rapidly in recent decades. While in [12], Zhang et al claimed that it is the collaborative representation but not the l1-norm sparsity that makes SRC powerful for face recognition These methods have shown their ability to face recognition, directly using the whole training samples as the dictionary atoms may make them not effective enough to represent the query images. (1) We propose a frontalization-based support vector guided dictionary learning (FSVGDL) method to cope with unconstrained face identification. To address other variations in unconstrained images, we learn a support vector-guided dictionary which can adaptively assign different weights to different pairs of coding vectors for face representation.

Related Works

Frontalization-Based Support Vector-Guided Dictionary Learning

Experiments

Methods

Findings

Conclusions