Abstract

There have been decades of research on face recognition, and the performance of many state-of-the-art face recognition algorithms under well-conditioned environments has become saturated. Accordingly, recent research efforts have focused on difficult but practical challenges. One such issue is the single sample per person (SSPP) problem, i.e., the case where only one training image of each person. While this problem is challenging because it is difficult to establish the within-class variation, working toward its solution is very practical because often only a few images of a person are available. To address the SSPP problem, we propose an efficient coupled bilinear model that generates virtual images under various illuminations using a single input image. The proposed model is inspired by the knowledge that the illuminance of an image is not sensitive to the poor quality of a subspace-based model, and it has a strong correlation to the image itself. Accordingly, a coupled bilinear model was constructed that retrieves the illuminance information from an input image. This information is then combined with the input image to estimate the texture information, from which we can generate virtual illumination conditions. The proposed method can instantly generate numerous virtual images of good quality, and these images can then be utilized to train the feature space for resolving SSPP problems. Experimental results show that the proposed method outperforms the existing algorithms.

Highlights

  • For the last 30 years, researchers have investigated issues associated with finding and developing effective methods of face recognition [1,2]

  • This paper focuses on addressing variations in illumination, i.e., we explore the possibility of synthesizing virtual images from a single image with various realistic illuminations while considering quality, efficiency, and practicality

  • In order to reduce the gap between these two conflicting goals, in this paper, we propose a novel face relighting method for single sample per person (SSPP) problems which instantly synthesizes various relit images with good qualities

Read more

Summary

Introduction

For the last 30 years, researchers have investigated issues associated with finding and developing effective methods of face recognition [1,2]. Many issues are yet to be resolved in order to expand the range of effective applications of face recognition. The difficulty in face recognition is caused by intrinsic factors, such as varied expressions, aging, makeup, or worn accessories, and extrinsic factors, such as varied poses or illumination that could change the facial image to resemble various other facial images. The methods that have been developed and employed for face recognition far can be divided into two types: those that use three-dimensional (3D) face models [3,4] and those that use the features from two-dimensional (2D) facial images [5,6]. Obtaining a 3D face model requires special equipments or needs numerous computational resources; limitations are unavoidable for Sensors 2019, 19, 43; doi:10.3390/s19010043 www.mdpi.com/journal/sensors

Objectives
Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call