Face recognition using Entropy Weighted Patch PCA Array under variation of lighting conditions from a single sample image per person

The sensitivity to illumination changes is one of the most important issues for the evaluation of face recognition systems. In this paper, we propose a new approach to recognize face images under variation of lighting conditions when only one sample image per person is available. In this approach, a face image is represented as an array of Patch PCA (PPCA) extracted from a partitioned face image containing information of local regions instead of holistic information of a face. In order to adjust the contribution of each local region of a face in terms of the richness of identity information, an entropy-based weighting technique is utilized to assign proper weights to PPCA features. The encouraging experimental results using AR face database demonstrate that the proposed method provides a new solution to the problem of robustly recognizing face images under different lighting conditions in single model databases.

I. Introduction

Many pattern recognition and computer vision systems utilize the appearance-based paradigm for object recognition. One primary benefit of appearance-based methods is that it is not essential to create representations or models for objects since, for a given object, its model is now implicitly defined by the selection of the sample images of the object [1]. In appearance-based methods, we usually represent an image of size n×m pixels by a vector with dimensionality of n.m. However, in practice, this dimension is very large to perform fast and robust object recognition. A common way to deal with this problem is by using dimensionality reduction techniques. Two of the most popular techniques for this aim are Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) which is also known as Fisher Discriminant Analysis (FDA).