Component-based Cascade Linear Discriminant Analysis for Face Recognition.pdf

Component-based Cascade Linear Discriminant Analysis for Face Recognition Wenchao Zhang1, Shiguang Shan2, Wen Gao2, Yizheng Chang1 and Bo Cao2 1 School of Computer Science and Technology, Harbin Institute of Technology 150001 Harbin, P.R.China {wczhang, yzchang}@, 2 ICT-ISVISION Joint RD Laboratory for Face Recognition, CAS, 100080 Beijing, P.R.China {sgshan, wgao, bcao}@ Abstract. This paper presents a novel face recognition method based on cascade Linear Discriminant Analysis (LDA) of the component-based face representation. In the proposed method, a face image is represented as four components with overlap at the neighboring area rather than a whole face patch. Firstly, LDA is conducted on the principal components of each component individually to extract component discriminant features. Then, these features are further concatenated to undergo another LDA to extract the final face descriptor, which actually have assigned different weights to different component features. Our experiments on the FERET face database have illustrated the effectiveness of the proposed method compared with the traditional Fisherface method both for face recognition and verification. 1 Introduction Over the past 20 years, numerous algorithms have been proposed for face recognition. See detailed surveys [1][2][3]. In the following we will give a brief overview of face recognition methods. In the early researches, methods based on geometric feature and template matching used to be popular technologies, which were compared in 1992 by Brunelli and Poggio. Their conclusion showed that template matching based algorithms outperformed the geometric feature based ones [4]. Therefore, since the 1990s, methods based on appearance have been dominant researches. In these methods, each pixel in a face image is considered as a coordinate in a high-dimensional space and the classification is carried out in a low-dimensional feature space projected from the image space