Identity management based on PCA and SVM Lixin Shen 1 & Hong Wang 2,3 & Li Da Xu 4 & Xue Ma 1 & Sohail Chaudhry 5 & Wu He 4 # Springer Science+Business Media New York 2015 Abstract A new approach for face recognition, based on ker- nel principal component analysis (KPCA) and support vector machines (SVMs), is presented to improve the recognition performance of the method based on principal component analysis (PCA). This method can simultaneously be applied to solve both the over-fitting problem and the small sample problem. The KPCA method is performed on every facial image of the training set to get the core facial features of the training samples. To ensure that the loss of the image infor- mation will be as less as possible, the facial data of high- dimensional feature space is projected into low-dimensional space, and then the SVM face recognition model is established to identify the low-dimensional space facial data. Our exper- imental results demonstrate that the approach proposed in this paper is efficient, and the recognition accuracy of the pro- posed method reaches 95.4 %. Keywords Kernel principal component analysis . Support vector machine . Kernel function methods . Face recognition . Pattern recognition 1 Introduction With the rapid development of science and technology, people can make use of many convenient and efficient tools to obtain images, videos, audio and other digital information and then, processing information through various multimedia software, and extract the information they are interested in. One of the areas that has made great progress is biometric identification. It uses face recognition as a means for identification and has a wide variety of applications, including access control systems, face registration system, identity certification, crime face de- tection, etc. It has become one of the most popular research areas in the past decade (Bartlett et al. 2002). This paper first investigates KPCA feature-extraction methods. These methods combine principal component anal- ysis and the kernel function, use nonlinear mapping to map each picture to the low-dimensional feature space, and then apply the support vector machine (SVM) classifier to the ker- nel feature of the feature space for training and classifying. Because support vector machines have unique advantages compared to other machine learning methods (Du and Lv 2013; Duan and Xu 2012; Ma et al. 2014; Pan et al. 2014; Wang et al. 2009; Xing et al. 2013; Xu 2013; Yuan et al. 2008), such as managing small samples and handling nonlin- ear and over-fitting learning issues, KPCA can extract better features in situations where there is less information. If we combine KPCA feature-extraction and SVM classification, the face recognition problem can be more optimally solved. Figure 1 shows the recognition process. 2 Kernel function Kernel Function Definition: Let n-dimensional random vector x be a non-empty set, F be an inner product space, and ϕ be the * Hong Wang HWANG@NCAT.EDU 1 College of Transportation Management, Dalian Maritime University, Dalian, China 2 School of Business and Economics, North Carolina A&T State University, Greensboro, NC 27411, USA 3 School of Information, Yunnan University of Finance and Economics, Kunming, China 4 Old Dominion University, Norfolk, VA 23529, USA 5 Villanova University, Villanova, PA 19085, USA Inf Syst Front DOI 10.1007/s10796-015-9551-8