Evaluation of face recognition techniques using PCA, wavelets and SVM

In this study, we present an evaluation of using various methods for face recognition. As feature extracting techniques we benefit from wavelet decomposition and Eigenfaces method which is based on Principal Component Analysis (PCA). After generating feature vectors, distance classifier and Support Vector Machines (SVMs) are used for classification step. We examined the classification accuracy according to increasing dimension of training set, chosen feature extractor–classifier pairs and chosen kernel function for SVM classifier. As test set we used ORL face database which is known as a standard face database for face recognition applications including 400 images of 40 people. At the end of the overall separation task, we obtained the classification accuracy 98.1% with Wavelet–SVM approach for 240 image training set. As a special study of pattern recognition, face recognition has had crucial effects in daily life especially for security purposes. Face recognition task is actively being used at airports, employee entries , criminal detection systems, etc. For this task many methods have been proposed and tested. Most of these methods have trade off's like hardware requirements, time to update image database, time for feature extraction, response time. Generally face recognition methods are composed of a feature extractor (like PCA, Wavelet decomposer) to reduce the size of input and a classifier like Neural Networks, Support Vector Machines, Nearest Distance Classifiers to find the features which are most likely to be looked for. In this study, we chose wavelet decomposition and Eigenfaces method which is based on Principal Component Analysis (PCA) as main techniques for data reduction and feature extraction. PCA is an efficient and long term studied method to extract feature sets by creating a feature space. PCA also has low computation time which is an important advantage. On the other hand because of being a linear feature extraction method, PCA is inefficient especially when nonlinearities are present in the underlying relationships (Kursun & Favorov, 2004). Wavelet decomposition is a multilevel dimension reduction process that makes time–space–frequency analysis. Unlike Fourier transform, which provides only frequency analysis of signals, wavelet transforms provide time–frequency analysis, which is particularly useful for pattern recognition (Gorgel, Sertbas, Kilic, Ucan, & Osman, 2009). In this study, we used available 40 classes in the ORL face recognition dataset (ORL Database of Faces, 1994). Eigenfaces and Discrete Wavelet Transform are used for feature extractor. For the classification step, we consider Support Vector Machines (SVM) and nearest distance classification …