Intellectually combined face recognition using curvelet based principle component analysis for feature extraction and Bayesian Classifier

Biometric based face recognition provides an facial detection and verification system. The system is a ´No Human Touch´ technology. Because of this feature, face recognition systems have an edge over other biometric security products. No human touch feature makes it less prone to physical damage and human errors. In this paper, a new face recognition method based on 2D Level 2 Wavelet decomposition, PCA (principal Component Analysis) with singular value decomposition, and Bayesian Classifier is proposed. This method consists of three steps: i) Preprocessing, ii) feature extraction using curvelet, PCA with Singular value decomposition iii) classification and recognition using Bayes´ algorithm. Combination of PCA, with Singular Value Decomposition and Bayesian classifier is used for improving the rate of recognition when a few samples of images are available. Bayesian classifier is used to reduce the number of an misclassification caused by non-linearly separable classes. The proposed method provides a fast computation, relatively simple and works well in an constrained environment. This type of recognition can play an important role for authentication purpose in security related areas such as airport, banking, and secret missions.