Hybrid Approach for Face Recognition Using DWT and LBP

Abstract

Authentication of individuals plays a vital role to check intrusions in any online digital system. Most commonly and securely used techniques are biometric fingerprint reader and face recognition. Face recognition is the process of identification of individuals by their facial images, as faces are rarely matched. Face recognition technique merely considering test images and compare this with number of trained images stored in database and then conclude whether the test images matches with any trained images. In this paper we have discussed two hybrid techniques local binary pattern (LBP) and Discrete Wavelet Transform (DWT) for face images to extract feature stored in database by applying principal component analysis for fusion and same process is done for test images. Then K-nearest neighbor (KNN) classifier is used to classify images and measure the accuracy. Our proposed model achieved 95% accuracy. The aim of this paper is to develop a robust method for face recognition and classification of individuals to improve the recognition rate, efficiency of the system and for lesser complexity.