Local Binary Pattern based face recognition system for automotive security

In this paper we present an approach for face recognition under varying poses, illumination variation and facial expressions. Illumination variation, Pose variation and facial expressions are the main challenges among the various factors that severely affects the performance of the face recognition. The main aim of this paper is to calculate and evaluate the performance of combination of Independent Component Analysis and local binary pattern approach for different face databases that contains number of images with illumination variation, varying poses and facial expressions. 3 images per subject are used for training purpose and remaining images for testing or recognition purpose. Uniform Local Binary Pattern is used for extracting the features. These extracted features reduced by independent component analysis and Euclidean distance is used for matching. We have calculated FRR and TSR parameter which gives accuracy of given method. Finally comparing the results for different face databases.