Circular sector DCT based feature extraction for enhanced face recognition using histogram based dynamic gamma intensity correction
R. Gagan, Hardik Sanghavi, Sakthi Prabha Ramachandran, K. Manikantan · 2012
Face Recognition (FR) under varying lighting conditions is challenging and exacting illumination invariant features is an effective approach to solve this problem. In this paper, we propose a novel illumination normalization method called Histogram based Dynamic Gamma Intensity Correction, HDGIC, wherein the value of Λ is made to vary dynamically depending on the image. Also we propose a Circular sector DCT based Feature Extraction for enhancing the performance of the FR system. Individual stages of the FR system are examined and an attempt is made to improve each stage. A Binary Particle Swarm Optimization(BPSO)-based feature selection algorithm is used to search the feature vector space for the optimal feature subset. Experimental results, show the promising performance of quadrant of circle based DCT extraction technique together with HDGIC pre-processing for face recognition on Extended Yale B, Color FERET and ORL databases.