Histogram equalization and its variants on principal component analysis based face recognition system
Principal Component Analysis (PCA) is an appearance-based feature extraction technique that is widely used in facial recognition system which suffers from illumination conditions, thus the challenge of knowing which illumination control methods to be used in facial recognition system based on PCA algorithm is very important. This paper compared the performance of Principal Component Analysis based face recognition using illumination normalization techniques: Histogram Equalization (HE), Adaptive Histogram Equalization (AHE) and Contrast Limited Adaptive Histogram Equalization (CLAHE). The evaluation of the performance of the developed system was done comparatively in terms of false acceptance rate, false rejection rate and recognition rate. The best recognition rate of 83.67% was achieved in AHE for ORL, the effective false acceptance rate (FAR) of 20% was recorded in CLAHE with FERET database and also effective false recognition rate (FRR) of 10% was obtained in CLAHE for FERET database.
Keywords: Histogram Equalization, Adaptive Histogram Equalization, Contrast Adaptive Limited Histogram Equalization, Principal Component Analysis