Comparative analysis of hybridizing histogram equalization techniques on the performance of face recognition algorithms
Illumination invariance has been one of the challenging factor faced by face recognition systems, among others. Various techniques have been proposed to combat this problem of illumination variation in face recognition, among which are histogram equalization (HE), adaptive histogram equalization (AHE), fuzzy histogram equalization, contrast limited adaptive histogram equalization (CLAHE), block-based histogram equalization (BHE), to mention a few. Each of these techniques have been shown to improve performance of face recognition system in terms of recognition rate, with most solving the limitations of the other. This paper investigates the efficacy of hybridizing histogram equalization techniques, in particular Histogram Equalization (HE) and Contrast Limited Adaptive Histogram Equalization (CLAHE) as illumination normalization filter, on the performance of face recognition algorithms. Performance was evaluated using Principal Component Analysis (PCA) and Linear Disciminant Analysis (LDA) and our result showed that our hybrid model outperform the popular CLAHE which according to past researches, outperforms its variants, Histogram Equalization (HE) and Adaptive Histogram Equalization (AHE)
Keywords: illumination invariance, face recognition system, Histogram Equalization, Adaptive Histogram Equalization, Contrast Limited Adaptive Histogram Equalization, Fuzzy Histogram Equalization, Principal Component Analysis, linear discriminant analysis..