Performance Improvement of Hyperspectral Face Recognition by Multimodal and Multi Algorithmic Feature Fusion of Hybrid and Kekre Wavelets Based Feature Vectors

The hyperspectral images are much different from normal images and contains huge data, so does its analysis requires deep understanding of exactly of which properties are to be considered and measured. Now a day's good access to such ready to use and quality data of such Hyperspectral face data has increased the opportunity of building such systems at ease. The hyperspectral face biometric is proving to be a successful addition to conventional face based biometric systems. The objective of proposed research is to recognize hyper spectral face images by considering and comparing the results obtained from the implementation of Multimodal system. The implementation of this Multimodal System is based on three main feature vector extraction algorithms based on Hybrid Wavelet Type I (HWI), Hybrid Wavelet Type II (HWII) and Kekre Wavelet (KW). These texture information Feature vectors are then put through Intra Class testing and Inter Class testing followed by performance analysis by Multi-instance Fusion and Multi-algorithmic Fusion. The performance comparison is based on its Performance Index (PI) and Security Performance Index (SPI) in different fusion options to obtain an optimized result.