PCA and DWT Based Multimodal Biometric System for Recognizing a Person Using Face, Ear and Finger

This paper proposes the noble multimodal biometric system using multiple biometric traits such as face, ear and finger. This combination of multimodal traits is unique combination offering a better authenticity and accuracy than the existing multimodal biometric systems. In this experimentation the discrete wavelet transforms (DWT) is used for extraction of features which is followed by the principle component analysis (PCA). The wavelet coefficients are being extracted with the help of different wavelets, where only approximation coefficients are being considered in this experimentation. PCA offers dimension reduction by generating the principle components in the form of eigen faces which is used further for recognition followed by the classification.