PCA and DWT based multimodal biometric recognition system

Abstract

This paper presents a distinctive approach of the authenticity of biometric system, improving the robustness with the help of multiple traits such as, face and ear using Discrete Wavelet Transform (DWT) and Principle Component Analysis (PCA). The wavelet coefficients are being extracted by using 'Haar' and 'Daubechies' wavelets, where only the approximation coefficients are being considered in this experimentation. The extracted approximation coefficients are further used to extract the principle components (eigenimages ) for recognition purpose. The experimentation with combination of face and ear will provide better results as explored in this paper.