Multispectral Image Fusion System Based on Wavelet Transformation for Secure Human Recognition

Abstract

 Advancements in Biometric system are contemporary to work with images in multispectral domain using wavelet. Fingerprint biometric is widely used for human identification for robust and accurate result. Wavelet transform method has been extensively used in image processing due to its important role in texture classification and image compression. In this paper a  framework has been proposed for a biometric system using multi –spectral fingerprint images and applying wavelet decomposition to improve the performance of biometric system as compared to the existing conventional methods. It has been observed that the overall performance of the system may be improved by firstly applying the wavelet decomposition function to the image and then extracting the key feature points of the resultant decomposed image which result in robust identification and high accuracy and the security of the biometric system is improved. Moreover, the proposed combination of wavelet transformation and feature extraction would protect the biometric system against the biometric attacks or fabrication.

Keywords: Fingerprint, Wavelet Decomposition, Multispectral, Feature point, Minutiae , LL band , LH band , HL band , HH band.