Fingerprint Recognition using Texture Features

This paper proposes an efficient scheme of fingerprint recognition for biometric identification of individuals. Three statistical features are extracted from fingerprint images and represented using a mathematical model. These features are (1) an entropy coefficient, computed from the intensity histogram of the image, (2) a correlation coefficient, computed by correlation operation between the original image and a filtered version of the image obtained using a 2D median filter, and (3) an energy coefficient, obtained by first subjecting the image to a 5-level wavelet decomposition and thereafter computing the percentage energy of the approximation coefficient obtained after the 5th level decomposition. The approach is tested over a dataset of 80 images divided into 10 classes and is seen to provide accurate recognition results.