Fingerprint Recognition using Daubauchi Wavelet and Radial Basis Function Neural Network

Fingerprint is a unique facility which is present in human anatomy. The ups and downs of the curvature present in the finger among human are different. The curvature present among male and female are also different. In general, the image of a finger either a thumb or index finger is scanned by a compact fingerprint scanner with high resolution. The fingerprint scanned w412ill go through preprocessing followed by wavelet decomposition. This paper implements wavelet decomposition for extracting features of fingerprint images. Subsequently, at the 5th level decomposition, statistical features are computed from the coefficients of approximation and detail. These features are used to train the radial basis function (RBF) neural network for identifying fingerprints. Sample finger prints are taken from database from the internet resource. The fingerprints are decomposed using daubauchi wavelet 1(db1) into 5 levels. The coefficients of approximation at the 5thlevel are used for calculating statistical features. These statistical features are used for training the RBF network.