A Robust Fingerprint Recognition Technique Applying Minutiae Extractors and Neural Network

Abstract

Fingerprint recognition is one of the interesting and mature biometric technique for personal identification or authentication application. The objective of the work is to retrieve personal data by using the fingerprint. In this paper, the fingerprint of a person is captured via Digital Persona U.are.U 4500 fingerprint scanner. The input fingerprint is firstly enhanced using contrast stretching and two morphological operations (dilation and erosion). Next, 160x160 size of the center region of the fingerprint is cropped and extracted minutiae such as ridges and bifurcations. These minutiae are used as the features of a particular fingerprint. These features are put into the Neural Network for person recognition. Various experiments will confirm and prove the accuracy and performance of the proposed fingerprint recognition system