Simple Verification of Low-resolution Fingerprint Using Non-Minutiae Feature

Abstract

It is desirable to have a simple, fast and reliable fingerprint verification application. The application can be build by avoiding long and complicated computation. However, most of the verification methods need some complicated preprocessing. Moreover, they need good quality fingerprints. In this paper, we propose a simple method to verify low-resolution fingerprint using Haar-like transformation feature. Fingerprints were captured by a commercial scanner. They were transformed using Haar-like transformation to generate feature vectors. The vector then verified to the reference vectors by using Hamming Distance. Our experiments results show that the proposed method has an accuracy of up to 80 %.