A fingerprint verification using coarse-to-fine strategy

Abstract

With the development of communication technology, we need a user authentication system that can be applied in a faceless environment. Therefore, there has been a lot of research on fingerprint verification systems with emphasis on security and convenience and their application in a user authentication system. Minutiae-based fingerprint verification has been widely used because of its small system and suitability for real-time processing. However, there is a problem with verification rate decrease if the quality of fingerprint image input reduces or the number of minutiae decreases. Therefore, this paper proposed the fingerprint verification method using fingerprint image. To compare the fingerprint image effectively, Principal Component Analysis (PCA) technique which is actively used in a face verification system was used. To apply PCA to fingerprint verification, the fingerprint verification was performed after suggesting the alignment method using geometric hashing and Region of Interest (Rol) extraction method. Also, to reduce run-time, we used coarse-to-fine strategy. We extracted three alignment data using minutiae matching first and performed PCA. From the result of experiment, we could see that Equal Error Rate (EER) was improved from 3.01% to 2.11%.