International Journal on Advanced Science, Engineering and Information Technology, Vol. 8 (2018) No. 5, pages: 1941-1947, DOI:10.18517/ijaseit.8.5.6239

A New Approach for Fingerprint Authentication in Biometric Systems Using BRISK Algorithm

Elaf J. Al Taee, Zainalabideen Abdulsamad

Abstract

Now a day, Authentication process in biometric system become most critical task with the expansive of individual information in the world. Where in many current applications, devices and commercial treatments required fingerprint identification process in order to verify the requested services. Most technologies also motivate to this direction. With the increasing of fingerprints uses, there is a need to provide a technique that able to handle the issues that exist in fingerprint acquisition and verification processes. Typically, fingerprint authenticated based on pick small amount of information from some points called Minutiae points. This approach suffers from many issues and provide poor results when the samples of fingerprints are degraded (scale, illumination, direction) changes. However, BRISK algorithm used to handle the previous issues and to extract the significant information from corner points in fingerprint. BRISK is invariant to scale, illumination, and direction changes and its able to pick large number of information when compared with minutiae points. In this paper, BRISK algorithm used based on image based approach, where current recognition matrices are developed and proposed new metrics without need for human interaction. UPEK dataset used to test the performance of proposed system, where the results show high accuracy rate in this dataset. Proposed system evaluated using FAR, FRR, EER and Accuracy and based on selected metrics the proposed system and methodology achieve high accuracy rate than others, and gives a novel modification in authentication task in biometric systems

Keywords:

fingerprint matching; fingerprint retrieving; UPEK; BRISK algorithm; FAST detector; biometric system.

Viewed: 82 times (since Sept 4, 2017)

cite this paper     download