A fingerprint recognition system for cyber security

Abstract

Modern society demands a higher level of accuracy, security and practicality in identification. Traditional identification methods no longer meet these needs, and the rich physiological and behavioural characteristics of the human body offer a reliable solution for this. This has attracted a great deal of attention from the international academic and commercial community. Fingerprint recognition is one of the earliest technologies to use biometric features for identification and is also a very mature identification technology. This project will investigate and construct a fingerprint identification system depending on feature extraction and feature matching, which is the most dominant fingerprint recognition algorithm and technique today. The method mainly involves pre-processing the sample, extracting feature points from the sample fingerprint image, and then matching the fingerprints based on the number of feature pairs of two related fingerprints. The specific implementation of this project includes steps such as graph strength enhancement, binarization, orientation calculation, refinement, Gabor filtering, feature extraction and feature matching. The project was coded using MATLAB, and a final match percentage value was generated to determine if it was the same finger.