Computational Algorithms for Fingerprint Recognition

Abstract

Biometrics, which recognizes a person's identity using his/her physiological or behavioral characteristics, is inherently more reliable and capable than traditional methods. Biometric signs include fingerprint, face, gait, iris, voice, signature, etc. Among them, fingerprint is the one, which has been researched for a long time and shows the most promising future in real-world applications. However, because of the complex distortions among the different impressions of the same finger, fingerprint recognition is still a challenging problem. In this dissertation, our objective is to develop effective and efficient computational algorithms for an automatic fingerprint recognition system. The algorithms we address include: (1) Templates based minutiae extraction algorithm; (2) Triplets of minutiae based fingerprint indexing algorithm; (3) Genetic Algorithm based fingerprint matching algorithm; (4) Genetic Programming based feature learning algorithm for fingerprint classification; (5) Comparison of classification and indexing in identification; and (6) Fundamental performance analysis of fingerprint matching. All the experimental results are demonstrated on standard fingerprint database, NIST-4 fingerprint database. Although the algorithms we have developed can achieve a good performance in fingerprint recognition, we believe that there are still some problems need to be worked on to make automatic fingerprint recognition system more effective and efficient in real-world applications. We believe that it needs incorporation of researchers from different fields, such as Computer Science, Electrical Engineering, Physiology, Statistics, Social Sciences, etc. So that, it is possible to achieve a better fingerprint recognition performance, which is close to theoretical bound.