Fingerprint recognition system

Journal Title
Journal ISSN
Volume Title
Fingerprint recognition is one of the most popular biometric techniques in personal identification. The widespread use of fingerprint recognition as a biometric is because each fingerprints pattern of ridges and valleys is unique and does not vary with time and age. While there are several algorithms or methods for fingerprint recognition systems, the quest to develop a robust fingerprint recognition system remains a significant research area. One major challenge in designing a system for fingerprint recognition is its ability to perform well on both full and partial fingerprint images. Most fingerprint recognition systems developed so far use minutiae-based algorithms which tend to perform well under full fingerprint but poorly under partial occlusion. In partial occlusion, the minutiae, which are the core points of the fingerprints, get completely distorted. The distortion of minutiae makes minutiae-based algorithms perform poorly. Therefore, this study proposed a novel non-minutiae-based algorithm that adapts the Fisherface and Eigenface method from facial recognition. The proposed algorithm is insensitive to partial occlusion. The Eigenface and Fisherface methods were tested on FVC 2002 Datasets and yielded an accuracy of 86.67% and 90% respectively. These accuracy results indicates that there is a possibility of recognizing fingerprint images using non-minutiae-based algorithms from different domain.
Undergraduate thesis submitted to the Department of Computer Science and Information Systems, Ashesi University, in partial fulfillment of Bachelor of Science degree in Management Information Systems, May 2022