Clonal Selection Algorithm for Low Quality Fingerprint Image Verification

Abstract

Fingerprint verification has drawn a lot of attention to its approach in biometric since it is one of the most important biometric technologies nowadays and is widely used in several different applications and areas. It is applied in the forensic science area in order to identify people who were involved in criminal scenes such as the victims and the suspects. A human’s fingerprint is unique and usually has its own patterns and ridges, which differs them from other’s fingerprints. However, there are some drawbacks that can cause low accuracy and low performance of the verification. This occurs when the fingerprint images used are of low-quality and the fingerprints may be slightly incomplete (partial). Clonal Selection Algorithm (CSA) is known to be good in pattern matching and optimization of problems. Hence, this paper discusses the finding of the implementation of CSA in fingerprint verification. There were two main processes involved, which are features extraction using minutiae-based method and also the implementation of the CSA algorithm. Study shows that the FNMR result is 33.33% and the FMR is 16.67%. Further studies can be carried out by using the same algorithm, but focusing more on the feature extraction methods to improve the extraction of fingerprints.