A hierarchical heterogeneous ant colony optimization based fingerprint recognition system

Personal identification is crucial to secure data against cyber-attacks. With increasing identity theft, fingerprint recognition systems have a growing importance in enforcing security and reliable identification. Although, most fingerprint recognitions systems use minutiae features for fingerprint matching, they require extensive preprocessing of fingerprints when the image quality is poor. This may introduce false ridge patterns, degrading the performance of the system. Moreover, fingerprint matching over a large database can be inefficient due to high computation time of fingerprint matching algorithms. This demand for fingerprint recognition systems that are fast and reliable. This paper proposes a computationally intelligent fingerprint recognition system that extracts ridge patterns from the fingerprint for matching. Hierarchical Heterogeneous Ant Colony Optimization based Fingerprint Matching (HHACOFM) algorithm has ant agents at different levels in the hierarchy to find a match between the input and stored ridge patterns. The algorithm was evaluated over four databases: a synthetic database generated using SFinGe tool, an internal database, SOCOFing database and FVC2004 database. Experimental results indicate that the proposal achieves high recognition rate compared to the existing approaches. HHACOFM algorithm achieves less EER than the state-of-art approaches. The results were validated using statistical tests. HHACOFM enables parallelism and thus reduces the response time. The proposal is scalable and suitable for real time applications demanding fast fingerprint verification.