Touchless Fingerprint Recognition based on Hierarchical Clustering
Rudresh Dwivedi, Varun Pandya, Parthiv Shah · 2021
Biometric verification and identification are being extensively used in real life applications as an access control mechanism. Among the various biometric verification techniques, fingerprint based recognition system is one the most commonly used practice. Since, the attacker may challenge the biometric security from touch based verification by illicit acquisition of user’s fingerprint from sensor. The proposed method uses agglomerative hierarchical clustering which utilizes the formation of clusters in the pre-processed fingerprint samples. Thereafter, cluster-based verification is performed with genuine and non-genuine samples by comparing the intra-cluster distances. Experiments which are carried out on the PolyU Contactless Fingerprint database confirm the efficacy of the proposed method. Though, the GAR of proposed approach is confirmed as 85.42% for the database. However, the proposed approach is comparable to the existing methods considering the security and privacy perspectives.