Fast match-on-card technique using in-matcher clustering with ISO minutia template by Tai-Pang Chen; Wei-Yun Yau; Xudong Jiang International Journal of Biometrics (IJBM), Vol. 7, No. 2, 2015
Abstract:
Fingerprint match-on-card is receiving more attention from government and the IT industry as it provides higher level of security. In addition, it also has less privacy concern as the enrolled template does not leave the card. To address the interoperability needs for match-on-card implementation, the ISO standard ISO/IEC19794-2, defines the finger minutiae data format that contains only the basic information for matching. However, with such limited basic information, fingerprint matching is a challenge for match-on-card with limited processing power, especially when dealing with deformation and common correspondence for alignment. In this paper, a novel in-matcher clustering method is proposed to search for the matched clusters of minutiae with the least deformation error. The proposed clustering method solves the deformation and alignment problems. However, matched minutia clusters may contain false or missing minutiae which may cause mismatch in some regions between the genuine user and imposter. Such mismatch may result in false acceptance. To alleviate this problem, a further matching step using Mahalanobis distance to measure the inter-cluster similarity is proposed to remove the wrongly matched clusters. Finally, the overall match score is generated by combining the minutia matching (local matching) in group (cluster) and the matching of the geometrical structure between groups (global matching). The proposed algorithm achieved an average EER