A Reproducible Performance Evaluation Method for Forged Fingerprint Detection Algorithm

As biometric technologies such as fingerprint recognition technology become more widespread, the danger posed by security attacks using forged fingerprints is increasing. To counter this threat, diverse technologies are being developed to detect fingerprint forgery. However, thus far there has no method of evaluating the performance of forged fingerprint detection algorithms that has been confirmed to be reliable by achieving reproducible performance test results. This paper proposes a method for objective performance evaluation of forged fingerprint technology. Unlike previous methods of evaluating the performance of forged fingerprint detection algorithms, the proposed performance evaluation method can reproduce evaluation results, and can be used for technology evaluation. To eliminate the ambiguity of the method and express it in a formal manner, the Evaluation Model of Forged Fingerprint Detection (EMFFD) has been developed. In addition, various metrics have been defined to measure the performance evaluation of forged fingerprints.