Design of bi̇ometri̇c identification systems

It is of great importance to provide the highest accuracy of fingerprint identification and verification systems, which is one of the largest biometric features with the largest acceptance rate in terms of identity identification in the world.Fingerprint recognition systems are more widely used than any other biometric feature recognition system, from criminal systems, government departments, laboratories, border crossings, etc., wherever identification is required.For this reason, the fingerprint recognition system must be fast and reliable in order to realize the separation of the false and live fingerprints, and provide higher sensitivity. In this study, the performance metrics of the results of the system were evaluated by comparing the false fingerprints produced by latex, wood glue, gelatine and silicon derivatives and live fingerprints and the results were compared with the other systems.Appropriate machine learning methods for comparing fake and live fingerprints were run on fake and live fingerprints in the LivDet2015 dataset and the results obtained were compared and discussed.