The Performance Of User Verification Using Two Fingerprint Based On Error Rate

Biometric technology, especially fingerprint, attract users around the world to use it to secure their places or properties. On the current implementation of fingerprint based person verification, the system face several problems such as noisy finger and fingerprint misplacement by the user. When the problem only effected a small part of fingerprint, it is solved or minimized on the image processing and pattern recognition phase. But, when the problem involves a larger part of fingerprint, another approach needs to be used. To solve or minimize this kind of problem, the approach uses two fingerprints on the verification process have been experimented on this research. By not only referring to the thumbs which are usually used, the potential of the other fingerprint have been studied to find the two best fingerprints as used in this process. Using several simple methods to classify the decisions from both fingerprints; the best classifier have been used to study the performance level of the system compared to the current system used in single fingerprint. In this study, total error rate has been used as an indicator to the performance level of the system. Although the problems have not been totally solved, but the total error rates for fingerprint verification system has been minimized by using this approach.