An excellent system in palmprint recognition

Abstract

According to Federal Bureau of Investigation (FBI) of United States, more than 30% of record of evidence in criminal cases are derived from the palmprint, instead of fingerprint. For that reason, the research on palmprint biometric continues to be developed. One of the effort to improve the system is to implement the right algorithms for all stages of biometric research. An excellent system in this paper consists of four powerful algorithms to identify and verify a person through palm of the hand particularly increasing the ROI display image with wavelet, the selection of Gabor parameters, the use of appropriate dimension reduction, and the choice of proper matching method. According to research findings, the method is considered reliable to estimate the error rate with value 0.273% and success level of verification 99.727%.