作者: Hailun Liu , Dongmei Sun , Ke Xiong , Zhengding Qiu
DOI: 10.1007/978-3-642-25449-9_29
关键词: Feature selection 、 Point (geometry) 、 Computer science 、 Euclidean distance 、 Scheme (programming language) 、 Pattern recognition 、 Measure (mathematics) 、 Set (abstract data type) 、 Artificial intelligence 、 Feature data 、 Biometrics 、 Data mining
摘要: Fuzzy vault scheme is one of the most popular biometric cryptosystems. However, designed for set differences while Euclidean distance often used in techniques. Multidimensional fuzzy (MDFVS) a modified version that can be easily implemented based on feature data. In MDFVS, every point vector, and measure genuine points filtering. To get better performances, step selection MDFVS algorithms very important should well designed. this paper we propose applying recognition rate to discrimination features selecting strong distinctive into points. Some principles compose are discussed. An implementation with also presented. Experimental results palmprint show proposed approach improves performances MDFVS.