Fingerprint Matching Using SVM-Based Similarity Measure  

作　　者：SU Fei XIE Xiaohui FENG Jianjiang CAI Anni 

机构地区：[1]School of Telecommunication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China

出　　处：《Chinese Journal of Electronics》2007年第3期459-463,共5页电子学报（英文版）

基　　金：This work is supported by the National Natural Science Foundation of China （No.60472069）.

摘　　要：A new fingerprint matching method combining the minutiae and ridges and a novel Support vector machine （SVM）-based similarity measure is introduced in this paper. Through combining two distinct sets of fingerprint information, minutiae and ridge image, a better matching performance is obtained. By constructing feature vector, a revised SVM-based learning process is adopted to compute the similarity between the query and template fingerprint images. Experimental results show that the proposed method improves the matching performance dramatically.

关 键 词：Fingerprint matching MINUTIAE Ridge images Support vector machine （SVM） Similarity measure 

分 类 号：TN99[电子电信—信号与信息处理]