刑侦领域高分辨率掌纹识别技术及快速匹配方法  

High-resolution automatic palmprint identification technology and quick matching method in criminal investigation

作　　者：吴春生[1,2,3] 冯才刚 迟学斌[2] 

机构地区：[1]北京市刑事科学技术研究所,北京100054 [2]中国科学院计算机网络信息中心,北京100090 [3]中国科学院大学,北京100190

出　　处：《中国科学院大学学报（中英文）》2014年第4期555-563,共9页Journal of University of Chinese Academy of Sciences

基　　金：国家863项目(2010AA012301);国家自然科学基金(11271350);北京市科技计划(Z121100000312099)资助

摘　　要：针对公安机关特别是刑侦领域的掌纹自动识别技术需求,讨论基于高分辨率的掌纹细节特征点识别方法必要性.将细节特征点匹配方法作为重点,以Cafis指掌纹系统为例,分析掌纹细节特征点匹配算法的瓶颈,并与传统硬件并行和最新的GPU并行等提高掌纹比对速度的技术方法进行对比.The basic concepts and methods of automatic palmprint recognition technology are described. In view of the requirements of the criminal investigation department,the importance of the palmprint minutia identification algorithm of high resolution is discussed. The matching method of minutia feature point is considered important. Matching algorithm is introduced by taking Cafis system as an example. The algorithm bottlenecks of palmprint feature point recognition are analyzed.These technologies have high speeds compared to the parallol technologies.

关 键 词：刑侦 掌纹识别 高分辨率 细节特征点 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]