非接触式掌脉图像的点集操作定位法  

Location Method of Touchless Palmvein Image Using Point Set Operation

作　　者：刘刚[1] 黎明[1] 冷璐[1,2] 

机构地区：[1]南昌航空大学无损检测技术教育部重点实验室,江西南昌330063 [2]延世大学电气电子工程学院,韩国首尔120749

出　　处：《计算机仿真》2015年第1期458-462,共5页Computer Simulation

基　　金：国家自然科学基金(61305010;61262019;61202112;61303199);韩国国家研究基金基础科学研究计划项目(2013006574);中国博士后科学基金(2013M531554);南昌航空大学博士启动基金(EA201308058)

摘　　要：非接触式手掌静脉采集系统中存在诸多干扰因素影响感兴趣区域定位的准确性。为了克服干扰,提高认证性能,将"点集操作"引入掌脉图像的预处理中,直接实现针对区域的分析和处理。通过对分割手掌区域点集进行平移、交、并运算,获取手掌指根区域,即指根点集。然后根据约束条件对指根点集元素进行筛选,得到四个指根点的准确位置。根据指根点建立相应坐标系,动态定位感兴趣区域。实验对比证明,上述方法具有更优的抗干扰能力和定位准确性。There are several interference factors in touchless palmvein capture system, which degrade the location accuracy of region of interest （ROI）. In order to overcome the interference factors and improve the verification performance, ＂point set operation＂ is used to preprocess palmvein images, which analyzes and processes the regions directly. The valley regions between fingers, i. e. , the valley point set, are detected through the translation, intersection and union of point set of segmented palm. Four valley points are selected from the valley point set according to constraint conditions. With the help of the four valley points, the corresponding coordinate system is established to dynamically locate ROI. The experimental results confirm that the proposed ＂point set operation＂ has strong capacity of resisting disturbance and high location accuracy.

关 键 词：点集操作 非接触式采集 掌脉认证 感兴趣区域定位 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]