基于图像类型与关键点生成模式组合的手背静脉识别研究  

Research on Hand Vein Recognition Based on Combination of Image Type and Key Point Generation Mode

作　　者：王一丁 康玲 郑璇 WANG Yiding;KANG Ling;ZHENG Xuan(Col.of Information,North China Univ.of Tech.,100144,Beijing,China)

机构地区：[1]北方工业大学信息学院

出　　处：《北方工业大学学报》2019年第5期57-61,76,共6页Journal of North China University of Technology

基　　金：国家自然科学基金项目“弱约束条件下的手背静脉身份识别研究”(61673021)

摘　　要：目前在手背静脉识别中,对最佳图像类型与关键点生成模式组合没有做大量的研究,因此本文针对手背静脉图像纹理的结构特点,提出了图像类型与关键点生成模式多重组合的手背静脉识别方法.通过调整各关键点生成模式中的相关参数,得出每种图像类型分别与3种关键点生成模式组合下的识别率,并画出识别率曲线.经过实验对比,得到最佳组合即二值分割图与DOG关键点检测子模式的组合具有最高的识别率,同时手背静脉识别具有更好的鲁棒性.Researches on the best image type and key point generation mode in the hand vein identifi-cation are inadequate.In this paper,a hand-reverse vein recognition method combining image type with key point generation mode is proposed based on the structural characteristics of hand vein image texture.This paper adjusts the correlation parameters in each key point generation mode to obtain the recognition rate of each image type combined with the three key point generation modes and draws the recognition rate curve.The experimental combination of the best combination,that is,the combination of the binary segmentation map with the DOG key point detection sub-mode has the highest recognition rate and the hand vein recognition has better robustness.

关 键 词：手背静脉识别 DOG关键点检测子 图像类型 关键点 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]