基于信息熵的局部二值模式手指静脉识别  

Finger Vein Recognition Based on Local Binary Pattern Based on Information Entropy

作　　者：黄艳国 杨训根 周满国 HUANG Yan-guo;YANG Xun-gen;ZHOU Man-guo(School of Electrical Engineering and Automation, Jiangxi University of Science and Technology, Ganzhou 341000, China)

机构地区：[1]江西理工大学电气工程与自动化学院,赣州341000

出　　处：《科学技术与工程》2022年第14期5723-5728,共6页Science Technology and Engineering

基　　金：国家自然科学基金(72061016);江西省教育厅科技项目(GJJ160608);留学基金委资助项目(201908360225)。

摘　　要：为进一步提升手指静脉识别算法的识别率与识别速度,在图像处理阶段,提取出手指图像的感兴趣区域(region of interest,ROI),减少手指周围区域的干扰。为提升识别率,在局部二值模式(local binary patterns,LBP)的基础上,引入像素邻域之间的关系,增强LBP的识别性能;然后将信息熵与局部平均二值模式(local average local binary patterns,LALBP)结合得到熵值加权的局部二值模式(entropy weighted local binary patterns,ELBP)特征;最后采用主成分分析(principal component analysis,PCA)降维,以减少识别时间,去除冗余特征。通过对比欧氏距离与曼哈顿距离构建的分类器,与其他主流特征比较,验证算法的识别性能。在SDUMLA数据库与天津市智能实验室采集指静脉图像数据库上,保证了算法的识别速度前提下,分别取得了99.53%和99.84%的识别率,与其他识别算法相比识别率有明显的提高。In order to further improve the recognition rate and recognition speed of the finger vein recognition algorithm,in the image processing stage,the region of interest(ROI)of the finger image was extracted to reduce the interference of the area around the finger.For improve the recognition rate,on the basis of local binary pattern(LBP),the relationship between pixel neighborhoods was introduced to enhance the recognition performance of LBP.Then the information entropy is combined with the local average binary pattern(LALBP)to obtain the entropy weighted local binary pattern feature(ELBP).Finally,principal component analysis(PCA)was used to reduce the dimensionality to reduce the recognition time and remove redundant features.The classifier constructed by comparing Euclidean distance and Manhattan distance was compared with other mainstream features to verify the recognition performance of the algorithm.On the SDUMLA database and the finger vein image database collected by the Tianjin Intelligent Laboratory,the recognition rate of 99.53%and 99.84%are achieve on the premise of ensuring the recognition speed of the algorithm.Compared with others recognition algorithm,the recognition rate has been significantly improved.

关 键 词：手指静脉识别 感兴趣区域 信息熵 局部二值模式 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]