面向模糊指静脉图像的动态识别方法  

Dynamic Identification Method for Fuzzy Finger Vein Image

作　　者：李凡[1] 沈雷[1] 杨航 吕葛梁 LI Fan;SHEN Lei;YANG Hang;L Geliang(School of Communication Engineering,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China)

机构地区：[1]杭州电子科技大学通信工程学院

出　　处：《杭州电子科技大学学报（自然科学版）》2019年第6期24-29,共6页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：指静脉识别过程中,为了减少低质量模糊图像带来的影响,提出一种面向模糊指静脉图像的动态识别方法。利用图像金字塔模型判断静脉图像模糊程度,根据模糊程度动态调整NiBlack分割的计算系数,并根据不同识别模式来设置相应的动态识别阈值。实验结果表明:0误识率下,识别率从78.00%提升至86.03%,系统的整体识别性能得到明显提升。Low quality images may be collected in the process of finger vein recognition.In order to reduce the influence of fuzzy images on finger vein recognition system,this paper proposes a dynamic identification method for fuzzy digital vein image,which uses the image pyramid model to judge the degree of vein image blur and dynamically adjusts the calculation coefficient of NiBlack segmentation according to the degree of blur.The corresponding dynamic recognition threshold is then set based on the recognition patterns.The experimental results show that under the zero false accept rate,the recognition rate increases from 78.00%to 86.03%,and the overall recognition performance of the system is significantly improved.

关 键 词：手指静脉 模糊分类 图像金字塔 NiBlack分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]