基于形态学与灰度分布的序列虹膜质量评价算法  

Sequence Iris Quality Evaluation Algorithm Based on Morphology and Gray Distribution

作　　者：刘帅 刘元宁[1,3] 朱晓冬[1,3] 刘文滔 程垚松 LIU Shuai;LIU Yuanning;ZHU Xiaodon;LIU Wentao;CHENG Yaosong(Key Laboratory of Symbolic Computation and Knowledge Engineerin g of Ministry of Education, Jilin University,Changchun 130012,China;College of Software,Jilin University,Changchun 130012,China;College of Computer Science and Technology,Jilin University,Changchun 130012,China)

机构地区：[1]吉林大学符号计算与知识工程教育部重点实验室,长春130012 [2]吉林大学软件学院,长春130012 [3]吉林大学计算机科学与技术学院,长春130012

出　　处：《吉林大学学报（理学版）》2018年第5期1156-1162,共7页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:61471181);吉林省自然科学基金(批准号:20140101194JC;20150101056JC)

摘　　要：基于形态学和灰度直方图中的灰度分布,提出一种虹膜质量评价算法.通过清晰度,基于二值形态学的活体检测,灰度直方图灰度分布的区域有效判定以及超界偏移度和斜视离心度的斜眼程度判断,解决了连续采集的序列虹膜质量评价中因判别指标严苛导致的有效虹膜幸存率过低的问题.在排除不合格图像的同时,尽可能地保留可以进行识别的虹膜,以提高有效虹膜幸存率,并且时间消耗相对较短.We proposed an iris quality evaluation algorithm based on morphology and gray distribution of gray histogram.Through the sharpness,live detection based on binary morphology,region validity determination bas ed on gray distribution in the gray histogram,and squint level based on ul tra-boundary deviation degree and strabismus eccentricity degree,it so lved the problem of the low effective iris survival rate caused by stringent discri minant index in continuous collection of sequence iris quality evaluation.While removing the unqualified images,the proposed algorithm preserved iris which could be identified as far as possible,so as to improve the effective iris survival rate,and time consumption was relatively short.

关 键 词：虹膜质量评价 二值形态学 区域有效判定 超界偏移度 斜视离心度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]