摘要
虹膜图像的定位是整个虹膜识别过程中最重要的环节,定位的速度和准确性决定了整个虹膜识别系统的方案是否实用可行。论文提出的算法从建立虹膜图像的灰度直方图入手;根据图像中虹膜内外边界区域灰度分布的特点,采用不同的滤波方式;分析虹膜边界的灰度域值,利用可能的边界点,依靠投票机制完成虹膜的定位;从不同角度提出了新的思路。
Iris localization is always the most kernel procession in an iris recognition system.The speed and the accu-rateness of the iris localization decide whether the whole system is possible and practical.There are some new ideas in this paper.Beginning with the construction of the histogram of monochram image,using different smoothing-filters accord-ing to different monochram location of the boundary area in an iris image,analyzing the regional value of the iris boundary and accomplishing the localization depending on the polling -system by computing the possible boundary points.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第30期54-56,共3页
Computer Engineering and Applications
基金
国家自然科学基金支持(编号:90104030)
关键词
虹膜定位
灰度直方图
投票机制
Iris Localization,Histogram of Monochram,Polling-System
