基于基元模式信息统计的虹膜卷缩轮提取  

Iris Collarette Extraction Based on Primitives Pattern Statistics

作　　者：黄静[1] 苑玮琦[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《计算机辅助设计与图形学学报》2014年第8期1326-1332,共7页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61271365)

摘　　要：人眼虹膜表面纹理特征变化能够体现身体健康状况,为了利用虹膜卷缩轮的变化间接反映人体肠部功能,提出一种基于模糊纹理统计的虹膜卷缩轮提取算法.首先利用纹理基元统计出纹理变化规律;然后定义边界模式与非边界模式,在窗口内统计这2种模式出现的频数;最后通过频数的变化规律找出包含卷缩轮的窗口,以此提取卷缩轮.实验结果表明,该算法可以有效地提取具有模糊纹理的虹膜卷缩轮,为基于虹膜的计算机辅助诊断提供有利的依据.Change of human iris texture may reflect the physical conditions.Iris collarette change indirectly reflects the human bowel function.This paper presents a novel approach to analyze the collarette information based on fuzzy textures and statistics,using texture primitive to count rhythmicity of texture variation,then counting the frequency modes in the windows by defining the boundary and non-boundary mode,and finally identifying the specified window which contains the collarette round to extract collarette by frequency variation.Experimental results indicate that extraction with fuzzy textures which can effectively turn the iris collarette,provide a favorable basis for iris-based computer-aided diagnosis.

关 键 词：虹膜卷缩轮 模糊纹理 纹理基元 计算机辅助诊断 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]