鼻咽胶原双光子显微图像的纹理研究  

Research on the Texture of Two-photon Microscopic Image for Nasopharyngeal Collagen

作　　者：王廷银[1] 陈冠楠[1] 林宏心[1] 陈荣[1] 苏颖[2] 黄祖芳[1] 

机构地区：[1]医学光电科学与技术教育部重点实验室,福建师范大学物理与光电信息科技学院,福建福州350007 [2]福建省肿瘤医院,福建福州350014

出　　处：《福建师范大学学报（自然科学版）》2011年第5期52-56,共5页Journal of Fujian Normal University：Natural Science Edition

基　　金：国家自然科学基金资助项目(60778046);福建省自然科学基金资助项目(2008J0016;2010J05130);卫生部科学研究基金资助项目(WKJ2008-2-046)

摘　　要：研究鼻咽胶原双光子显微图像纹理特征以区分正常胶原与病变胶原.首先采用纹理的边缘和求熵算法来描述它的分布规则性;然后根据三阶矩对纹理亮度直方图进行分析,并采用密度分布和绿色素的占有率来反映纹理的内部结构,且采用梯度向量值提取特征;最后提出4个因素构成的判定公式.通过实验证明该方法效果显著,可应用于医院临床辅助诊断.The textures of nasopharyngeal collagen in two-photon microscopic image are analyzed in this article to distinguish the normal collagen from lesioned ones. There are five steps in the analyzing process： Firstly, the edge of textures and the entropy algorithm are used to describe its distribution rule; Secondly, the analysis is laid on the brightness his- togram according to the third moment ; thirdly, the density and occupation ratio of green pigment are used to reflect the inner structure of textures; Fourthly, the gradient vector value is used for the texture feature extracting; Finally, a judging formula formed by four different factors is put forth in the article. Experiment results show that the method is efficient in distinguishing normal collagen from lesioned ones, and the method can be extended into clinical usage in forms of auxiliary diagnose.

关 键 词：双光子显微图像 鼻咽胶原 纹理特征提取 

分 类 号：R318.51[医药卫生—生物医学工程]