基于特征信息和灰度信息的眼底图像配准  

Fundus Image Registration Based on Feature Information and Gray Information

作　　者：张志顺[1] ZHANG Zhi-shun(School of Medicine,Jiaying University,Meizhou 514031)

机构地区：[1]嘉应学院医学院,梅州514031

出　　处：《现代计算机》2020年第31期54-57,共4页Modern Computer

基　　金：广东省医学科学技术研究基金(No.B2015148);梅州市科技计划项目(No.2014B42)。

摘　　要：针对眼底图像配准存在配准精度不高、速度不快的问题,提出一种新的眼底图像配准算法。用血管中心线进行特征结点检测,用多分辨分层匹配算法实现平移变换,用最大互信息法实现眼底图像精确匹配。实验结果表明,采用灰度信息和特征信息相结合的方法进行配准,精度较高、速度较快,稳定性和适用性更好。Aiming at the problems of low accuracy and low speed in fundus image registration,a new fundus image registration algorithm is proposed.The blood vessel centerline is used to detect the feature nodes,the multi-resolution hierarchical matching algorithm is used to realize the translation transformation,and the maximum mutual information method is used to realize the accurate matching of fundus images.The ex⁃perimental results show that the method of combining gray information and feature information has higher accuracy,faster speed,better stability and applicability.

关 键 词：特征信息 灰度信息 配准、检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R770.4[自动化与计算机技术—计算机科学与技术]