免散瞳眼底图像在糖尿病视网膜病变自动筛查中的应用  

Research of Non-dilated Fundus Images for Automated Diabetic Retinopathy Screening

作　　者：高玮玮[1] 程武山[1] 沈建新[2] 左晶[3] 张爱华[1] 

机构地区：[1]上海工程技术大学机械工程学院,上海201620 [2]南京航空航天大学机电学院,江苏南京210016 [3]江苏省中医院眼科中心,江苏南京210029

出　　处：《激光生物学报》2015年第4期335-340,381,共7页Acta Laser Biology Sinica

基　　金：中央高校基本科研业务费专项(NJ20120007);江苏省普通高校研究生科研创新计划项目(CXLX11_0218)资助

摘　　要：为实现糖尿病视网膜病变(糖网)的自动筛查,建立了基于免散瞳眼底图像的糖网自动筛查方法。该方法包括视盘定位及提取、糖网白色病灶(硬性渗出、棉绒斑)自动检测以及微动脉瘤与视网膜内出血的自动检测。在此基础上设计并实现了基于免散瞳眼底图像的糖网自动筛查系统。利用已实现的系统对临床采集的7 687个样本共15 374幅眼底图像进行糖网自动筛查,对样本个体的检测结果为:灵敏度96.46%,特异性96.07%,平均处理时间57.87 s。测试结果表明,所构建的基于免散瞳眼底图像的糖网自动筛查系统满足英国糖尿病协会提出的糖网自动筛查标准(最低灵敏度80%,最低特异性95%)。In order to realize automatic screening for diabetic retinopathy（ DR）,an automated diabetic retinopathy screening system based on non-dilated fundus images is proposed and studied. Firstly,the composition and key technology of the system including locating and extracting optic disc and automated detection of hard exudates,cotton wool spots,hemorrhages as well as microaneurysms were analyzed and reviewed. Then the system based on non-dilated fundus images was developed and tested on an image dataset collected in clinic which include 7 687 samples in total of15 374 fundus images. With an exam-based criterion,sensitivity of 96. 46%,specificity of 96. 07% are achieved. Furthermore,the average time cost in processing a sample is 57. 87 seconds. Results suggest that the performance of DR screening system meet the minimum standard of 80% sensitivity and 95% specificity for the detection of sight-threatening DR which was recommended by British Diabetic Association guidelines.

关 键 词：免散瞳眼底图像 糖尿病视网膜病变 自动检测 自动筛查 图像分割 

分 类 号：R318[医药卫生—生物医学工程] TP391.41[医药卫生—基础医学]