糖尿病患者通常会伴有视网膜结构的改变，通过数字眼底照相机可以无创地获得视网膜图像，从而为糖尿病的诊断和监测提供重要信息。目前，不同糖尿病信息管理系统中视网膜图像信息不能很好地访问和分享，而且对糖尿病计算机辅助诊断提供的信息有限。现有的计算机辅助诊断结构化报告(CAD-SR)模型不能很好地表征视网膜图像中的信息。本课题提出了视网膜血管分析需要结合血管解剖拓扑结构的科学假设，并拟在课题组前期居民电子健康档案系统糖尿病信息管理模块和PACS系统的基础上，基于DICOM和WADO标准解决视网膜图像的互操作问题。利用计算机技术自动分析血管拓扑结构并提取相应层次的形态学参数，在现有CAD-SR信息对象定义的基础上，集成拓扑层次和形态学参数的数值信息，更加有效地对视网膜图像进行CAD-SR建模。本课题的开展对于大数据环境下糖尿病信息管理系统中的图像互操作和基于糖尿病信息系统的数据挖掘具有重要意义。

Those patients who have diabetes often accompany with retinal vessel changes, which images could be non-invasively obtained with digital fundus camera, thus providing important information for diabetes diagnosis and monitoring. At present, such images could not be visited and shared conveniently in different diabetes information management systems (DIMS), and the information they provided for diabetes computer aided diagnosis (CAD) is deficient. The current CAD-structure report (CAD-SR) model might not be able to present retinal image information well. Therefore, we hypothesize that the analysis on retinal image should consider the topology of retinal vessels, and aim to study the image interoperability issues in compliance with DICOM and WADO standards based on our previous research on DIMS of electronic health record system and picture archive and communication system (PACS). Besides, we will design algorithms to investigate the vessels' topological structure and obtain the correlated geometric features from those images. Then, we will define information object definition of CAD-SR and implement topology orders and geometric numeral entry to model CAD-SR effectively. This project is of great significance for the image interoperability in DIMS and for the data mining based on which under big data environment.