基于眼前房角OCT影像质量分级的原发性闭角型青光眼辅助自动诊断研究

项目名称： 基于眼前房角OCT影像质量分级的原发性闭角型青光眼辅助自动诊断研究

项目编号： No.61501154

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 武薇

作者单位： 杭州电子科技大学

项目金额： 19万元

中文摘要： 基于医学图像的青光眼诊断在临床中有着广泛应用。为解决传统方法存在效率低下以及主观性影响较大等问题，本项目提出了基于图像质量评价的前房角光学断层扫描(OCT)影像中原发性闭角型青光眼分级自动诊断方法。主要研究内容包括：①将图像质量评价统计模型和人类视觉感知特性相结合，研究针对OCT图像质量的客观无参型评价方法，解决传统评价方法与主观评价间存在不一致的问题；②基于Schwalbe线自动检测的房角相关参数自动测量算法，以提高测量效率，保证测量结果的客观性和稳定性，以及对房角分级的准确性；③基于局部的点、线、面三种特征的房角分类方法，以确保在质量较差的图像中，也能对房角分类，为房角状态的评估提供新方法。通过本项目的研究，有望提高前房角关键参数的测量精度和速度，实现对房角的准确客观评级，为原发性闭角型青光眼的排查与诊治提供可供参考的客观评价准则。

中文关键词： 计算机辅助诊断；光学相干断层成像；原发性闭角型青光眼；图像质量评估；特征提取与分类

英文摘要： Glaucoma diagnosis based on medical images has been applied in clinic widely. Traditional AOD measurements from optical coherence tomography (OCT) are based on manual or semi-automatic methods, which are time consuming and also operator-dependent. To solve these problems, we proposed a hierarchical method used for primary angle closure glaucoma diagnosis in anterior chamber angle OCT images, which is based on image quality assessment. This project will focus in three main researches. Firstly, to resolve the problem of inconsistency between the objective and subjective image quality assessment in OCT image, a new objective non-reference OCT image quality assessment algorithm will be researched. It combines together on the basis of statistical model and human visual perception characteristics. Secondly, an advanced automatic measurement method of angle parameters based on Schwalbe’s line is presented. It can improve the measurement efficiency; guarantee the objectivity and stability of the measurement results, and the accuracy of the angle grading. Thirdly，an angle grading method is researched, which is based on the features in local point, local line and local region. It can grade the angle accurately, even in the image with poor quality. This project can improve the measurement accuracy and speed, and achieve the accurate grading for the angle. Also the angle grading performance can be improved even in the image with low signal-to-noise ratio.

英文关键词： computer-aided diagnosis;optical coherence tomography;primary angle-closure glaucoma;image quality assessment;feature extraction and classification