基于Canny边缘检测和外观特征的微血管瘤检测方法

发布时间：2018-02-28 15:07

  本文关键词： 糖尿病性视网膜病变 微血管瘤 Canny算子 区域填充　出处：《现代电子技术》2017年06期 　论文类型：期刊论文

【摘要】：糖尿病性视网膜病变进行早期筛查可以减少疾病的发展并且阻止随后的视力损害。微血管瘤是糖尿病性视网膜病变的早期临床症状,可以通过微血管瘤检测对糖尿病性视网膜病变进行早期筛查。针对眼底图像中视网膜血管、视盘、渗出物以及微血管瘤之间的相互关系,在红色通道和绿色通道加权图上定位出视盘,在绿色通道上采用基于简单统计的自适应双阈值Canny算子进行边缘检测,并进行封闭区域的填充。设定阈值消除大面积对象并移除视网膜血管、视盘和渗出物得到微血管瘤的候选区域,最后根据形状特征和颜色特征从候选区域中得到真正的视网膜微血管瘤。实验结果表明,该算法能够有效提取视网膜眼底图像中的微血管瘤,敏感性和阳性预测值分别达到92%和86%,优于现有一些典型的微血管瘤检测方法,能够精确地检测出微血管瘤,可用在糖尿病性视网膜病变早期筛查中。
[Abstract]:Early screening for diabetic retinopathy can reduce the progression of the disease and prevent subsequent visual impairment. Microangioma is an early clinical symptom of diabetic retinopathy. Early screening for diabetic retinopathy can be performed by microangioma detection. The relationship between retinal vessels, optic disks, exudates and microangiomas in fundus images, The video disk is located on the weighted map of the red channel and the green channel, and the edge detection is carried out by using the adaptive double-threshold Canny operator based on simple statistics on the green channel. Set a threshold to eliminate large areas of objects and remove retinal blood vessels, optic disks and exudates to obtain candidate areas for microangioma, Finally, the real retinal microangioma is obtained from the candidate regions according to the shape and color features. The experimental results show that the algorithm can effectively extract the microhemangioma from the retinal fundus image. The sensitivity and positive predictive values were 92% and 86, respectively, which were superior to some typical microhemangioma detection methods, and could be used in the early screening of diabetic retinopathy.
【作者单位】： 福建农林大学计算机与信息学院;福州大学物理与信息工程学院;厦门大学信息科学与技术学院;
【基金】：国家自然科学基金项目(60827002) 福建省自然科学基金项目(2012J05116) 福建省教育厅项目(JA13037)
【分类号】：R587.2;R774.1;TP391.41

【相似文献】

中国硕士学位论文全文数据库 前1条

1 李艳丽;基于数学形态学的最大二维信息熵及Canny边缘检测的图像分割的研究[D];信阳师范学院;2011年

，

本文编号：1547817