基于形态特征和k均值聚类的黄斑检测与定位

发布时间：2018-05-07 13:56

本文选题：黄斑检测 + 黄斑定位　；参考：《中国生物医学工程学报》2017年06期

【摘要】：彩色眼底图像已经广泛地应用于眼科相关疾病的辅助诊断和筛查。眼底图像中的黄斑区域检测和中心定位是眼科疾病分级、诊疗的重要步骤。提出一种有效检测与定位黄斑的方法,通过分析黄斑的低亮度和趋于圆形的形态特征,可以不依赖视盘和血管信息,在二值化眼底图像中实现黄斑检测,确定黄斑区域。改进k均值聚类方法,引入图像的空间信息,优化聚类对象,获取黄斑的边缘信息,实现黄斑中心的有效定位。在公开的眼底图像数据库上验证方法的性能,具有较高的准确率。对正常和存在病变的眼底图像的黄斑中心有效定位,可达到96.11%和92.12%,平均准确率达到93.92%。实验表明,提出的基于形态特征和k均值聚类的黄斑检测与定位方法简单、高效,对眼科疾病的计算机辅助诊断有实用价值。
[Abstract]:Color fundus images have been widely used in the diagnosis and screening of ophthalm-related diseases. Macular area detection and central location in fundus images are important steps in classification, diagnosis and treatment of ophthalmic diseases. An effective method to detect and locate macula is presented. By analyzing the low brightness and circular shape of macula, we can detect the macula in binary fundus image without depending on the information of optic disc and blood vessel, and determine the macular area. The k-means clustering method is improved, the spatial information of the image is introduced, the clustering object is optimized, the edge information of the macula is obtained, and the effective location of the macular center is realized. The performance of the method is verified on the image database of the eye fundus, and the accuracy of the method is high. The effective localization of macular center in normal and diseased fundus images can reach 96.11% and 92.12% respectively, with an average accuracy of 93.92%. The experimental results show that the method of macular detection and localization based on morphological features and k-means clustering is simple and efficient and has practical value for computer-aided diagnosis of ophthalmic diseases.
【作者单位】：福州大学物理与信息工程学院;福建省信息处理与智能控制重点实验室(闽江学院);
【基金】：福建省中青年教师教育科研项目(JAT160398) 福建省高校自然基金青年重点项目(JZ160467) 福州市科技计划项目(2016-S-116)
【分类号】：R770.4;TP391.41

【相似文献】