眼底图像的双向扫描中央凹定位方法

Fovea Localization Based on Bidirectional Scanning for Fundus Images

中央凹定位是计算机辅助眼底疾病诊断的关键步骤。为提高中央凹定位的准确性,提出双向扫描定位方法。在检测出血管和视盘的基础上,以视盘中心为起点的纵向扫描线从上向下扫描,根据扫描线穿过血管像素最少估计中央凹的纵坐标;根据所估计的纵坐标确定一条带状区域,并作通过垂直于带状区域的从左向右扫描的横向扫描线确定横坐标。该方法利用中央凹低亮度、无血管的区域特征来降低计算复杂度,并通过增强对比度来避免不均匀光照产生的干扰。在MESSIDOR数据集和KAGGLE数据集上的实验结果表明,该方法具有较好的定位效果,平均高斯相似度分别为0.980和0.997。

Fovea localization is a key step in computer-aided diagnosis of retinopathy.In order to improve the accuracy of fovea localization,a bidirectional scanning method is proposed in this paper.After the detection of blood vessels and optic disc,a longitudinal scan line starting from the center of the optic disc is made to scan the image from top to bottom,and the vertical coordinate of the fovea is estimated according to the minimum pixel obtained by the scan line passing through the blood vessels.A strip is determined according to the estimated vertical coordinate,and the horizontal coordinate is determined by a transverse scan line perpendicular to the strip and scanning the image from left to right.The proposed method takes advantage of the fact that the fovea is the low-brightness vascular-free area in the fundus image to reduce the computational complexity,and avoids the interference caused by uneven illumination by enhancing the contrast.The experimental results on MESSIDOR data set and KAGGLE data set show that the proposed method has good localization effect,and the average Gaussian similarity is 0.980 and 0.997,respectively.