视网膜血管直径交互式测量方法
Interactive measurement of retinal vessel diameter

以DRIVE和STARE公开眼底库以及临床采集的眼底图像为实验对象，引入盒式滤波优化算法进行眼底图像自适应对比度增强，提出一种感兴趣测量区域视网膜动静脉管径的交互式测量方法. 该方法采用2D Gabor滤波提取眼底血管轮廓进行阈值分割和细化提取骨架，对骨架上的分叉点进行连通域判断确定最终血管网络，采用交互式测量方法得到血管的直径. 实验结果表明，本算法在眼底图像增强和眼底血管网络提取等方面效果较好，血管直径的交互式测量结果与人工测量结果相比较为接近，且计算速度较快，误差较低.
The proposed approach is evaluated against DRIVE、STRAVE database and some clinical fundus images in this paper. Adaptive contrast enhanced method combining box filter optimal algorithm is adopted to quicken the computing speed and an interactive method for measuring retinal arteriovenous is presented. The vessel contour is extracted by 2D Gabor filter and the skeleton through the binary image. The vessel network can be obtained through judging the branch and end points’ connectivity. The interactive method is introduced to measure the diameter of the vessel. The experimental results show that the proposed algorithm performs well in the image enhancement and the extracting of the vessel network；the measurement results from the interactive method and artificial method are close