基于脉冲耦合神经网络结合U-Net的眼底血管分割

Retinal Blood Vessel Segmentation Based on Pulse-Coupled Neural Network Combined with U-Net

视网膜血管分割作为一种现代医学诊断的基础性方法,在眼部疾病及相关病症的筛查和诊断中起着重要作用。针对血管分割分割过程中细节丢失、连通性差等问题,提出一种新的分割算法。该算法基于改进的自适应阈值SSPCNN进行眼底血管图像分割,并应用改进SE模块的密集可变形卷积U-Net进行前期图像增强。为该算法设计实验,检测其在DRIVE和STARE数据集上的实际分割效果检测。实验结果表明该模型在DRIVE和STARE数据集上的灵敏度、特异性和准确度均优于现有算法。

As a basic method of modern medical diagnosis,retinal blood vessel segmentation plays an important role in the screening and diagnosis of eye diseases and related diseases.Aiming at the problems of missing details and poor connectivity in blood vessel segmentation,a new segmentation algorithm is proposed.Based on the improved adaptive threshold SSPCNN,the algorithm segments the fundus blood vessel image,and uses the improved SE module's dense deformable convolution U-Net to enhance the pre-vious image.Experiments for this algorithm is designed to detect its actual segmentation effect on DRIVE and STARE data sets.The experimental results show that the sensitivity,specificity and accuracy of the model are better than the existing algorithms on DRIVE and STARE data sets.