Abstract
The accurate segmentation of retinal blood vessels in fundus is of great practical significance to help doctors diagnose fundus diseases. Aiming to solve the problems of serious segmentation errors and low accuracy in traditional retinal segmentation, a scheme based on the combination of U-Net and Dense-Net was proposed. Firstly, the vascular feature information was enhanced by fusion limited contrast histogram equalization, median filtering, data normalization and multi-scale morphological transformation, and the artifact was corrected by adaptive gamma correction. Secondly, the randomly extracted image blocks are used as training data to increase the data and improve the generalization ability. Thirdly, stochastic gradient descent was used to optimize the Dice loss function to improve the segmentation accuracy. Finally, the Dense-U-net model was used for segmentation. The specificity, accuracy, sensitivity and AUC of this algorithm are 0.9896, 0.9698, 0.7931, 0.8946 and 0.9738, respectively. The proposed method improves the segmentation accuracy of vessels and the segmentation of small vessels.
Highlights
The results indicate that the accuracy, specificity and positive predictive value of the existing algorithms for vessel segmentation on the DRIVE dataset are lower than those of the algorithm in this paper, indicating that the algorithm in this paper has a better comprehensive segmentation performance on this dataset
In order to solve problems of low segmentation accuracy and the incomplete segmentation of small vessels, this paper proposes a network based on the combination of U-Net and Dense-Net
Extracted image blocks were used as training data, Dense-U-Net was used as training network model, random gradient descent was used to optimize Dice loss function, and random transformation was used to expand training data and improve the generalization ability
Summary
The retinal vascular system provides rich information about the state of the eye and is the only non-invasive imaging method to obtain visible blood vessels from the human body. SAFFARZADEH et al [10] proposed a new method of retinal image vascular segmentation based on multi-scale line operator and K-means clustering. Traditional retinal vascular segmentation algorithms cannot accudeep learning-based vascularHowever, segmentation methods have been proposed. KOWSKI et al [17] proposed a supervised segmentation technique that uses a deep neural number learning-based vascular segmentation methods have been proposed [15]. [20] regarded retinal vascular segmentation as a boundary detection task, used multi-scale multi-scale features refine blocks (MSFRB), and an attention mechanism (AM). Context information and a side output layer in the network to learn the rich hierarchical regarded retinal vascular segmentation as a boundary detection task, used multi-scale structure, and used conditional random fields to model thethelong-term context information and a side output layer in network todependence learn the richbehierarchical tween pixels.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
