ProNet Adaptive Retinal Vessel Segmentation Algorithm Based on Improved UperNet Network

作     者：Sijia Zhu Pinxiu Wang Ke Shen 

作者机构：CW Chu CollegeJiangsu Normal UniversityXuzhou221000China Biomedical Engineering CollegeSouthern Medical UniversityGuangzhou510080China 

出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)

年 卷 期：2024年第78卷第1期

页      面：283-302页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 1002[医学-临床医学] 081104[工学-模式识别与智能系统] 08[工学] 100212[医学-眼科学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 10[医学] 

基　　金：Children’s Heart and Health Research Centre 

主　　题：Retinal segmentation multifaceted optimization cross-fusion data enhancement focal loss 

摘      要：This paper proposes a new network structure,namely the ProNet *** medical image segmentation can help clinical diagnosis of related eye diseases and is essential for subsequent rational *** baseline model of the ProNet network is UperNet(Unified perceptual parsing Network),and the backbone network is ConvNext(Convolutional Network).A network structure based on depth-separable convolution and 1×1 convolution is used,which has good performance and *** further optimise ProNet mainly in two *** is data enhancement using increased noise and slight angle rotation,which can significantly increase the diversity of data and help the model better learn the patterns and features of the data and improve the model’s ***,it can effectively expand the training data set,reduce the influence of noise and abnormal data in the data set on the model,and improve the accuracy and reliability of the *** is the loss function aspect,and we finally use the focal loss *** focal loss function is well suited for complex tasks such as object *** function will penalise the loss carried by samples that the model misclassifies,thus enabling better training of the model to avoid these errors while solving the category imbalance problem as a way to improve image segmentation density and segmentation *** the experimental results,the evaluation metrics mIoU(mean Intersection over Union)enhanced by 4.47%,and mDice enhanced by 2.92% compared to the baseline *** generalization effects and more accurate image segmentation are achieved.