首页|基于ReSE-AG-Net的青光眼视盘视杯分割方法

基于ReSE-AG-Net的青光眼视盘视杯分割方法

扫码查看
分割视盘视杯是诊断青光眼的主要依据,因此提出了一种基于改进U-Net网络的青光眼视盘视杯分割方法,称为ReSE-AG-Net.并将ReSE-AG-Net、U-Net、M-Net、CE-Net和CDED-Net分别在Drishti-GS1数据集和RIM-ONE-R3数据集上进行测试,对得到分割后的视盘视杯的Dice系数和IoU进行对比.实验结果表明,在Drishti-GS1数据集上,分割视盘的Dice系数和IoU分别是0.9891和0.9590,分割视杯的Dice系数和IoU分别是0.9310和0.8710,在RIM-ONE-R3数据集上,分割视盘的Dice系数和IoU分别是0.9789和0.9587,分割视杯的Dice系数和IoU分别是0.8637和0.7601,对比U-Net模型和现有的分割模型均有一定的提升,证明了该方法的可行性.
ReSE-AG-Net Based Optic Disk Optic Cup Segmentation for Glaucoma
Segmentation of optic disk optic cup is the main basis for diagnosing glaucoma. Therefore a glaucoma optic disk optic cup segmentation method based on improved U-Net network called ReSE-AG-Net is proposed. ReSE-AG-Net,U-Net,M-Net,CE-Net and CDED-Net are tested on Drishti-GS1 dataset and RIM-ONE-R3 dataset respectively,and the segmented optic disc optic cups obtained are tested and compared. Dice coefficients and IoU of optic disk optic cups are compared. The experimental results show that the Dice coefficient and IoU of the segmented optic disk are 0.9891 and 0.9590,and the Dice coefficient and IoU of the segmented videocup are 0.9310 and 0.8710 on Drishti-GS1 dataset,and the Dice coefficient and IoU of the segmented optic disk on RIM-ONE-R3 dataset are 0.9789 and 0.9587,and the Dice coefficient and IoU for segmenting the optic cup are 0.8637 and 0.7601,respectively,which compare the U-Net model and the existing segmentation models,proving the feasibility of the method.

glaucomasegmentation of optic disk optic cupU-NetReSE-AG-Net

李豪杰、郭子洋、刘伟志、罗益、曹秒

展开 >

长春理工大学 生命科学与技术学院,长春 130022

国网吉林电力有限公司超高压公司,长春 130022

青光眼 视盘视杯分割 U-Net ReSE-AG-Net

吉林省科技厅项目

20240401073YY

2024

长春理工大学学报(自然科学版)
长春理工大学

长春理工大学学报(自然科学版)

CSTPCD
影响因子:0.432
ISSN:1672-9870
年,卷(期):2024.47(4)