贵州医科大学学报

2020, v.45;No.235(04) 432-437

[打印本页] [关闭]
本期目录(Current Issue) | 过刊浏览(Past Issue) | 高级检索(Advanced Search)

基于深度学习的视网膜病变眼底图视盘自动定位与分割研究
Automated Localization and Segmentation of Retinopathy Fundus Map Using Deep Learning

杨帆;陈睿诗;莫阳;缪月红;段平;谷浩;王丽会;
YANG Fan;CHEN Ruishi;MO Yang;MIAO Yuehong;DUAN Ping;GU Hao;WANG Lihui;Key Laboratory of Biology and Medical Engineering,Guizhou Medical University;Department of Biomedical Engineering,School of Biology and Engineering,Guizhou Medical University;Department of Ophthalmology,the Affiliated Hospital of Guizhou Medical University;Key Laboratory of Intelligent Medical Image Analysis and Precise Diagnosis of Guizhou Province,School of Computer Science and Technology,Guizhou University;

摘要(Abstract):

目的:研究视网膜病变眼底图像中视盘的自动定位与分割。方法:收集Messidor数据库中1 200张眼底图,其中900张用于训练,300张用于测试;采用2个深度学习网络分别实现视盘在图像上的精确定位与分割,采用准确率、Dice系数、Jaccard系数、敏感性、特异性指标评估结果。结果:视盘定位准确率为100%,视盘分割准确率为97. 38%,平均Dice系数、Jaccard系数、敏感性、特异性为0. 919 4、0. 854 9、0. 995 9、0. 970 0,平均处理时间为0. 25 s。结论:深度学习方法能够实现高精度的视盘定位及视盘分割,有助于眼底图像的实时自动分析。
Objective: To investigate the automatic positioning and segmentation of the optic disc in the fundus image of retinopathy. Methods: We collected 1,200 fundus images from the Messidor database,of which 900 image were used for training and 300 images were used for testing. Two deep learning networks were used to achieve precise positioning and segmentation of the video disc on the image. Accuracy,Dice coefficient,Jaccard coefficient,sensitivity and specificity index were used to evaluate results. Results: The accuracy rate of optic disc positioning reached 100%,and the accuracy rate of optic disc segmentation was 97. 38%. The average Dice coefficient,Jaccard coefficient,sensitivity,and specificity were 0. 919 4,0. 854 9,0. 995 9,and 0. 970 0,respectively,and the average processing time was 0. 25 s. Conclusion: Deep learning can realize high-precision optic disc positioning and optic disc segmentation,which is helpful for real-time automatic analysis of fundus images.

关键词(KeyWords): 视网膜病变眼底图;视盘定位;视盘分割;深度学习,U-Net
fundus map of retinopathy;optic disc localization;optic disc segmentation;deep learning,U-Net

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金(81660298,61661010);; 2017贵州省大学生创新创业训练计划资助项目(201710660033);; 贵州省科学技术基金[黔科合J字(2015)2044号]

作者(Author): 杨帆;陈睿诗;莫阳;缪月红;段平;谷浩;王丽会;
YANG Fan;CHEN Ruishi;MO Yang;MIAO Yuehong;DUAN Ping;GU Hao;WANG Lihui;Key Laboratory of Biology and Medical Engineering,Guizhou Medical University;Department of Biomedical Engineering,School of Biology and Engineering,Guizhou Medical University;Department of Ophthalmology,the Affiliated Hospital of Guizhou Medical University;Key Laboratory of Intelligent Medical Image Analysis and Precise Diagnosis of Guizhou Province,School of Computer Science and Technology,Guizhou University;

Email:

DOI: 10.19367/j.cnki.1000-2707.2020.04.011

参考文献(References):

文章评论(Comment):

序号(No.) 时间(Time) 反馈人(User) 邮箱(Email) 标题(Title) 内容(Content)
反馈人(User) 邮箱地址(Email)
反馈标题(Title)
反馈内容(Content)
扩展功能
本文信息
服务与反馈
本文关键词相关文章
本文作者相关文章
中国知网
分享