深度学习在糖尿病视网膜病变检测和分类中的研究  

The Study of Deep Learning in Diabetic Retinopathy Detection and Classification

作　　者：苗玥[1] 邢俊凤[1] 杜鹏程[2] MIAO Yue;XING Junfeng;DU Pengcheng(Inner Mongolia University of Science and Technology Department of Computer Science and Technology,Baotou Medical College,Baotou Inner Mongolia 014010;Inner Mongolia University of Science and Technology Department of Ophthalmology,First Affiliated Hospital of Baotou Medical College,Baotou Inner Mongolia 014010)

机构地区：[1]内蒙古科技大学包头医学院计算机科学与技术系,内蒙古包头014010 [2]内蒙古科技大学包头医学院一附院眼科,内蒙古包头014010

出　　处：《中国科技纵横》2023年第22期67-69,135,共4页China Science & Technology Overview

基　　金：内蒙古自然科学基金项目(2021MS06010);包头医学院科创融合教育研究项目(BYJJ-KCRH202302);包头医学院创新团队项目(bycxtd-16);包头医学院2022年课程思政专项教学改革项目(2022KCSZ-19)。

摘　　要：糖尿病视网膜病变在早期诊断中不易察觉,如果错过了治疗的最佳时期,得不到及时治疗,将会导致失明。深度学习技术可以高效辅助当前糖尿病视网膜的大规模影像早期筛查和诊断。基于此,介绍深度学习在糖尿病视网膜中的研究价值和意义,研究应用于糖尿病视网膜图像检测和严重性分类的算法,讨论未解决的问题。Diabetes retinopathy is difficult to detect in the early diagnosis.If the best period of treatment is missed and not treated in time,blindness will result.Deep learning technology can effectively assist the current large-scale image early screening and diagnosis of diabetes retina.This study introduced the research value and significance of deep learning in diabetes retina,studied the algorithms applied to diabetes retina image detection and severity classification,and discussed the unresolved problems.

关 键 词：深度学习 糖尿病视网膜 分类 检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]