摘要
本文针对传统的视网膜图像处理过程繁琐、鲁棒性差的缺点,提出设计了一种基于深度卷积神经网络(Convolutional Neural Network,CNN)的视网膜图像自动识别系统。首先,对图像预处理包括去除噪声、数值归一化、数据量扩增;然后,设计提出了一种新的神经网络模型——XNet,XNet中和了LeNet和Inception网络的深度,网络参数根据训练样本进行适应性调整;最后,针对不同的网络结构进行了准确率和迭代次数的比较。实验结果表明,XNet网络的结构要优于LeNet和Inception,准确率可以达到91%;并通过实验证实了数据扩增的必要性。
In this research,a retina image automatic recognition system based on Convolutional Neural Network(CNN)is proposed for the disadvantages of the traditional retina image processing process which is cumbersome and poor in robustness.First,image preprocessing includes noise removal,numerical normalization,and data volume amplification;then,a new neural network model,XNet,is designed.XNet inherits the advantages of LeNet and Inception networks.The network parameters are based on training.The samples were adjusted adaptively.Finally,the comparison of accuracy and number of iterations was performed for different network structures.The experimental results show that the XNet network structure is superior to LeNet and Inception,and the accuracy rate can reach 91%;and the necessity of data amplification is confirmed through experiments.
作者
陈宇
周雨佳
丁辉
CHEN Yu;ZHOU Yu-jia;DING Hui(School of Information and Computer Engineering of Northeast Forestry University,Harbin 150040,China;Hainan Eye Hospital,Haikou 570100,China)
出处
《哈尔滨理工大学学报》
CAS
北大核心
2020年第1期73-79,共7页
Journal of Harbin University of Science and Technology
基金
中央高校基本科研业务费(2572015DY07)
国家自然科学基金(60972127)
黑龙江省教育厅科学技术研究项目(12513016)
黑龙江省博士后基金
黑龙江省自然科学基金(F201347)
哈尔滨市科技创新人才专项资金(2013RFQXJ100)。
关键词
卷积神经网络
深度学习
视网膜分类
糖尿病视网膜图像
convolution neural network
deep learning
retinal images classification
Diabetic Retinal Images
作者简介
周雨佳(1995—),女,硕士研究生;丁辉(1982—),女,博士,主治医师;通讯作者:陈宇(1975—),男,博士后,副教授,E-mail:nefu_chenyu@163.com。
