摘要
目的 基于眼底彩照,分别应用经典卷积神经网络 DenseNet121 和轻量级网络 GhostNet 训练糖尿病视网膜病变(diabetic retinopathy,DR)的诊断模型(将DR和正常眼底做区分)和鉴别诊断(将DR和其他眼底病做区分)模型,评价基于轻量级网络GhostNet的DR诊断模型的应用价值.方法 收集大样本的眼底彩照29 535 张(含DR 9 883 张、正常眼底2 000 张、用于做鉴别诊断的其他致盲性眼底病 17 652 张).分别采用经典卷积神经网络DenseNet121 和轻量级网络GhostNet建模,并借助迁移学习做模型训练.采用受试者工作特征(receiver operating characteristic,ROC)曲线及其曲线下面积(area under the curve,AUC)、灵敏度、特异度、准确率评价模型性能.结果 与基于DenseNet121 的模型相比,基于GhostNet的模型对单张眼底照的诊断时间缩短了 60.3%.在DR的诊断方面,基于GhostNet的模型的AUC值、灵敏度、特异度、准确率分别为 0.911、0.888、0.934、91.3%,基于DenseNet121 的模型的AUC值、灵敏度、特异度、准确率分别为 0.954、0.921、0.986、95.5%.在DR与其他眼底病的鉴别诊断方面,基于GhostNet的模型的AUC值、灵敏度、特异度、准确率分别为 0.862、0.856、0.901、87.8%;基于DenseNet121 的模型的AUC值、灵敏度、特异度、准确率分别为 0.899、0.871、0.935、90.2%.结论 基于GhostNet轻量级神经网络构建的DR诊断模型和鉴别诊断模型,其诊断效率较经典模型DenseNet121 有显著提升,并且模型兼具较高的准确率.对于社区医院等缺乏眼科医师且设备性能不高的基层医疗机构,可考虑应用该技术开展DR的初筛.
Abstract
Objective To evaluate the value of GhostNet lightweight network based on fundus color imaging in the diagnosis of diabetic retinopathy(DR),and the value of differential diagnosis between DR and other fundus diseases compared with classical convolutional neural network DenseNet121.Methods Totally 29 535 color fundus photographs(including 9 883 DR,2 000 normal fundus and 17 652 other blinding fundus diseases for differential diagnosis)were collected from large samples.Classical convolutional neural network(DenseNet121)and lightweight network(GhostNet)were used to train the model by transfer learning.Receiver operating characteristic(ROC)curve,sensitivity,specificity,positive predictive value,negative predictive value and area under the curve(AUC)were used to evaluate the performance of the two models.Results Compared with model based on DenseNet121,the diagnostic time of single fundus photo by GhostNet model was shortened by 60.3%.In the diagnosis of DR,AUC value,sensitivity and specificity,accuracy reached 0.911,0.888,0.934,and 91.3%respectively.It is slightly lower than 0.954,0.921,0.986,and 95.5%of model based on DenseNet121.In the differential diagnosis of DR and other fundus diseases,the AUC value,sensitivity,specificity,accuracy of model based on GhostNet were 0.862,0.856,0.901,and 87.8%respectively.For model based on DenseNet121,the values were 0.899,0.871,0.935,and 90.2%,respectively.Conclusions The operation speed of DR diagnosis model and differential diagnosis model based on GhostNet lightweight neural network is significantly faster than that of classical model DenseNet121,and the accuracy is satisfying.For primary medical institutions with lack of ophthalmologists and low equipment performance,such as community hospitals,we can consider using this technology to carry out primary screening of DR.
基金项目
北京市医院管理中心"青苗计划"专项(QMS20210215)
NETL
NSTL
万方数据