基于眼底彩色照相的常见眼底疾病智能辅助诊断轻量化模型研究  

Research on lightweight model of intelligent-assisted diagnosis of common fundus diseases based on fundus color photography

作　　者：陆兵 吴茂念[1,2] 郑博 朱绍军 郝秀兰 陈楠[3] 侯泽江 蒋沁[3] 杨卫华 Lu Bing;Wu Maonian;Zheng Bo;Zhu Shaojun;Hao Xiulan;Chen Nan;Hou Zejiang;Jiang Qin;Yang Weihua(School of Information Engineering,Huzhou University,Huzhou 313000,China;Zhejiang Province Key Laboratory of Smart Management&Application of Modern Agricultural Resources,Huzhou University,Hangzhou 313000,China;Ophthalmology Artificial Intelligence Big Data Laboratory,Affiliated Eye Hospital of Nanjing Medical University,Nanjing 210029,China)

机构地区：[1]湖州师范学院信息工程学院,湖州313000 [2]浙江省现代农业资源智慧管理与应用研究重点实验室,杭州313000 [3]南京医科大学附属眼科医院眼科人工智能大数据实验室,南京210029

出　　处：《中华眼底病杂志》2022年第2期146-152,共7页Chinese Journal of Ocular Fundus Diseases

基　　金：国家自然科学基金(61906066);浙江省自然科学基金(LQ18F020002);湖州市科技计划项目(2016YZ02);南京市企业专家团队工作室项目。

摘　　要：目的观察基于眼底彩色照相的常见眼底疾病六分类智能辅助诊断轻量化模型的诊断价值。方法应用研究。采集南京医科大学附属眼科医院和浙江省数理医学学会智能眼科数据库的2400张彩色眼底像数据集,该数据集经脱敏处理及眼底病专科医师标注,包括糖尿病视网膜病变、青光眼、视网膜静脉阻塞、高度近视、老年性黄斑变性、正常眼底像各400张。训练时使用迁移学习方法,将经典分类模型VGGNet16、ResNet50、DenseNet121和轻量化分类模型MobileNet3、ShuffleNet2、GhostNet在ImageNet数据集上训练获得的参数迁移到六分类常见眼底疾病智能辅助诊断模型作为初始化参数,进行训练并获得最新模型。选取临床患者彩色眼底像1315张作为测试集。评价指标包括灵敏度、特异性、准确度(Accuracy)、F1-Score和诊断试验的一致性(Kappa值);比较不同模型的受试者工作特征曲线以及曲线下面积值。结果与经典分类模型比较,3种轻量化分类模型的储存大小和参数量均显著降低,其中ShuffleNetV2平均每张识别时间较经典分类模型VGGNet16快438.08 ms。3个轻量化分类模型的准确度均>80.0%;Kappa值均>70.0%,具有显著一致性;对正常眼底图像诊断的灵敏度、特异性、F1-Score均≥98.0%;宏观F1分别为78.2%、79.4%、81.5%。结论基于眼底彩色照相的常见眼底疾病智能辅助诊断轻量化模型识别准确率高、速度快;储存大小和参数量均较经典分类模型显著降低。Objective To observe the diagnostic value of six classification intelligent auxiliary diagnosis lightweight model for common fundus diseases based on fundus color photography.Methods A applied research.A dataset of 2400 color fundus images from Nanjing Medical University Eye Hospital and Zhejiang Mathematical Medical Society Smart Eye Database was collected,which was desensitized and labeled by a fundus specialist.Of these,400 each were for diabetic retinopathy,glaucoma,retinal vein occlusion,high myopia,age-related macular degeneration,and normal fundus.The parameters obtained from the classical classification models VGGNetl6,ResNet50,DenseNetl21 and lightweight classification models MobileNet3,ShuffleNet2,GhostNet trained on the ImageNet dataset were migrated to the six-classified common fundus disease intelligent aid diagnostic model using a migration learning approach during training as initialization parameters for training to obtain the latest model.1315 color fundus images of clinical patients were used as the test set.Evaluation metrics included sensitivity,specificity,accuracy,FI-Score and agreement of diagnostic tests(Kappa value);comparison of subject working characteristic curves as well as area under the curve values for different models.Result Compared with the classical classification model,the storage size and number of parameters of the three lightweight classification models were significantly reduced,with ShuffleNetV2 having an average recognition time per sheet 438.08 ms faster than the classical classification model VGGNet 16.All 3 lightweight classification models had Accuracy>80.0%;Kappa values>70.0%with significant agreement;sensitivity,specificity,and FI-Score for the diagnosis of normal fundus images were≥98.0%;Macro-FI was 78.2%,79.4%,and 81.5%,respectively.Conclusion The intelligent assisted diagnosis of common fundus diseases based on fundus color photography is a lightweight model with high recognition accuracy and speed;the storage size and number of parameters are significantly r

关 键 词：视网膜疾病 人工智能 诊断 计算机辅助 轻量化模型 

分 类 号：R77[医药卫生—眼科]