|
|||||
|
|
| 基于深度学习的翼状胬肉自动分类系统研究 | |
| 引用本文： | 何楷,吴茂念,郑博,杨卫华,朱绍军,金玲. 基于深度学习的翼状胬肉自动分类系统研究[J]. 国际眼科杂志, 2022, 22(5): 711-715 |
| 作者姓名： | 何楷 吴茂念 郑博 杨卫华 朱绍军 金玲 |
| 作者单位： | 中国浙江省湖州市,湖州师范学院信息工程学院; 中国浙江省湖州市,浙江省现代农业资源智慧管理与应用研究重点实验室,中国浙江省湖州市,湖州师范学院信息工程学院; 中国浙江省湖州市,浙江省现代农业资源智慧管理与应用研究重点实验室,中国浙江省湖州市,湖州师范学院信息工程学院; 中国浙江省湖州市,浙江省现代农业资源智慧管理与应用研究重点实验室,中国江苏省南京市,南京医科大学附属眼科医院 眼科人工智能大数据实验室,中国浙江省湖州市,湖州师范学院信息工程学院; 中国浙江省湖州市,浙江省现代农业资源智慧管理与应用研究重点实验室,中国江苏省南京市,南京医科大学附属眼科医院 眼科人工智能大数据实验室 |
| 基金项目： | 国家自然科学青年基金项目(No.61906066); 浙江省自然科学基金项目(No.LQ18F020002); 浙江省教育厅一般科研项目(No.Y202147191); 湖州市科技计划项目(No.2016YZ02); 南京市企业专家团队工作室项目; 湖州师范学院研究生科研创新项目(No.2020KYCX21) |
| 摘 要： | 目的:评估基于深度学习的翼状胬肉自动分类诊断系统的应用价值。方法:在2020-05/2021-04期间，从南京医科大学附属眼科医院共收集750张翼状胬肉正常、观察期和手术期眼前节图片。在原始数据集和增强数据集上分别训练7个三分类模型。测试临床470张图片，比较数据增强前后模型的泛化能力，确定可用于翼状胬肉自动分类系统的最好模型。结果:在原始数据集上训练最好模型的灵敏度平均值为92.55%,特异度平均值为96.86%,AUC平均值为94.70%。数据增强后，不同模型灵敏度、特异度和AUC平均提升3.7%、1.9%和2.7%。在增强数据集上训练的EfficientNetB7模型灵敏度平均值为93.63%,特异度平均值为97.34%,AUC平均值为95.47%。结论:在增强数据集上训练的EfficientNetB7模型取得最好的分类效果，可用于翼状胬肉自动分类系统。该自动分类系统能较好地诊断翼状胬肉疾病，有望成为基层医疗的有效筛查工具，也为翼状胬肉的细化分级研究提供参考。 |
| 关 键 词： | 人工智能 深度学习 翼状胬肉 分类模型 数据增强 迁移学习 |
| 收稿时间： | 2021-07-21 |
| 修稿时间： | 2022-03-29 |
Research on the automatic classification system of pterygium based on deep learning |
|
| Kai He,Mao-Nian Wu,Bo Zheng,Wei-Hua Yang,Shao-Jun Zhu and Ling Jin. Research on the automatic classification system of pterygium based on deep learning[J]. International Eye Science, 2022, 22(5): 711-715 | |
| Authors: | Kai He Mao-Nian Wu Bo Zheng Wei-Hua Yang Shao-Jun Zhu Ling Jin |
| Affiliation: | School of Information Engineering, Huzhou University, Huzhou 313000, Zhejiang Province, China; Zhejiang Province Key Laboratory of Smart Management and Application of Modern Agricultural Resources, Huzhou University, Huzhou 313000, Zhejiang Province, China,School of Information Engineering, Huzhou University, Huzhou 313000, Zhejiang Province, China; Zhejiang Province Key Laboratory of Smart Management and Application of Modern Agricultural Resources, Huzhou University, Huzhou 313000, Zhejiang Province, China,School of Information Engineering, Huzhou University, Huzhou 313000, Zhejiang Province, China; Zhejiang Province Key Laboratory of Smart Management and Application of Modern Agricultural Resources, Huzhou University, Huzhou 313000, Zhejiang Province, China,Big Data Laboratory of Ophthalmic Artificial Intelligence;the Affiliated Eye Hospital of Nanjing Medical University, Nanjing 210029, Jiangsu Province, China,School of Information Engineering, Huzhou University, Huzhou 313000, Zhejiang Province, China; Zhejiang Province Key Laboratory of Smart Management and Application of Modern Agricultural Resources, Huzhou University, Huzhou 313000, Zhejiang Province, China and Big Data Laboratory of Ophthalmic Artificial Intelligence;the Affiliated Eye Hospital of Nanjing Medical University, Nanjing 210029, Jiangsu Province, China |
| Abstract: | AIM: To evaluate the application value of the automatic classification and diagnosis system of pterygium based on deep learning. METHODS: A total of 750 images of normal, observational and operative anterior sections of pterygium were collected from the Affiliated Eye Hospital of Nanjing Medical University between May 2020 and April 2021. Seven triclassification models were respectively trained with original data set and enhanced data set. Totally 470 clinical images were tested, and the generalization ability of the model before and after data enhancement was compared to determine the best model for the automatic classification system of pterygium. RESULTS:The average sensitivity, specificity and AUC of the best model trained on the original data set were 92.55%, 96.86% and 94.70% respectively. After data was enhanced, the sensitivity, specificity and AUC of different models were increased by 3.7%, 1.9% and 2.7% on average. The sensitivity, specificity and AUC of the EfficientNetB7 model trained on the enhanced data set were 93.63%, 97.34% and 95.47% respectively. CONDLUSION: The EfficientNetB7 model, which was trained on the enhanced data set, achieves the best classification result and can be used in the automatic classification system of pterygium.This automatic classification system can diagnose diseases about pterygium better, and it is expected to be an effective screening tool for primary medical care. It also provides reference for the research on the refinement and grading of pterygium. |
| Keywords: | artificial intelligence deep learning pterygium classification model data-enhancement transfer learning |
| 点击此处可从《国际眼科杂志》浏览原始摘要信息 | |
| 点击此处可从《国际眼科杂志》下载全文 | |