Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging
Summary: Retinopathy of prematurity (ROP) is currently one of the leading causes of infant blindness worldwide. Recently significant progress has been made in deep learning-based computer-aided diagnostic methods. However, deep learning often requires a large amount of annotated data for model optim...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-01-01
|
| Series: | iScience |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2589004223025932 |
| _version_ | 1797371382467657728 |
|---|---|
| author | Wei Feng Qiujing Huang Tong Ma Lie Ju Zongyuan Ge Yuzhong Chen Peiquan Zhao |
| author_facet | Wei Feng Qiujing Huang Tong Ma Lie Ju Zongyuan Ge Yuzhong Chen Peiquan Zhao |
| author_sort | Wei Feng |
| collection | DOAJ |
| description | Summary: Retinopathy of prematurity (ROP) is currently one of the leading causes of infant blindness worldwide. Recently significant progress has been made in deep learning-based computer-aided diagnostic methods. However, deep learning often requires a large amount of annotated data for model optimization, but this requires long hours of effort by experienced doctors in clinical scenarios. In contrast, a large number of unlabeled images are relatively easy to obtain. In this paper, we propose a new semi-supervised learning framework to reduce annotation costs for automatic ROP staging. We design two consistency regularization strategies, prediction consistency loss and semantic structure consistency loss, which can help the model mine useful discriminative information from unlabeled data, thus improving the generalization performance of the classification model. Extensive experiments on a real clinical dataset show that the proposed method promises to greatly reduce the labeling requirements in clinical scenarios while achieving good classification performance. |
| first_indexed | 2024-03-08T18:19:00Z |
| format | Article |
| id | doaj.art-e23d4045559c4eef9a90a2b853c4c6f8 |
| institution | Directory Open Access Journal |
| issn | 2589-0042 |
| language | English |
| last_indexed | 2024-03-08T18:19:00Z |
| publishDate | 2024-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | iScience |
| spelling | doaj.art-e23d4045559c4eef9a90a2b853c4c6f82023-12-31T04:26:32ZengElsevieriScience2589-00422024-01-01271108516Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity stagingWei Feng0Qiujing Huang1Tong Ma2Lie Ju3Zongyuan Ge4Yuzhong Chen5Peiquan Zhao6Beijing Airdoc Technology Co., Ltd, Beijing 100089, China; Faculty of Engineering, Monash University, Melbourne, VIC 3000, AustraliaDepartment of Ophthalmology, Xinhua Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200092, China; Department of Ophthalmology, Rainbow Children’s Clinic, Shanghai 200010, China; Corresponding authorBeijing Airdoc Technology Co., Ltd, Beijing 100089, ChinaBeijing Airdoc Technology Co., Ltd, Beijing 100089, China; Faculty of Engineering, Monash University, Melbourne, VIC 3000, AustraliaFaculty of Engineering, Monash University, Melbourne, VIC 3000, AustraliaBeijing Airdoc Technology Co., Ltd, Beijing 100089, ChinaDepartment of Ophthalmology, Xinhua Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200092, China; Corresponding authorSummary: Retinopathy of prematurity (ROP) is currently one of the leading causes of infant blindness worldwide. Recently significant progress has been made in deep learning-based computer-aided diagnostic methods. However, deep learning often requires a large amount of annotated data for model optimization, but this requires long hours of effort by experienced doctors in clinical scenarios. In contrast, a large number of unlabeled images are relatively easy to obtain. In this paper, we propose a new semi-supervised learning framework to reduce annotation costs for automatic ROP staging. We design two consistency regularization strategies, prediction consistency loss and semantic structure consistency loss, which can help the model mine useful discriminative information from unlabeled data, thus improving the generalization performance of the classification model. Extensive experiments on a real clinical dataset show that the proposed method promises to greatly reduce the labeling requirements in clinical scenarios while achieving good classification performance.http://www.sciencedirect.com/science/article/pii/S2589004223025932Health technologyApplied computing |
| spellingShingle | Wei Feng Qiujing Huang Tong Ma Lie Ju Zongyuan Ge Yuzhong Chen Peiquan Zhao Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging iScience Health technology Applied computing |
| title | Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging |
| title_full | Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging |
| title_fullStr | Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging |
| title_full_unstemmed | Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging |
| title_short | Development and validation of a semi-supervised deep learning model for automatic retinopathy of prematurity staging |
| title_sort | development and validation of a semi supervised deep learning model for automatic retinopathy of prematurity staging |
| topic | Health technology Applied computing |
| url | http://www.sciencedirect.com/science/article/pii/S2589004223025932 |
| work_keys_str_mv | AT weifeng developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT qiujinghuang developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT tongma developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT lieju developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT zongyuange developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT yuzhongchen developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging AT peiquanzhao developmentandvalidationofasemisuperviseddeeplearningmodelforautomaticretinopathyofprematuritystaging |