Multi-Model Domain Adaptation for Diabetic Retinopathy Classification
Diabetic retinopathy (DR) is one of the most threatening complications in diabetic patients, leading to permanent blindness without timely treatment. However, DR screening is not only a time-consuming task that requires experienced ophthalmologists but also easy to produce misdiagnosis. In recent ye...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-07-01
|
| Series: | Frontiers in Physiology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fphys.2022.918929/full |
| _version_ | 1811227020552568832 |
|---|---|
| author | Guanghua Zhang Guanghua Zhang Bin Sun Zhaoxia Zhang Jing Pan Weihua Yang Yunfang Liu |
| author_facet | Guanghua Zhang Guanghua Zhang Bin Sun Zhaoxia Zhang Jing Pan Weihua Yang Yunfang Liu |
| author_sort | Guanghua Zhang |
| collection | DOAJ |
| description | Diabetic retinopathy (DR) is one of the most threatening complications in diabetic patients, leading to permanent blindness without timely treatment. However, DR screening is not only a time-consuming task that requires experienced ophthalmologists but also easy to produce misdiagnosis. In recent years, deep learning techniques based on convolutional neural networks have attracted increasing research attention in medical image analysis, especially for DR diagnosis. However, dataset labeling is expensive work and it is necessary for existing deep-learning-based DR detection models. For this study, a novel domain adaptation method (multi-model domain adaptation) is developed for unsupervised DR classification in unlabeled retinal images. At the same time, it only exploits discriminative information from multiple source models without access to any data. In detail, we integrate a weight mechanism into the multi-model-based domain adaptation by measuring the importance of each source domain in a novel way, and a weighted pseudo-labeling strategy is attached to the source feature extractors for training the target DR classification model. Extensive experiments are performed on four source datasets (DDR, IDRiD, Messidor, and Messidor-2) to a target domain APTOS 2019, showing that MMDA produces competitive performance for present state-of-the-art methods for DR classification. As a novel DR detection approach, this article presents a new domain adaptation solution for medical image analysis when the source data is unavailable. |
| first_indexed | 2024-04-12T09:35:26Z |
| format | Article |
| id | doaj.art-3c2ba28e6590458f9db532ac10dcdf30 |
| institution | Directory Open Access Journal |
| issn | 1664-042X |
| language | English |
| last_indexed | 2024-04-12T09:35:26Z |
| publishDate | 2022-07-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Physiology |
| spelling | doaj.art-3c2ba28e6590458f9db532ac10dcdf302022-12-22T03:38:15ZengFrontiers Media S.A.Frontiers in Physiology1664-042X2022-07-011310.3389/fphys.2022.918929918929Multi-Model Domain Adaptation for Diabetic Retinopathy ClassificationGuanghua Zhang0Guanghua Zhang1Bin Sun2Zhaoxia Zhang3Jing Pan4Weihua Yang5Yunfang Liu6Department of Intelligence and Automation, Taiyuan University, Taiyuan, ChinaGraphics and Imaging Laboratory, University of Girona, Girona, SpainShanxi Eye Hospital, Taiyuan, ChinaShanxi Eye Hospital, Taiyuan, ChinaDepartment of Materials and Chemical Engineering, Taiyuan University, Taiyuan, ChinaAffiliated Eye Hospital, Nanjing Medical University, Nanjing, ChinaThe First Affiliated Hospital of Huzhou University, Huzhou, ChinaDiabetic retinopathy (DR) is one of the most threatening complications in diabetic patients, leading to permanent blindness without timely treatment. However, DR screening is not only a time-consuming task that requires experienced ophthalmologists but also easy to produce misdiagnosis. In recent years, deep learning techniques based on convolutional neural networks have attracted increasing research attention in medical image analysis, especially for DR diagnosis. However, dataset labeling is expensive work and it is necessary for existing deep-learning-based DR detection models. For this study, a novel domain adaptation method (multi-model domain adaptation) is developed for unsupervised DR classification in unlabeled retinal images. At the same time, it only exploits discriminative information from multiple source models without access to any data. In detail, we integrate a weight mechanism into the multi-model-based domain adaptation by measuring the importance of each source domain in a novel way, and a weighted pseudo-labeling strategy is attached to the source feature extractors for training the target DR classification model. Extensive experiments are performed on four source datasets (DDR, IDRiD, Messidor, and Messidor-2) to a target domain APTOS 2019, showing that MMDA produces competitive performance for present state-of-the-art methods for DR classification. As a novel DR detection approach, this article presents a new domain adaptation solution for medical image analysis when the source data is unavailable.https://www.frontiersin.org/articles/10.3389/fphys.2022.918929/fulldiabetic retinopathy classificationmulti-modeldomain adaptationconvolutional neural networkdeep learning |
| spellingShingle | Guanghua Zhang Guanghua Zhang Bin Sun Zhaoxia Zhang Jing Pan Weihua Yang Yunfang Liu Multi-Model Domain Adaptation for Diabetic Retinopathy Classification Frontiers in Physiology diabetic retinopathy classification multi-model domain adaptation convolutional neural network deep learning |
| title | Multi-Model Domain Adaptation for Diabetic Retinopathy Classification |
| title_full | Multi-Model Domain Adaptation for Diabetic Retinopathy Classification |
| title_fullStr | Multi-Model Domain Adaptation for Diabetic Retinopathy Classification |
| title_full_unstemmed | Multi-Model Domain Adaptation for Diabetic Retinopathy Classification |
| title_short | Multi-Model Domain Adaptation for Diabetic Retinopathy Classification |
| title_sort | multi model domain adaptation for diabetic retinopathy classification |
| topic | diabetic retinopathy classification multi-model domain adaptation convolutional neural network deep learning |
| url | https://www.frontiersin.org/articles/10.3389/fphys.2022.918929/full |
| work_keys_str_mv | AT guanghuazhang multimodeldomainadaptationfordiabeticretinopathyclassification AT guanghuazhang multimodeldomainadaptationfordiabeticretinopathyclassification AT binsun multimodeldomainadaptationfordiabeticretinopathyclassification AT zhaoxiazhang multimodeldomainadaptationfordiabeticretinopathyclassification AT jingpan multimodeldomainadaptationfordiabeticretinopathyclassification AT weihuayang multimodeldomainadaptationfordiabeticretinopathyclassification AT yunfangliu multimodeldomainadaptationfordiabeticretinopathyclassification |