Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy
This study is aimed at evaluating a deep transfer learning-based model for identifying diabetic retinopathy (DR) that was trained using a dataset with high variability and predominant type 2 diabetes (T2D) and comparing model performance with that in patients with type 1 diabetes (T1D). The Kaggle d...
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2021-01-01
|
| Series: | Journal of Diabetes Research |
| Online Access: | http://dx.doi.org/10.1155/2021/2751695 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832566212273373184 |
|---|---|
| author | Jui-En Lo Eugene Yu-Chuan Kang Yun-Nung Chen Yi-Ting Hsieh Nan-Kai Wang Ta-Ching Chen Kuan-Jen Chen Wei-Chi Wu Yih-Shiou Hwang Fu-Sung Lo Chi-Chun Lai |
| author_facet | Jui-En Lo Eugene Yu-Chuan Kang Yun-Nung Chen Yi-Ting Hsieh Nan-Kai Wang Ta-Ching Chen Kuan-Jen Chen Wei-Chi Wu Yih-Shiou Hwang Fu-Sung Lo Chi-Chun Lai |
| author_sort | Jui-En Lo |
| collection | DOAJ |
| description | This study is aimed at evaluating a deep transfer learning-based model for identifying diabetic retinopathy (DR) that was trained using a dataset with high variability and predominant type 2 diabetes (T2D) and comparing model performance with that in patients with type 1 diabetes (T1D). The Kaggle dataset, which is a publicly available dataset, was divided into training and testing Kaggle datasets. In the comparison dataset, we collected retinal fundus images of T1D patients at Chang Gung Memorial Hospital in Taiwan from 2013 to 2020, and the images were divided into training and testing T1D datasets. The model was developed using 4 different convolutional neural networks (Inception-V3, DenseNet-121, VGG1, and Xception). The model performance in predicting DR was evaluated using testing images from each dataset, and area under the curve (AUC), sensitivity, and specificity were calculated. The model trained using the Kaggle dataset had an average (range) AUC of 0.74 (0.03) and 0.87 (0.01) in the testing Kaggle and T1D datasets, respectively. The model trained using the T1D dataset had an AUC of 0.88 (0.03), which decreased to 0.57 (0.02) in the testing Kaggle dataset. Heatmaps showed that the model focused on retinal hemorrhage, vessels, and exudation to predict DR. In wrong prediction images, artifacts and low-image quality affected model performance. The model developed with the high variability and T2D predominant dataset could be applied to T1D patients. Dataset homogeneity could affect the performance, trainability, and generalization of the model. |
| format | Article |
| id | doaj-art-6dbf1ab63a6b4d289165e185d9f863d1 |
| institution | Kabale University |
| issn | 2314-6753 |
| language | English |
| publishDate | 2021-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Journal of Diabetes Research |
| spelling | doaj-art-6dbf1ab63a6b4d289165e185d9f863d12025-02-03T01:04:46ZengWileyJournal of Diabetes Research2314-67532021-01-01202110.1155/2021/2751695Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic RetinopathyJui-En Lo0Eugene Yu-Chuan Kang1Yun-Nung Chen2Yi-Ting Hsieh3Nan-Kai Wang4Ta-Ching Chen5Kuan-Jen Chen6Wei-Chi Wu7Yih-Shiou Hwang8Fu-Sung Lo9Chi-Chun Lai10School of MedicineDepartment of OphthalmologyDepartment of Computer Science and Information Engineering National Taiwan UniversityDepartment of OphthalmologyDepartment of OphthalmologyDepartment of OphthalmologyDepartment of OphthalmologyDepartment of OphthalmologyDepartment of OphthalmologyDivision of Pediatric Endocrinology and GeneticsDepartment of OphthalmologyThis study is aimed at evaluating a deep transfer learning-based model for identifying diabetic retinopathy (DR) that was trained using a dataset with high variability and predominant type 2 diabetes (T2D) and comparing model performance with that in patients with type 1 diabetes (T1D). The Kaggle dataset, which is a publicly available dataset, was divided into training and testing Kaggle datasets. In the comparison dataset, we collected retinal fundus images of T1D patients at Chang Gung Memorial Hospital in Taiwan from 2013 to 2020, and the images were divided into training and testing T1D datasets. The model was developed using 4 different convolutional neural networks (Inception-V3, DenseNet-121, VGG1, and Xception). The model performance in predicting DR was evaluated using testing images from each dataset, and area under the curve (AUC), sensitivity, and specificity were calculated. The model trained using the Kaggle dataset had an average (range) AUC of 0.74 (0.03) and 0.87 (0.01) in the testing Kaggle and T1D datasets, respectively. The model trained using the T1D dataset had an AUC of 0.88 (0.03), which decreased to 0.57 (0.02) in the testing Kaggle dataset. Heatmaps showed that the model focused on retinal hemorrhage, vessels, and exudation to predict DR. In wrong prediction images, artifacts and low-image quality affected model performance. The model developed with the high variability and T2D predominant dataset could be applied to T1D patients. Dataset homogeneity could affect the performance, trainability, and generalization of the model.http://dx.doi.org/10.1155/2021/2751695 |
| spellingShingle | Jui-En Lo Eugene Yu-Chuan Kang Yun-Nung Chen Yi-Ting Hsieh Nan-Kai Wang Ta-Ching Chen Kuan-Jen Chen Wei-Chi Wu Yih-Shiou Hwang Fu-Sung Lo Chi-Chun Lai Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy Journal of Diabetes Research |
| title | Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy |
| title_full | Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy |
| title_fullStr | Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy |
| title_full_unstemmed | Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy |
| title_short | Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy |
| title_sort | data homogeneity effect in deep learning based prediction of type 1 diabetic retinopathy |
| url | http://dx.doi.org/10.1155/2021/2751695 |
| work_keys_str_mv | AT juienlo datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT eugeneyuchuankang datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT yunnungchen datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT yitinghsieh datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT nankaiwang datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT tachingchen datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT kuanjenchen datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT weichiwu datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT yihshiouhwang datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT fusunglo datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy AT chichunlai datahomogeneityeffectindeeplearningbasedpredictionoftype1diabeticretinopathy |