Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models
This study focuses on overcoming challenges in classifying eye diseases using color fundus photographs by leveraging deep learning techniques, aiming to enhance early detection and diagnosis accuracy. We utilized a dataset of 6392 color fundus photographs across eight disease categories, which was l...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-08-01
|
| Series: | Healthcare |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-9032/11/15/2228 |
| _version_ | 1797586743233347584 |
|---|---|
| author | Jing-Zhe Wang Nan-Han Lu Wei-Chang Du Kuo-Ying Liu Shih-Yen Hsu Chi-Yuan Wang Yun-Ju Chen Li-Ching Chang Wen-Hung Twan Tai-Been Chen Yung-Hui Huang |
| author_facet | Jing-Zhe Wang Nan-Han Lu Wei-Chang Du Kuo-Ying Liu Shih-Yen Hsu Chi-Yuan Wang Yun-Ju Chen Li-Ching Chang Wen-Hung Twan Tai-Been Chen Yung-Hui Huang |
| author_sort | Jing-Zhe Wang |
| collection | DOAJ |
| description | This study focuses on overcoming challenges in classifying eye diseases using color fundus photographs by leveraging deep learning techniques, aiming to enhance early detection and diagnosis accuracy. We utilized a dataset of 6392 color fundus photographs across eight disease categories, which was later augmented to 17,766 images. Five well-known convolutional neural networks (CNNs)—efficientnetb0, mobilenetv2, shufflenet, resnet50, and resnet101—and a custom-built CNN were integrated and trained on this dataset. Image sizes were standardized, and model performance was evaluated via accuracy, Kappa coefficient, and precision metrics. Shufflenet and efficientnetb0demonstrated strong performances, while our custom 17-layer CNN outperformed all with an accuracy of 0.930 and a Kappa coefficient of 0.920. Furthermore, we found that the fusion of image features with classical machine learning classifiers increased the performance, with Logistic Regression showcasing the best results. Our study highlights the potential of AI and deep learning models in accurately classifying eye diseases and demonstrates the efficacy of custom-built models and the fusion of deep learning and classical methods. Future work should focus on validating these methods across larger datasets and assessing their real-world applicability. |
| first_indexed | 2024-03-11T00:26:29Z |
| format | Article |
| id | doaj.art-768472179dbb4d828dc9f223cd181998 |
| institution | Directory Open Access Journal |
| issn | 2227-9032 |
| language | English |
| last_indexed | 2024-03-11T00:26:29Z |
| publishDate | 2023-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Healthcare |
| spelling | doaj.art-768472179dbb4d828dc9f223cd1819982023-11-18T22:57:17ZengMDPI AGHealthcare2227-90322023-08-011115222810.3390/healthcare11152228Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN ModelsJing-Zhe Wang0Nan-Han Lu1Wei-Chang Du2Kuo-Ying Liu3Shih-Yen Hsu4Chi-Yuan Wang5Yun-Ju Chen6Li-Ching Chang7Wen-Hung Twan8Tai-Been Chen9Yung-Hui Huang10Department of Information Engineering, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 84001, TaiwanDepartment of Medical Imaging and Radiological Science, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 82445, TaiwanDepartment of Information Engineering, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 84001, TaiwanDepartment of Radiology, E-DA Cancer Hospital, I-Shou University, No. 21, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 82445, TaiwanDepartment of Information Engineering, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 84001, TaiwanDepartment of Medical Imaging and Radiological Science, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 82445, TaiwanSchool of Medicine for International Students, I-Shu University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 84001, TaiwanSchool of Medicine for International Students, I-Shu University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 84001, TaiwanDepartment of Life Sciences, National Taitung University, No. 369, Sec. 2, University Road, Taitung City 95048, TaiwanDepartment of Medical Imaging and Radiological Science, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 82445, TaiwanDepartment of Medical Imaging and Radiological Science, I-Shou University, No. 8, Yida Road, Jiao-Su Village, Yan-Chao District, Kaohsiung City 82445, TaiwanThis study focuses on overcoming challenges in classifying eye diseases using color fundus photographs by leveraging deep learning techniques, aiming to enhance early detection and diagnosis accuracy. We utilized a dataset of 6392 color fundus photographs across eight disease categories, which was later augmented to 17,766 images. Five well-known convolutional neural networks (CNNs)—efficientnetb0, mobilenetv2, shufflenet, resnet50, and resnet101—and a custom-built CNN were integrated and trained on this dataset. Image sizes were standardized, and model performance was evaluated via accuracy, Kappa coefficient, and precision metrics. Shufflenet and efficientnetb0demonstrated strong performances, while our custom 17-layer CNN outperformed all with an accuracy of 0.930 and a Kappa coefficient of 0.920. Furthermore, we found that the fusion of image features with classical machine learning classifiers increased the performance, with Logistic Regression showcasing the best results. Our study highlights the potential of AI and deep learning models in accurately classifying eye diseases and demonstrates the efficacy of custom-built models and the fusion of deep learning and classical methods. Future work should focus on validating these methods across larger datasets and assessing their real-world applicability.https://www.mdpi.com/2227-9032/11/15/2228color fundus photographsCNNdeep learning |
| spellingShingle | Jing-Zhe Wang Nan-Han Lu Wei-Chang Du Kuo-Ying Liu Shih-Yen Hsu Chi-Yuan Wang Yun-Ju Chen Li-Ching Chang Wen-Hung Twan Tai-Been Chen Yung-Hui Huang Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models Healthcare color fundus photographs CNN deep learning |
| title | Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models |
| title_full | Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models |
| title_fullStr | Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models |
| title_full_unstemmed | Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models |
| title_short | Classification of Color Fundus Photographs Using Fusion Extracted Features and Customized CNN Models |
| title_sort | classification of color fundus photographs using fusion extracted features and customized cnn models |
| topic | color fundus photographs CNN deep learning |
| url | https://www.mdpi.com/2227-9032/11/15/2228 |
| work_keys_str_mv | AT jingzhewang classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT nanhanlu classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT weichangdu classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT kuoyingliu classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT shihyenhsu classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT chiyuanwang classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT yunjuchen classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT lichingchang classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT wenhungtwan classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT taibeenchen classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels AT yunghuihuang classificationofcolorfundusphotographsusingfusionextractedfeaturesandcustomizedcnnmodels |