A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence
Abstract Diabetic retinopathy (DR), the main cause of irreversible blindness, is one of the most common complications of diabetes. At present, deep convolutional neural networks have achieved promising performance in automatic DR detection tasks. The convolution operation of methods is a local cross...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-02-01
|
| Series: | CAAI Transactions on Intelligence Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/cit2.12155 |
| _version_ | 1797311454322360320 |
|---|---|
| author | Xiaoling Luo Wei Wang Yong Xu Zhihui Lai Xiaopeng Jin Bob Zhang David Zhang |
| author_facet | Xiaoling Luo Wei Wang Yong Xu Zhihui Lai Xiaopeng Jin Bob Zhang David Zhang |
| author_sort | Xiaoling Luo |
| collection | DOAJ |
| description | Abstract Diabetic retinopathy (DR), the main cause of irreversible blindness, is one of the most common complications of diabetes. At present, deep convolutional neural networks have achieved promising performance in automatic DR detection tasks. The convolution operation of methods is a local cross‐correlation operation, whose receptive field determines the size of the local neighbourhood for processing. However, for retinal fundus photographs, there is not only the local information but also long‐distance dependence between the lesion features (e.g. hemorrhages and exudates) scattered throughout the whole image. The proposed method incorporates correlations between long‐range patches into the deep learning framework to improve DR detection. Patch‐wise relationships are used to enhance the local patch features since lesions of DR usually appear as plaques. The Long‐Range unit in the proposed network with a residual structure can be flexibly embedded into other trained networks. Extensive experimental results demonstrate that the proposed approach can achieve higher accuracy than existing state‐of‐the‐art models on Messidor and EyePACS datasets. |
| first_indexed | 2024-03-08T01:59:54Z |
| format | Article |
| id | doaj.art-882c5b8b7418452791f9d7c7e3197a53 |
| institution | Directory Open Access Journal |
| issn | 2468-2322 |
| language | English |
| last_indexed | 2024-03-08T01:59:54Z |
| publishDate | 2024-02-01 |
| publisher | Wiley |
| record_format | Article |
| series | CAAI Transactions on Intelligence Technology |
| spelling | doaj.art-882c5b8b7418452791f9d7c7e3197a532024-02-14T05:37:37ZengWileyCAAI Transactions on Intelligence Technology2468-23222024-02-019115316610.1049/cit2.12155A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependenceXiaoling Luo0Wei Wang1Yong Xu2Zhihui Lai3Xiaopeng Jin4Bob Zhang5David Zhang6Shenzhen Key Laboratory of Visual Object Detection and Recognition Harbin Institute of Technology Shenzhen ChinaShenzhen Key Laboratory of Visual Object Detection and Recognition Harbin Institute of Technology Shenzhen ChinaShenzhen Key Laboratory of Visual Object Detection and Recognition Harbin Institute of Technology Shenzhen ChinaShenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen ChinaCollege of Big Data and Internet Shenzhen Technology University Shenzhen ChinaThe Department of Computer and Information Science University of Macau Macao Macau ChinaThe Chinese University of Hong Kong (Shenzhen) Shenzhen ChinaAbstract Diabetic retinopathy (DR), the main cause of irreversible blindness, is one of the most common complications of diabetes. At present, deep convolutional neural networks have achieved promising performance in automatic DR detection tasks. The convolution operation of methods is a local cross‐correlation operation, whose receptive field determines the size of the local neighbourhood for processing. However, for retinal fundus photographs, there is not only the local information but also long‐distance dependence between the lesion features (e.g. hemorrhages and exudates) scattered throughout the whole image. The proposed method incorporates correlations between long‐range patches into the deep learning framework to improve DR detection. Patch‐wise relationships are used to enhance the local patch features since lesions of DR usually appear as plaques. The Long‐Range unit in the proposed network with a residual structure can be flexibly embedded into other trained networks. Extensive experimental results demonstrate that the proposed approach can achieve higher accuracy than existing state‐of‐the‐art models on Messidor and EyePACS datasets.https://doi.org/10.1049/cit2.12155image classificationmedical image processingpattern recognition |
| spellingShingle | Xiaoling Luo Wei Wang Yong Xu Zhihui Lai Xiaopeng Jin Bob Zhang David Zhang A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence CAAI Transactions on Intelligence Technology image classification medical image processing pattern recognition |
| title | A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence |
| title_full | A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence |
| title_fullStr | A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence |
| title_full_unstemmed | A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence |
| title_short | A deep convolutional neural network for diabetic retinopathy detection via mining local and long‐range dependence |
| title_sort | deep convolutional neural network for diabetic retinopathy detection via mining local and long range dependence |
| topic | image classification medical image processing pattern recognition |
| url | https://doi.org/10.1049/cit2.12155 |
| work_keys_str_mv | AT xiaolingluo adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT weiwang adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT yongxu adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT zhihuilai adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT xiaopengjin adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT bobzhang adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT davidzhang adeepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT xiaolingluo deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT weiwang deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT yongxu deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT zhihuilai deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT xiaopengjin deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT bobzhang deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence AT davidzhang deepconvolutionalneuralnetworkfordiabeticretinopathydetectionviamininglocalandlongrangedependence |