EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy
Present architecture of convolution neural network for diabetic retinopathy (DR-Net) is based on normal convolution (NC). It incurs high computational cost as NC uses a multiplicative weight that measures a combined correlation in both cross-channel and spatial dimension of layer’s inputs...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Institute of Electrical and Electronics Engineers Inc.
2022
|
| Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85132695820&doi=10.1109%2fTBCAS.2022.3182907&partnerID=40&md5=2f47cbd2430b2c0bfdb893d962b85891 http://eprints.utp.edu.my/33165/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| id |
my.utp.eprints.33165 |
|---|---|
| record_format |
eprints |
| spelling |
my.utp.eprints.331652022-07-06T08:05:05Z EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy Aujih, A.B. Shapiai, M.I. Meriaudeau, F. Tang, T.B. Present architecture of convolution neural network for diabetic retinopathy (DR-Net) is based on normal convolution (NC). It incurs high computational cost as NC uses a multiplicative weight that measures a combined correlation in both cross-channel and spatial dimension of layer’s inputs. This might cause the overall DR-Net architecture to be over-parameterised and computationally inefficient. This paper proposes EDR-Net – a new end-to-end, DR-Net architecture with depth-wise separable convolution module. The EDR-Net architecture was trained with DRKaggle-train dataset (35,126 images), and tested on two datasets, i.e. DRKaggle-test (53,576 images) and Messidor-2 (1,748 images). Results showed that the proposed EDR-Net achieved predictive performance comparable with current state-of-the-arts in detecting referable diabetic retinopathy (rDR) from fundus images and outperformed other light weight architectures, with at least two times less computation cost. This makes it more amenable for mobile device based computer-assisted rDR screening applications. IEEE Institute of Electrical and Electronics Engineers Inc. 2022 Article NonPeerReviewed https://www.scopus.com/inward/record.uri?eid=2-s2.0-85132695820&doi=10.1109%2fTBCAS.2022.3182907&partnerID=40&md5=2f47cbd2430b2c0bfdb893d962b85891 Aujih, A.B. and Shapiai, M.I. and Meriaudeau, F. and Tang, T.B. (2022) EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy. IEEE Transactions on Biomedical Circuits and Systems . pp. 1-12. http://eprints.utp.edu.my/33165/ |
| institution |
Universiti Teknologi Petronas |
| building |
UTP Resource Centre |
| collection |
Institutional Repository |
| continent |
Asia |
| country |
Malaysia |
| content_provider |
Universiti Teknologi Petronas |
| content_source |
UTP Institutional Repository |
| url_provider |
http://eprints.utp.edu.my/ |
| description |
Present architecture of convolution neural network for diabetic retinopathy (DR-Net) is based on normal convolution (NC). It incurs high computational cost as NC uses a multiplicative weight that measures a combined correlation in both cross-channel and spatial dimension of layer’s inputs. This might cause the overall DR-Net architecture to be over-parameterised and computationally inefficient. This paper proposes EDR-Net – a new end-to-end, DR-Net architecture with depth-wise separable convolution module. The EDR-Net architecture was trained with DRKaggle-train dataset (35,126 images), and tested on two datasets, i.e. DRKaggle-test (53,576 images) and Messidor-2 (1,748 images). Results showed that the proposed EDR-Net achieved predictive performance comparable with current state-of-the-arts in detecting referable diabetic retinopathy (rDR) from fundus images and outperformed other light weight architectures, with at least two times less computation cost. This makes it more amenable for mobile device based computer-assisted rDR screening applications. IEEE |
| format |
Article |
| author |
Aujih, A.B. Shapiai, M.I. Meriaudeau, F. Tang, T.B. |
| spellingShingle |
Aujih, A.B. Shapiai, M.I. Meriaudeau, F. Tang, T.B. EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| author_facet |
Aujih, A.B. Shapiai, M.I. Meriaudeau, F. Tang, T.B. |
| author_sort |
Aujih, A.B. |
| title |
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| title_short |
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| title_full |
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| title_fullStr |
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| title_full_unstemmed |
EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy |
| title_sort |
edr-net: lightweight deep neural network architecture for detecting referable diabetic retinopathy |
| publisher |
Institute of Electrical and Electronics Engineers Inc. |
| publishDate |
2022 |
| url |
https://www.scopus.com/inward/record.uri?eid=2-s2.0-85132695820&doi=10.1109%2fTBCAS.2022.3182907&partnerID=40&md5=2f47cbd2430b2c0bfdb893d962b85891 http://eprints.utp.edu.my/33165/ |
| _version_ |
1738657464377671680 |
| score |
13.211869 |