Home > Research > Publications & Outputs > Multiscale sequential convolutional neural netw...

Links

Text available via DOI:

View graph of relations

Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc. / Al-Bander, Baidaa; Al-Nuaimy, Waleed; Williams, Bryan M. et al.
In: Biomedical Signal Processing and Control, Vol. 40, 01.02.2018, p. 91-101.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Al-Bander, B, Al-Nuaimy, W, Williams, BM & Zheng, Y 2018, 'Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc', Biomedical Signal Processing and Control, vol. 40, pp. 91-101. https://doi.org/10.1016/j.bspc.2017.09.008

APA

Vancouver

Al-Bander B, Al-Nuaimy W, Williams BM, Zheng Y. Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc. Biomedical Signal Processing and Control. 2018 Feb 1;40:91-101. Epub 2017 Sept 23. doi: 10.1016/j.bspc.2017.09.008

Author

Al-Bander, Baidaa ; Al-Nuaimy, Waleed ; Williams, Bryan M. et al. / Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc. In: Biomedical Signal Processing and Control. 2018 ; Vol. 40. pp. 91-101.

Bibtex

@article{b3b7772357cf4e03bd7ed33e35093fb4,
title = "Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc",
abstract = "Detecting the locations of the optic disc and fovea is a crucial task towards developing automatic diagnosis and screening tools for retinal disease. We propose to address this challenging problem by investigating the potential of applying deep learning techniques to this field. In the proposed method, simultaneous detection of the centers of the fovea and the optic disc (OD) from color fundus images is considered as a regression problem. A deep multiscale sequential convolutional neural network (CNN) is designed and trained. The publically available MESSIDOR and Kaggle datasets are used to train the network and evaluate its performance. The centers of the fovea and the OD in each image were marked by expert graders as the ground truth. The proposed method achieves an accuracy of 97%, 96.7% for the detection of the OD center and 96.6%, 95.6% for the detection of the foveal center of the MESSIDOR and Kaggle test sets respectively. Our promising results demonstrate the excellent performance of the proposed CNNs in simultaneously detecting the centers of both the fovea and OD without human intervention or handcrafted features. Moreover, we can localize the landmarks of an image in 0.007s. This approach could be used as a crucial part of automated diagnosis systems for better management of eye disease.",
keywords = "Convlutional neural networks, Diabetes, Fovea detection, Optic disc detection",
author = "Baidaa Al-Bander and Waleed Al-Nuaimy and Williams, {Bryan M.} and Yalin Zheng",
year = "2018",
month = feb,
day = "1",
doi = "10.1016/j.bspc.2017.09.008",
language = "English",
volume = "40",
pages = "91--101",
journal = "Biomedical Signal Processing and Control",
issn = "1746-8094",
publisher = "Elsevier BV",

}

RIS

TY - JOUR

T1 - Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc

AU - Al-Bander, Baidaa

AU - Al-Nuaimy, Waleed

AU - Williams, Bryan M.

AU - Zheng, Yalin

PY - 2018/2/1

Y1 - 2018/2/1

N2 - Detecting the locations of the optic disc and fovea is a crucial task towards developing automatic diagnosis and screening tools for retinal disease. We propose to address this challenging problem by investigating the potential of applying deep learning techniques to this field. In the proposed method, simultaneous detection of the centers of the fovea and the optic disc (OD) from color fundus images is considered as a regression problem. A deep multiscale sequential convolutional neural network (CNN) is designed and trained. The publically available MESSIDOR and Kaggle datasets are used to train the network and evaluate its performance. The centers of the fovea and the OD in each image were marked by expert graders as the ground truth. The proposed method achieves an accuracy of 97%, 96.7% for the detection of the OD center and 96.6%, 95.6% for the detection of the foveal center of the MESSIDOR and Kaggle test sets respectively. Our promising results demonstrate the excellent performance of the proposed CNNs in simultaneously detecting the centers of both the fovea and OD without human intervention or handcrafted features. Moreover, we can localize the landmarks of an image in 0.007s. This approach could be used as a crucial part of automated diagnosis systems for better management of eye disease.

AB - Detecting the locations of the optic disc and fovea is a crucial task towards developing automatic diagnosis and screening tools for retinal disease. We propose to address this challenging problem by investigating the potential of applying deep learning techniques to this field. In the proposed method, simultaneous detection of the centers of the fovea and the optic disc (OD) from color fundus images is considered as a regression problem. A deep multiscale sequential convolutional neural network (CNN) is designed and trained. The publically available MESSIDOR and Kaggle datasets are used to train the network and evaluate its performance. The centers of the fovea and the OD in each image were marked by expert graders as the ground truth. The proposed method achieves an accuracy of 97%, 96.7% for the detection of the OD center and 96.6%, 95.6% for the detection of the foveal center of the MESSIDOR and Kaggle test sets respectively. Our promising results demonstrate the excellent performance of the proposed CNNs in simultaneously detecting the centers of both the fovea and OD without human intervention or handcrafted features. Moreover, we can localize the landmarks of an image in 0.007s. This approach could be used as a crucial part of automated diagnosis systems for better management of eye disease.

KW - Convlutional neural networks

KW - Diabetes

KW - Fovea detection

KW - Optic disc detection

U2 - 10.1016/j.bspc.2017.09.008

DO - 10.1016/j.bspc.2017.09.008

M3 - Journal article

AN - SCOPUS:85029711683

VL - 40

SP - 91

EP - 101

JO - Biomedical Signal Processing and Control

JF - Biomedical Signal Processing and Control

SN - 1746-8094

ER -