Accelerating retinal fundus image classification using artificial neural networks (ANNs) and reconfigurable hardware (FPGA)

Arfan Ghani, Chan H. See, V Sudhakaran, J Ahmad, Raed Abd-Alhameed

Research output: Contribution to journalArticle

2 Downloads (Pure)

Abstract

Diabetic retinopathy (DR) and glaucoma are common eye diseases that aect a blood vessel in the retina and are two of the leading causes of vision loss around the world. Glaucoma is a common eye condition where the optic nerve that connects the eye to the brain becomes damaged, whereas DR is a complication of diabetes caused by high blood sugar levels damaging the back of the eye. In order to produce an accurate and early diagnosis, an extremely high number of retinal images needs to be processed. Given the required computational complexity of image processing algorithms and the need for high-performance architectures, this paper proposes and demonstrates the use of fully parallel field programmable gate arrays (FPGAs) to overcome the burden of real-time
computing in conventional software architectures. The experimental results achieved through software implementation were validated on an FPGA device. The results showed a remarkable improvement in terms of computational speed and power consumption. This paper presents various preprocessing methods to analyse fundus images, which can serve as a diagnostic tool for detection of glaucoma and diabetic retinopathy. In the proposed adaptive thresholding-based preprocessing method, features were selected by calculating the area of the segmented optic disk, which was further classified using a feedforward neural network (NN). The analysis was carried out using feature extraction
through existing methodologies such as adaptive thresholding, histogram and wavelet transform. Results obtained through these methods were quantified to obtain optimum performance in terms of classification accuracy. The proposed hardware implementation outperforms existing methods and oers a significant improvement in terms of computational speed and power consumption.
Original languageEnglish
Article number1522
Number of pages17
JournalElectronics (Switzerland)
Volume8
Issue number12
Early online date11 Dec 2019
DOIs
Publication statusE-pub ahead of print - 11 Dec 2019

Fingerprint

Reconfigurable hardware
Image classification
Field programmable gate arrays (FPGA)
Optics
Electric power utilization
Neural networks
Feedforward neural networks
Blood vessels
Software architecture
Medical problems
Sugars
Wavelet transforms
Computational complexity
Brain
Image processing
Blood
Hardware

Bibliographical note

This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited

Keywords

  • Neural network
  • Machine learning
  • Glaucoma
  • Diabetic Retinopathy
  • Adaptive Threshold
  • FPGA
  • loTs
  • Smart healthcare

Cite this

Accelerating retinal fundus image classification using artificial neural networks (ANNs) and reconfigurable hardware (FPGA). / Ghani, Arfan; See, Chan H.; Sudhakaran, V; Ahmad, J; Abd-Alhameed, Raed.

In: Electronics (Switzerland), Vol. 8, No. 12, 1522, 11.12.2019.

Research output: Contribution to journalArticle

Ghani, Arfan ; See, Chan H. ; Sudhakaran, V ; Ahmad, J ; Abd-Alhameed, Raed. / Accelerating retinal fundus image classification using artificial neural networks (ANNs) and reconfigurable hardware (FPGA). In: Electronics (Switzerland). 2019 ; Vol. 8, No. 12.
@article{e40dc371fcbc479a9cb6b84baa0b5147,
title = "Accelerating retinal fundus image classification using artificial neural networks (ANNs) and reconfigurable hardware (FPGA)",
abstract = "Diabetic retinopathy (DR) and glaucoma are common eye diseases that aect a blood vessel in the retina and are two of the leading causes of vision loss around the world. Glaucoma is a common eye condition where the optic nerve that connects the eye to the brain becomes damaged, whereas DR is a complication of diabetes caused by high blood sugar levels damaging the back of the eye. In order to produce an accurate and early diagnosis, an extremely high number of retinal images needs to be processed. Given the required computational complexity of image processing algorithms and the need for high-performance architectures, this paper proposes and demonstrates the use of fully parallel field programmable gate arrays (FPGAs) to overcome the burden of real-timecomputing in conventional software architectures. The experimental results achieved through software implementation were validated on an FPGA device. The results showed a remarkable improvement in terms of computational speed and power consumption. This paper presents various preprocessing methods to analyse fundus images, which can serve as a diagnostic tool for detection of glaucoma and diabetic retinopathy. In the proposed adaptive thresholding-based preprocessing method, features were selected by calculating the area of the segmented optic disk, which was further classified using a feedforward neural network (NN). The analysis was carried out using feature extractionthrough existing methodologies such as adaptive thresholding, histogram and wavelet transform. Results obtained through these methods were quantified to obtain optimum performance in terms of classification accuracy. The proposed hardware implementation outperforms existing methods and oers a significant improvement in terms of computational speed and power consumption.",
keywords = "Neural network, Machine learning, Glaucoma, Diabetic Retinopathy, Adaptive Threshold, FPGA, loTs, Smart healthcare",
author = "Arfan Ghani and See, {Chan H.} and V Sudhakaran and J Ahmad and Raed Abd-Alhameed",
note = "This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited",
year = "2019",
month = "12",
day = "11",
doi = "10.3390/electronics8121522",
language = "English",
volume = "8",
journal = "Electronics (Switzerland)",
issn = "2079-9292",
publisher = "MDPI",
number = "12",

}

TY - JOUR

T1 - Accelerating retinal fundus image classification using artificial neural networks (ANNs) and reconfigurable hardware (FPGA)

AU - Ghani, Arfan

AU - See, Chan H.

AU - Sudhakaran, V

AU - Ahmad, J

AU - Abd-Alhameed, Raed

N1 - This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited

PY - 2019/12/11

Y1 - 2019/12/11

N2 - Diabetic retinopathy (DR) and glaucoma are common eye diseases that aect a blood vessel in the retina and are two of the leading causes of vision loss around the world. Glaucoma is a common eye condition where the optic nerve that connects the eye to the brain becomes damaged, whereas DR is a complication of diabetes caused by high blood sugar levels damaging the back of the eye. In order to produce an accurate and early diagnosis, an extremely high number of retinal images needs to be processed. Given the required computational complexity of image processing algorithms and the need for high-performance architectures, this paper proposes and demonstrates the use of fully parallel field programmable gate arrays (FPGAs) to overcome the burden of real-timecomputing in conventional software architectures. The experimental results achieved through software implementation were validated on an FPGA device. The results showed a remarkable improvement in terms of computational speed and power consumption. This paper presents various preprocessing methods to analyse fundus images, which can serve as a diagnostic tool for detection of glaucoma and diabetic retinopathy. In the proposed adaptive thresholding-based preprocessing method, features were selected by calculating the area of the segmented optic disk, which was further classified using a feedforward neural network (NN). The analysis was carried out using feature extractionthrough existing methodologies such as adaptive thresholding, histogram and wavelet transform. Results obtained through these methods were quantified to obtain optimum performance in terms of classification accuracy. The proposed hardware implementation outperforms existing methods and oers a significant improvement in terms of computational speed and power consumption.

AB - Diabetic retinopathy (DR) and glaucoma are common eye diseases that aect a blood vessel in the retina and are two of the leading causes of vision loss around the world. Glaucoma is a common eye condition where the optic nerve that connects the eye to the brain becomes damaged, whereas DR is a complication of diabetes caused by high blood sugar levels damaging the back of the eye. In order to produce an accurate and early diagnosis, an extremely high number of retinal images needs to be processed. Given the required computational complexity of image processing algorithms and the need for high-performance architectures, this paper proposes and demonstrates the use of fully parallel field programmable gate arrays (FPGAs) to overcome the burden of real-timecomputing in conventional software architectures. The experimental results achieved through software implementation were validated on an FPGA device. The results showed a remarkable improvement in terms of computational speed and power consumption. This paper presents various preprocessing methods to analyse fundus images, which can serve as a diagnostic tool for detection of glaucoma and diabetic retinopathy. In the proposed adaptive thresholding-based preprocessing method, features were selected by calculating the area of the segmented optic disk, which was further classified using a feedforward neural network (NN). The analysis was carried out using feature extractionthrough existing methodologies such as adaptive thresholding, histogram and wavelet transform. Results obtained through these methods were quantified to obtain optimum performance in terms of classification accuracy. The proposed hardware implementation outperforms existing methods and oers a significant improvement in terms of computational speed and power consumption.

KW - Neural network

KW - Machine learning

KW - Glaucoma

KW - Diabetic Retinopathy

KW - Adaptive Threshold

KW - FPGA

KW - loTs

KW - Smart healthcare

UR - http://www.scopus.com/inward/record.url?scp=85076766230&partnerID=8YFLogxK

U2 - 10.3390/electronics8121522

DO - 10.3390/electronics8121522

M3 - Article

VL - 8

JO - Electronics (Switzerland)

JF - Electronics (Switzerland)

SN - 2079-9292

IS - 12

M1 - 1522

ER -