Exudate Regeneration for Automated Exudate Detection in Retinal Fundus Images

Muhammad Hussain; Hussain Al-Aqrabi; Muhammad Munawar; Richard Hill; Simon Parkinson

doi:10.1109/ACCESS.2022.3205738

Exudate Regeneration for Automated Exudate Detection in Retinal Fundus Images

Muhammad Hussain, Hussain Al-Aqrabi, Muhammad Munawar, Richard Hill, Simon Parkinson

Research output: Contribution to journal › Article › peer-review

15 Citations (Scopus)

Abstract

This paper presents a framework for the automated detection of Exudates, an early sign of Diabetic Retinopathy. The paper introduces a classification-extraction-superimposition (CES) mechanism for enabling the generation of representative exudate samples based on limited open-source samples. The paper demonstrates how the manipulation of Yolov5M output vector can be utilized for exudate extraction and super-imposition, segueing into the development of a custom CNN architecture focused on exudate classification in retinal based fundus images. The performance of the proposed architecture is compared with various state-of-the-art image classification architectures on a wide range of metrics, including the simulation of post deployment inference statistics. A self-label mechanism is presented, endorsing the high performance of the developed architecture, achieving 100% on the test dataset.

Original language	English
Article number	9885192
Number of pages	11
Journal	IEEE Access
DOIs	https://doi.org/10.1109/ACCESS.2022.3205738
Publication status	Published - 12 Sep 2022

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/ACCESS.2022.3205738Licence: CC BY

Cite this

@article{6d51b50f9a554f80b2003147f7db9be9,

title = "Exudate Regeneration for Automated Exudate Detection in Retinal Fundus Images",

abstract = "This paper presents a framework for the automated detection of Exudates, an early sign of Diabetic Retinopathy. The paper introduces a classification-extraction-superimposition (CES) mechanism for enabling the generation of representative exudate samples based on limited open-source samples. The paper demonstrates how the manipulation of Yolov5M output vector can be utilized for exudate extraction and super-imposition, segueing into the development of a custom CNN architecture focused on exudate classification in retinal based fundus images. The performance of the proposed architecture is compared with various state-of-the-art image classification architectures on a wide range of metrics, including the simulation of post deployment inference statistics. A self-label mechanism is presented, endorsing the high performance of the developed architecture, achieving 100% on the test dataset.",

keywords = "Exudate, Exudates detection, Self-labelling, Convolutional Neural Network (CNN), Data fusion, Convolutional Networks, Visualization, Optical imaging, Image segmentation, Data Fusion, Data integration, Training data, Computer architecture, Feature extraction, Diabetes, Convolutional neural networks, Exudates Detection, Labeling, Biomedical imaging",

author = "Muhammad Hussain and Hussain Al-Aqrabi and Muhammad Munawar and Richard Hill and Simon Parkinson",

note = "Publisher Copyright: Author",

year = "2022",

month = sep,

day = "12",

doi = "10.1109/ACCESS.2022.3205738",

language = "English",

journal = "IEEE Access",

issn = "2169-3536",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Exudate Regeneration for Automated Exudate Detection in Retinal Fundus Images

AU - Hussain, Muhammad

AU - Al-Aqrabi, Hussain

AU - Munawar, Muhammad

AU - Hill, Richard

AU - Parkinson, Simon

N1 - Publisher Copyright: Author

PY - 2022/9/12

Y1 - 2022/9/12

N2 - This paper presents a framework for the automated detection of Exudates, an early sign of Diabetic Retinopathy. The paper introduces a classification-extraction-superimposition (CES) mechanism for enabling the generation of representative exudate samples based on limited open-source samples. The paper demonstrates how the manipulation of Yolov5M output vector can be utilized for exudate extraction and super-imposition, segueing into the development of a custom CNN architecture focused on exudate classification in retinal based fundus images. The performance of the proposed architecture is compared with various state-of-the-art image classification architectures on a wide range of metrics, including the simulation of post deployment inference statistics. A self-label mechanism is presented, endorsing the high performance of the developed architecture, achieving 100% on the test dataset.

AB - This paper presents a framework for the automated detection of Exudates, an early sign of Diabetic Retinopathy. The paper introduces a classification-extraction-superimposition (CES) mechanism for enabling the generation of representative exudate samples based on limited open-source samples. The paper demonstrates how the manipulation of Yolov5M output vector can be utilized for exudate extraction and super-imposition, segueing into the development of a custom CNN architecture focused on exudate classification in retinal based fundus images. The performance of the proposed architecture is compared with various state-of-the-art image classification architectures on a wide range of metrics, including the simulation of post deployment inference statistics. A self-label mechanism is presented, endorsing the high performance of the developed architecture, achieving 100% on the test dataset.

KW - Exudate

KW - Exudates detection

KW - Self-labelling

KW - Convolutional Neural Network (CNN)

KW - Data fusion

KW - Convolutional Networks

KW - Visualization

KW - Optical imaging

KW - Image segmentation

KW - Data Fusion

KW - Data integration

KW - Training data

KW - Computer architecture

KW - Feature extraction

KW - Diabetes

KW - Convolutional neural networks

KW - Exudates Detection

KW - Labeling

KW - Biomedical imaging

UR - http://www.scopus.com/inward/record.url?scp=85139406122&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2022.3205738

DO - 10.1109/ACCESS.2022.3205738

M3 - Article

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 9885192

ER -

Exudate Regeneration for Automated Exudate Detection in Retinal Fundus Images

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Feature Mapping for Rice Leaf Defect Detection Based on a Custom Convolutional Architecture

PV-CrackNet Architecture for Filter Induced Augmentation and Micro-Cracks Detection within a Photovoltaic Manufacturing Facility

Cite this