Advancement in Diabetic Retinopathy Diagnosis Techniques: Automation and Assistive Tools
Ritu Sharma1, Bhawna Goyal1, *, Ayush Dogra2, 3
Identifiers and Pagination:Year: 2023
E-location ID: e187444002308080
Publisher ID: e187444002308080
Article History:Received Date: 18/11/2022
Revision Received Date: 02/05/2023
Acceptance Date: 26/06/2023
Electronic publication date: 04/09/2023
Collection year: 2023
open-access license: This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International Public License (CC-BY 4.0), a copy of which is available at: https://creativecommons.org/licenses/by/4.0/legalcode. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Diabetic retinopathy (DR) is a condition in which the retina of the human eye is impaired. If not properly treated, the condition may spread throughout the eye and result in irreversible blindness. There is a shortage of qualified ophthalmologists in developing nations, and there is very low knowledge of this condition. However, it is feasible to provide initial treatment to patients by integrating automated instruments and treating the illness before it reaches an incurable stat.
Identifying the various processes involved in identifying diabetic retinopathy, including preprocessing, feature extraction, segmentation, and classification. The work being discussed focuses on different detection approaches for diabetic retinopathy, specifically on CAD (Computer-Aided Diagnosis) models that are based on various AI (Artificial Intelligence) and image processing technologies.
Computer-aided diagnosis (CAD) systems are designed to assist medical professionals in the diagnosis of various diseases, including diabetic retinopathy. CAD systems use various algorithms and image processing techniques to analyze medical images and extract relevant features for classification. By using AI and image processing technologies, CAD systems can help to automate and standardize the diagnostic process, improving accuracy and efficiency.
To analyze the different strategies for detecting diabetic retinopathy, reputable publications and conferences are mined for research materials. The papers will be from the most recent years and will be categorized depending on the methods employed, such as machine learning, deep learning, etc. Each study is evaluated based on its technique, findings, dataset, and pros and cons.
The data shows that journals account for most of the work in this study (51%), with conferences accounting for 40% of the work and book chapters accounting for 9%. In addition, the data depicts a year-by-year study of work relevant to diabetic retinopathy detection. The major data is available on Google Scholar compared to Elsevier and Science Direct for research precision. Google Scholar has 60% data, Elsevier has approximately 10%, and Science Direct has approx 30% data on diabetic retinopathy detection.
The conclusion suggests that to achieve high accuracy in detecting diabetic retinopathy, hybrid techniques should be developed in the future. This may involve combining multiple approaches, such as machine learning algorithms, image processing techniques, and human expertise, to create a more comprehensive and accurate diagnostic tool. By using a hybrid approach, it may be possible to overcome individual techniques' limitations and improve the accuracy of diabetic retinopathy detection. Further research and development in this area may lead to more effective screening and treatment for diabetic retinopathy, ultimately improving patient outcomes.