A Critical Insight into Pragmatic Manifestation of Diabetic Retinopathy Grading and Detection
Muhammad Samer Sallam1, Rashidah Funke Olanrewaju2, Ani Liza Asnawi3
1Muhammad Samer Sallam*, Department of Computer and Information Engineering, International Islamic University, Kuala Lumpur, Malaysia.
2Rashidah Funke Olanrewaju, Department of Computer and Information Engineering, International Islamic University, Kuala Lumpur, Malaysia.
3Ani Liza Asnawi, Department of Computer and Information Engineering, International Islamic University, Kuala Lumpur, Malaysia.
Manuscript received on November 15, 2019. | Revised Manuscript received on 20 November, 2019. | Manuscript published on December 10, 2019. | PP: 2040-2049 | Volume-9 Issue-2, December 2019. | Retrieval Number: B8001129219/2019©BEIESP | DOI: 10.35940/ijitee.B8001.129219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Nowadays, artificial intelligence applications invade all of the fields including medical applications field. Deep learning, a subfield of artificial intelligence, in particular, Convolutional Neural Networks (CNN), have quickly become the first choice for processing and analyzing medical images due to its performance and effectiveness. Diabetic retinopathy is a vision loss disease that infects people with diabetes. This disease damages the blood vessels in the retina, hence, leads to blindness. Due to the sensitivity and complications involved in managing diabetics, designing and developing automated systems to detect and grade diabetic retinopathy is considered one of the recent research areas in the world of medical image applications. In this paper, the aspects of deep learning field related to diabetic retinopathy have been discussed. Various concepts in deep learning including traditional Artificial Neural Network (ANN) algorithm, ANN drawbacks in context of computer vision and image processing applications, and the best algorithm to overcome ANN drawbacks, CNN, have been elucidated along with the architecture. The paper also reviews an extensive summary of some works in the current research trend and future applications of the DL algorithms in medical image analysis for DR detection and grading. Furthermore, various research gabs related to building such automated systems for medical image analysis have been conferred – such as imbalance dataset which is considered one of the main performance issues that should be handled, the need of high performance computational resources to train deep and efficient models and others. This is quite beneficial for researchers working in the domain of medical image analysis to handle DR.
Keywords: Convolutional Neural Networks, Retinal Fundus Images Classification, Diabetic Retinopathy
Scope of the Article: Classification