Severity Grading and Early Retinopathy Lesion Detection through Hybrid Inception-ResNet Architecture.

Sana Yasin,Klaudia Proniewska,Samar Alqhtani,Adam Glowacz,Frantisek Brumercik,Tariq Ali,Ali Alqahtani,Umar Draz,Lukasz Wzorek,Abdul Rehman,Muhammad Irfan,Nasrullah Iqbal

doi:10.3390/s21206933

Abstract

Diabetic retinopathy (DR) is a diabetes disorder that disturbs human vision. It starts due to the damage in the light-sensitive tissues of blood vessels at the retina. In the beginning, DR may show no symptoms or only slight vision issues, but in the long run, it could be a permanent source of impaired vision, simply known as blindness in the advanced as well as in developing nations. This could be prevented if DR is identified early enough, but it can be challenging as we know the disease frequently shows rare signs until it is too late to deliver an effective cure. In our work, we recommend a framework for severity grading and early DR detection through hybrid deep learning Inception-ResNet architecture with smart data preprocessing. Our proposed method is composed of three steps. Firstly, the retinal images are preprocessed with the help of augmentation and intensity normalization. Secondly, the preprocessed images are given to the hybrid Inception-ResNet architecture to extract the vector image features for the categorization of different stages. Lastly, to identify DR and decide its stage (e.g., mild DR, moderate DR, severe DR, or proliferative DR), a classification step is used. The studies and trials have to reveal suitable outcomes when equated with some other previously deployed approaches. However, there are specific constraints in our study that are also discussed and we suggest methods to enhance further research in this field.

Highlights

IntroductionDetection can avert the danger of vision loss
In order to examine the usefulness of every preprocessing step, i.e., normalization of data in the first place augmentation of data and lastly the balancing of data in the suggested network, results of the network are matched for the initial classification phase, among the six categories:
Including all the anticipated steps increases the accuracy of the proposed model from classification accuracy = 69.34% to classification = 88.10%; these outcomes express the effect of the anticipated steps of preprocessing in order to enhance the performance

Summary

Introduction

Detection can avert the danger of vision loss. Automated categorization of cardiovascular and ophthalmologic ofinfections ysis of fundus images has become a well-known exercise in the field of telemedic vious methods were composed of manual separation; it was tiresome, t suming, difficult, and skilled manpower isophthalmologic mandatory [1]. Co Automated categorization of cardiovascular and infections by other analysis of fundus irregularities realistic, impartial, a ofaided fundusidentification images has become a well-known exercise in is theeconomical, field of telemedicine

Objectives

Methods

Results

Conclusion