•  
  •  
 

Applied Mathematics & Information Sciences

Author Country (or Countries)

India

Abstract

Retina is an important layer of tissue in the back of the eye. The retina’s primary job is to gather light that the lens has focused, convert the light into neural signals, and send those signals to the brain for optical compensation. This crucial tissue may appear damaged due to retinal detachment (RD). Such conditions can impair vision and undoubtedly have the potential to be severe enough to cause blindness. It can be challenging to identify the damage since the layers and the nerve connections are too delicate and thin, and they may be mistaken for another illness. IMRCNN, an improved mask recurrent convolutional neural network, was proposed to test its efficiency in RD detection. The effectiveness of the suggested approach is evaluated by looking at the implementation measures Sensitivity (S), Accuracy (A), Specificity (SP), and F-score (F). For 54,000 retinographic images, the average S, SP, A, and F values were 92.20%, 98%, 95.10%, and 93%, respectively. This suggested model performs better at classifying RD-related lesions and classifying the intensity levels on various retinal pictures. This type of analysis methodology concentrates on breaking down images into pixels and studying the data from the bottom up. It offers greater analysis and more precisely detects RD. This study presents essential knowledge and cutting-edge machine learning methodologies in the field of medical image processing methods and analysis. The main objectives of this work are to define and apply the identified and addressed important principles as well as to provide research on medical image processing.

Digital Object Identifier (DOI)

http://dx.doi.org/10.18576/amis/170301

Download

Share

COinS