Deep Learning Techniques On Fundus, B-Scan And Oct Imaging Modalities For Retinal Detachment Classification

Abstract

Retinal Detachment Is An Eye Disorder. It Detaches Itself From The Other Layers Of The Eye.
It Occurs Due To Loss Of Vision In The Eye. In Image Processing Applications, State-Of-TheArt-Technology Deep Learning In Artificial Intelligence Plays An Important Role In Biomedical
And Satellite Image Processing For Object Identification And Object Recognition. The Proposed
Method Is To Detect Retinal Abnormality By Using Computer-Aided Diagnosis From Advanced
Imaging Modalities Such As Optical Coherence Tomography (OCT) System, Fundus Image And
B-Scan System Have Emerged As A Powerful New Methodology. This Method Employs A Deep
Learning Convolutional Neural Network By Rdnet For Automated Abnormality Identification.
Deep Network Rdnet Architecture Is Applied To The Fundus, OCT And B-Scan Systems. This
Data Set Consists Of 2 Categories Including A Normal And Abnormal Image Of The Retina. The
Proposed Deep Learning Model Is Compared With Other Deep Learning Models Such As Resnet,
Inception-V3 On Fundus, OCT Systems. The Proposed Model Was Trained On The Agarwal Eye
Hospital’s Image Dataset Consists Of 1117 Images. The Proposed Method Achieves An
Accuracy Of 96%, The Precision Of 93%, Recall Of 1% And F1 Score Of 96%.