A Hybrid Model for Early Prediction of Retinopathy Diseases based on Radial Basis Function Neural Networks with Multi Verse Optimization

Abstract

OCT is a non-invasive real-time, high-resolution imaging of highly scattering tissues and also a developing
biomedical imaging technology. It is broadly utilized in ophthalmology to perform diagnostic imaging on the
structure of the anterior eye and the retina. This paper aims to classify the optical coherence tomography
images to identify diseased retina image from a normal retina image. An automatic method of classification
technique for various retinal eye diseases like Macular Hole, Macular Edema, Vitreous Macular Traction,
Epiretinal Membrane and Age-related Macular Degeneration is proposed and validated in this paper. First
the feature extraction technique based on shape, texture and statistical features is implemented and the
feature values from the segmented OCT images are extracted for further classification purpose. Next the
classification is carried out using Radial Basis Function Neural Network (RBFNN). To improvise the
classification efficiency RBFNN combined with Multi Verse Optimization (MVO) is implemented to classify
the normal and abnormal OCT images. The result shows that the proposed method achieves a precision of
84.722%, recall of 84.722%, sensitivity of 84.722%, and specificity of 96.667% and finally the accuracy
achieved is 94.527%. The result shows that the proposed classification method achieves improvement in
classification accuracy over different retina diseases and in future it can be used to improve the
classification of other OCT related image diseases