A Review on Recent Work On OCT Image Classification for Disease Detection

Abstract

Optical Coherence Tomography (OCT) is a standard tool for cross-sectional imaging of retinal tissues. A large number of OCT scans are performed yearly, and ophthalmologists examine the OCT scans to detect eye diseases. Computer-assisted diagnosis (CAD) or decision support system is needed to screen OCT scans for disease detection. Earlier machine learning models were based on finding the most discriminative features by domain experts and then building the machine learning classifier. But recent research shows deep learning models are more suitable and give promising results. This paper summarizes the major deep learning approaches for OCT image classification for the detection of the most common retinal diseases, namely, AMD (Age-related Macular degeneration), DME (Diabetic Macular Edema), and CNV (Choroidal Neovascularization). We have discussed the advantages and challenges of deep learning models for OCT image classification, which can give directions for future research.