Automatic Classification of Healthy/AMD Class Retinal OCT Images with Deep-Learning: A Study
Sivakumar Rajendran
Abstract
Advancements in medical imaging have greatly increased the volume and complexity of data, often requiring expert interpretation and decision-making. However, the shortage of available experts in clinical settings can delay timely diagnosis. Automated diagnostic techniques can aid clinicians in accurately diagnosing and managing diseases. This work proposes a binary classification model using Deep Learning (DL) for classifying and analyzing the severity of Age-related Macular Degeneration (AMD) diseases in Optical Coherence Tomography (OCT) images. This model consists of three stages: binary classification, deep feature mining with selected algorithm and image collection and resizing. In this work, various algorithms are used for classification and Xception algorithm shows better result than other similar algorithms with the detected accuracy of 0.99 for classification.