A Novel Fusion deep Learning Approach for Glaucoma Severity Classification – An Optical Coherence Tomography-Based Bimodal Artificial Intelligence Model

Abstract

Abstract From the inception of artificial intelligence (AI) in optical coherence tomography (OCT) for glaucoma diagnosis, the emphasis has always been on models focusing on visual data. The visual signals are extrapolated from the optic disc and retinal nerve fibre layer (RNFL) region and are processed visually. This manuscript introduces a novel bimodal AI model that combines visual data from OCT with numerical metrics such as RNFL thickness, vCDR, rim area, and cup volume. Furthermore, our unique model utilises an algorithm that goes one step further than existing models in the literature by providing multi-classification (normal vs mild/moderate vs severe) as opposed to binary classification.