Exploring the Potentials of Explainable AI for Early Cataract Detection to Foster Accessible Healthcare

Explainable AI (XAI) aims to provide transparency and comprehensibility in ML models’ decision-making processes, enabling users to understand why a model makes specific predictions or decisions. This transparency feature is crucial for building trust in Ai systems, especially in high-stake applications such as healthcare, finance, and autonomous vehicles etc. In this research article, we aim to explore the capabilities of explainable AI for one of the critical use cases in the domain of ophthalmology for cataract prediction. Cataract is a common yet prevalent eye disease which is affecting millions of individuals across the world. It is essential to detect and treat this disease early to prevent vision loss and enhance the quality of life. By advancing diagnosis accuracy and efficiency, artificial intelligence (AI) has the potential to detect cataract based on the fundus image of an eye and other possible parameters. This can act as a decision support system for an ophthalmologist or general medical practitioner, especially in remote areas which are lacking high-end scientific equipment. However, the opacity and complexity of advanced AI models make them difficult to adapt such systems in actual clinical practice. Explainable AI (XAI) techniques if applied to such advanced ML models can address this issue by allowing healthcare practitioners to comprehend the decision-making process of AI models. This article reviews the critical explainable AI techniques and their possible application for one of the use cases – cataract detection.