Recent advances in corneal specular microscopy image analysis through artificial intelligence

Although conventional automated analysis of corneal specular microscopy images has historically been limited by reproducibility challenges in the presence of corneal guttae, recent advances in artificial intelligence (AI) have significantly enhanced its diagnostic potential in such cases. This review explores the integration of AI techniques for analyzing specular microscopy images, emphasizing the shift from classical to advanced AI methods. We highlight AI-based methodologies-supervised and unsupervised learning-that have significantly enhanced the accuracy of in vivo human corneal endothelium analysis. The paper also discusses the challenges in data collection, emphasizing ethical considerations and the need for high-quality datasets. Additionally, we explore novel AI-derived metrics and their implications in enhancing diagnostic precision, particularly in Fuchs endothelial corneal dystrophy. The review concludes with insights into the future direction of AI in specular microscopy, highlighting its increasing relevance in ocular healthcare and the potential to overcome longstanding limitations in the field.