Applications of Artificial Intelligence in Vision Sciences

Abstract

Relevance: To analyze the current impact of artificial intelligence in vision sciences, highlighting its utility in the diagnosis and classification of ocular pathologies.

Abstract: The purpose of this study was to evaluate, through a bibliographic search, the applications of artificial intelligence in vision sciences, identifying its most relevant contributions to diagnostic processes and its current limitations. A literature search was conducted in the PubMed database following the PRISMA strategy, including scientific articles published over the last twenty years. A total of 431 articles were initially identified, of which 40 met the inclusion criteria after the selection process and were finally analyzed.

The analysis of the selected studies indicates that, in certain ocular pathologies, artificial intelligence represents a significant advancement in both diagnosis and classification, as demonstrated by the high AUC, sensitivity, and specificity values reported for various deep learning models. These findings suggest that artificial intelligence may serve as a valuable supportive tool for clinicians, contributing to improved diagnostic accuracy, enhanced patient quality of life, and a reduced risk of blindness.

However, despite its promising potential, further research is required incorporating more diverse clinical datasets and a larger volume of multicategorical disease images, in order to bring model performance closer to real-world clinical practice conditions.