What will AI Ophthalmology v2.0 look like?

Abstract

It is fair to say that the field of artificial intelligence (AI) has moved from the 'gee whiz' phase to the 'wow' phase. In the 'gee whiz' phase, a computer could be programmed to play checkers as well as a human player. Jumping ahead to the 'wow' phase, convoluted neural networks (CNNs) facilitate the detection of clinically significant levels of glaucomatous optic neuropathy, retinopathy of prematurity, and diabetic retinopathy with precision that rivals and even exceeds that of skilled clinicians. CNNs are computer algorithms that are organized in ways that allow for layered processing of data with a hierarchical architecture that mimics the cerebral visual cortex. The FDA has approved two AI algorithms -IDx 1 and EyeArt 2 -for their ability to detect referable diabetic retinopathy as well as a reading center manned by experts using robust clinical standards. The new journal, Modeling and Artificial Intelligence in Ophthalmology, will chronicle AI Ophthalmology version 2.0. What will AI Ophthalmology v2.0 look like?

Ähnliche Arbeiten