Glaucoma detection in myopic eyes using deep learning autoencoder-based regions of interest.

Summary: Diagnosing glaucoma in people who are nearsighted (myopic) can be tricky. Doctors usually look at the thickness of the eye's nerve layer, but this standard measurement isn't always accurate for nearsighted eyes. In a new study, researchers tested a smart computer program—a "deep learning" AI—to look at 3D eye scans in a new way. They trained the AI on scans from both healthy eyes and eyes with glaucoma. This "dual AI" was much better at spotting the disease than the standard methods, scoring a 92% accuracy rating! This means doctors might soon have a powerful new tool to catch glaucoma early and save people's eyesight.