HANNOVER MESSE 2020, 20 - 24 April
switch to:
Artificial Intelligence

AI recognizes eye damage in diabetics

Adaptive algorithms help analyze scans and X-rays in ever more areas of healthcare.

01 Jan. 2018
Claudia Witte
American Academy of Ophthalmology
AI recognizes eye damage in diabetics (graphic: American Academy of Ophthalmology)

The Medical University of Vienna and the General Hospital (AKH) will begin using AI technology in its Diabetology outpatient clinic as of January 2018. Their team of physicians uses optical coherence tomography ( OCT ) to perform 40,000 retinal scans in 1.2 seconds, creating 65 million voxels, i.e. points in a three-dimensional grid. The OCT device used is about the size of a coffee machine and was developed at MedUni Vienna.

One of the long-term consequences of diabetes is blindness due to damaged blood vessels in the retina, symptoms that are often not recognized in the initial stage. A machine-learning procedure developed at the University of Vienna examines the OCT scans for corresponding indications and suggests a treatment method. The Head of the Department of Ophthalmology and Optometry, Dr. Ursula Schmidt-Erfurth, was awarded the J. Donald Glass Medal of the US Macula Society for this development.

The company contextflow , a spin-off of MedUni Vienna, also works with AI systems and intelligent image recognition , albeit in radiology. It uses machine learning to evaluate CT and MRI image data. The software is particularly useful in cases of doubt and, once an image area has been marked, finds visually similar comparison cases and relevant specialist journal articles.