PROJECT TEAM
Jakob Grauslund, Leadin investigator of IMPRESS, Jan Erik Henriksen, Maganemnet support, Kurt Højlund, Clinical lead in diabetes), Kristina Lyngsø, project koordinator, Dr. Sebastian Dinesen, developer, Dr. Simon Lowater, Developer, Dr. Lars Skollerud, Professor Thiusiius Savarimuthu, Thecnical lead of IMPRESS, Peter Ratgen, software designer, Andrea Núñez, software developer, Morten La Cours, Managment support, Javad Hajarim, Clinical Lead RH
THE NEED
Denmark is a leader in national diabetic retinopathy screening, but current programs face challenges. Manual grading is resource-intensive, the ICDR scale only reflects the most severe lesion, and no formal training is required. Moreover, tracking disease progression is difficult due to a lack of tools to mark and follow specific lesions. This highlights the need for decision support tools to improve efficiency and diagnostic accuracy.
THE SOLUTION
The solution NOAH, is an autonomous screening algorithm based on artificial intelligence that streamlines and enhances eye screening for diabetic retinopathy. The system analyzes a real-time image of the eye, accurately identifies the location of lesions, and automatically classifies disease severity. This frees up valuable time and resources for healthcare professionals, allowing them to focus more on patient care—while significantly improving diagnostic accuracy.
IMPRESS - Odense University Hospital
Call 3 - 2022 | Call 6 - 2024
500.000 DKK
Clinical Area
Diabetic retinopathy
Technology
AI Decision Support Tool
PROJECT SUMMARY
NOAH is an autonomous AI-based DR screening tool that analyzes retinal images in real time, accurately segments lesions, and classifies disease severity. It improves diagnostic precision and frees up clinical resources for patient care
CLINICAL IMPACT
The solution improves diagnostic accuracy and reduces the risk of overlooked diabetic eye lesions, leading to earlier detection and better patient outcomes. At the same time, NOAH significantly eases the DR grading burden on ophthalmologists by enabling task delegation to certified staff. This allows for more efficient resource use and supports a more sustainable and scalable diagnostic workflow.