Mediwhale Presents Retinal AI Technology at International Conference

[Financial News] Medical AI specialist company Mediwhale announced on the 14th that it will host a machine learning model development competition and present a total of 3 research abstracts at the International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2025), the world's most prestigious medical imaging conference.
Among these, 2 papers were selected for the 'CLINCCAI' program, which only allows research where a clinician is the lead author, proving the clinical value of the research.
MICCAI 2025 is a representative academic event where medical imaging experts from around the world participate, and it will be held in Daejeon from September 23 to 27.
At this MICCAI, Mediwhale will host the 'MUCARD' competition based on its own multi-center retinal data. The purpose of this competition is to develop a machine learning model that overcomes the differences in quality and format between retinal images taken with cameras from various manufacturers and demonstrates consistent performance.
Through this competition, researchers worldwide will experiment with the generalization performance in various camera environments using Mediwhale's data. Mediwhale plans to share the results and know-how of this research and seek future collaborative research opportunities.
Mediwhale will disclose research on △Introduction of personalized retinal AI prediction model △Presentation of the basis for Dr. Noon CVD judgment △Analysis of the correlation between image clarity and prediction accuracy.
The first abstract introduces the basic AI model 'PRETI' that enables personalized retinal image analysis by reflecting basic information such as the patient's age and gender. The second abstract quantitatively analyzes the basis on which Mediwhale's cardiovascular disease prediction solution 'Dr. Noon CVD', currently used in hospitals, predicts risk using the 'SHAP' technique. The third abstract analyzes that the image enhancement technique, which clarifies blurry retinal images using deep learning technology, can distort key health indicators such as the patient's age and gender that can be inferred from the retina. 
Taegeun Choi, CEO of Mediwhale, said, "It is very meaningful that Mediwhale directly hosts challenges and receives recognition for various research achievements at the world-renowned academic conference MICCAI," adding, "In order to create AI that medical professionals can trust and use, we will continue research and development to ensure not only the accuracy of the technology but also its use in more medical settings."
Mediwhale is the first in the world to develop and commercialize a medical AI solution that can prevent cardiovascular diseases with a simple eye exam. It is preparing to enter the US market with the goal of obtaining De Novo approval from the US Food and Drug Administration (FDA) in 2026.
vrdw88@fnnews.com Kang Joongmo Reporter