Detection of endometriosis lesions from MRI images



Main objectives

Develop AI software to detect and characterize endometriosis on pelvic MRI


Endometriosis is a disease that is still poorly known and poorly diagnosed. The average length of time before a patient is diagnosed is 7 years after the onset of the first symptoms and 75% of patients receive at least one erroneous diagnosis. Diagnostic errors are partly due to the difficulty of analyzing pelvic MRI images and the lack of specialized radiologists, who represent less than 1% of radiologists in France and are poorly distributed throughout the country.

Methodology/Technology used

The project employs a set of machine learning technologies to recognize and model the organs of the female pelvis and classify the different types of endometriosis lesions according to their location.



Elise Mekkaoui () & Raphaelle Taub ()


AP-HP, Inria

Useful links

Other projects



US Caractérisation et prédiction de la survenue de formes graves ou létales du COVID-19 à partir des données issues de l’EDS de l’AP-HP

Names of partners involved
AP-HP, Inria & Centrale Supélec

read more