Latent Diffusion Models For Cardiac Attribute Regularization

Abstract:

Diffusion models have recently caught the attention of the medical imaging community by producing realistic synthetic images. Recent efforts have focused on improving model controllability of the generation process by allowing selective modifications of data attributes, such as altering the gender of a person in an image. Latent Diffusion Models (LDMs) can be used to generate realistic data of brain MRI controlled by attributes such as age, sex, and brain structure volumes.

This project aims to use latent diffusion models to generate realistic cardiac MRI and control the generation process by given attributes such as age, cardiac volumes, etc.

Maxime Di Folco
Maxime Di Folco
Research Scientist

My research interest is the study of the cardiac function via machine learning methods, in particular representation learning methods that aim to acquire low dimensional representation of high dimensional data. I have a strong interest in cardiac remodelling (adaptation of the heart to its environment or a disease), notably the study of the deformation and shape aspects.