Segmentation, processing and advanced visualization of atrial data from 4D-flow MRI
4D-flow MR images of the heart are emerging as a revolutionary way to visualize and understand blood flow patterns in normal and abnormal cases. Most of research on 4D-flow MRI have been focused on aortic applications. Therefore, there is a need to adapt the current protocols and data processing/visualization tools to work in other interesting sub-structures of the heart such as the left atria. The main goal of this project is to develop new image processing techniques, based on deep learning, to segment the left atria from 4D-flow MRI acquisitions. Transfer learning algorithms to cope with scanner differences will also be studied. In addition, advanced techniques will be required to visualize and characterize the resulting blood flow patterns. This project will be performed in collaboration with Hospital Clínic de Barcelona and Hospital de la Vall d'Hebron.
Supervisors: Oscar Camara, Xabier Morales