Cardiac Segmentation on DE-MRI with deep learning-based tools Cardiac Segmentation on DE-MRI with deep learning-based tools

Catheter ablation of ventricular tachycardia (VT) currently has an important role in the treatment of ventricular tachycardia. In the last years, delayed-enhanced cardiac magnetic resonance (DE-MRI) has emerged as the gold standard imaging technique for defining myocardial scar. The first step for processing the DE-MRI image consists on manually segmenting the left ventricle. However, this manual step is time-consuming and is subject to intra- and inter-observer variability. This project aims at developing an automatic cardiac segmentation method for DE-MRI images with deep learning and/or adversarial techniques. The project will be mainly carried out at Galgo Medical / Adas3D offices and will be done in collaboration with several local hospitals and the Universitat Pompeu Fabra. At the end of the project, a software prototype written in C++ will have to be delivered, together with an evaluation of its accuracy.

Supervisor: Lluís Serra (Adas3D), Xavier Planes (Adas3D), Oscar Camara