Template matching with phase dependent multiple templates for 3D tracking using cine MRI
Luis Fernando Paredes (PUC, DKFZ, U. Heidelberg)
Martes 12 de abril de 2016, 15:30, auditorio Prof. Jorge Krause
The aim of this work is to increase the localization accuracy of template matching tracking algorithms for 3D localization of structures using 2D real time cine MR images. The main objective is to investigate the feasibility and implementation of tracking with phase-dependent multiple templates extracted from parallel cine MRI images (or 4D data generation) for overcome the out of plane motion and target shape changes influenced by the respiratory motion in liver. An important key point is the use of the vessels like reliable landmarks due its excellent soft tissue contrast in the T2 weighted cine MRI. The result of this work is 3D target localization estimated from incoming 2D real-time cine MR images using 4D information as training data, without additional dose to the patient and non invasive technique. Finally the idea is test this kind of tracking algorithms for the incoming MRI-Linac to the DKFZ facility.