These scripts provide a method to fuse information from multipe imaging sequences to build a patient-specific finite-element (FE) model of the left ventricle (LV). Description of this work was published in "Open-Source Routines for Building Personalized Left Ventricular Models from Cardiac MRI Data".
This project provides a generic pipeline for mapping imaging data onto a LV FE mesh using a prolate, spheroidal coordinate system. Specifically, there are two scripts that follow the same pipeline to project scar and mechanical activation information onto a FE mesh.
The data files needed to run the example scripts are uploaded on SimTK.
The scripts for mapping are hosted here on github.
The MATLAB data fusion routine uses the built-in scatteredInterpolant function to interpolate data from imaging across the LV.
- LVMAP_Shell.m projects scar from LGE MRI onto the LV, FE mesh
- LVMAP_Shell_DENSE.m projects mechanical activation from DENSE MRI onto the LV FE mesh
The Python (3.0+) data fusion routine uses the scipy radial basis function to interpolate data from imaging across the LV. While RBF is a different method from scatteredInterpolant, they both approximate similar results when projecting the scar from LGE MRI.
For any questions regarding the library you can contact TK at phung@virginia.edu