FIM_Eikonal

Extension of SCI-Solver_Eikonal by Fu et al. for simulation of anisotropic wavefront propagation in the heart. See SCI-Solver_Eikonal/README.md for a description of the Fast Iterative Method, references to the detailed papers, and author information for the original code.

tetFIM uses the fast iterative method (FIM) on tetrahedral meshes to solve the anisotropic eikonal equation

with at the seed points.
is a 3x3 matrix encoding the speed information.
The result are activation times at the nodes of the mesh.
If speed = 1 and anisotropy = 1, then and represents the geodesic distance with respect to the seed points.

Syntax:

tetFIM -i inFile.vtu -p seedsFile.txt -o outFile.txt {additional options}

Available options:

-i  or -inFile              VTK unstructured grid dataset (.vtu or .vtk)
-p  or -seedsFile           .txt
-o  or -outFile             .bin or .txt
-a  or -anisotropy          default: 1.0
-s  or -speed               default: 1.0
-m  or -maxIterations       default: 2000
-b  or -maxBlocks           default: 16
-vb or -maxVertsPerBlock    default: 24
-v  or -verbose
-h  or -help

Hints:

• The seeds file is used to provide IDs of starting points (value zero in the solution). It may consist of multiple lines defining different 'seed configurations', each of which is solved individually. Within each line, point IDs have to be separated by a space.

• If the output file has the extension '.bin', the output will be binary, otherwise ascii. For binary, the first value will be the number of points as int32, followed by the values for each seed configuration as float. For ascii, there will be one column for each point in the mesh and one row for each seed configuration.

• Cell data arrays 'Phi' and 'Theta' containing angles in radian are required in the VTK dataset for anisotropy != 1.

• Cell data arrays 'Speed' and/or 'Anisotropy' may be provided in the VTK dataset to define different speeds / anisotropy ratios for each tetrahedron.

• The speed is defined with respect to the transversal direction, i.e. the effective speed in fiber direction is anisotropy * speed.

• Try to reduce maxVertsPerBlock if you get 'cudaCheckError() : invalid configuration argument'. This parameter determines maxNumTotalTets, which must be smaller than the max number of threads per block for your CUDA device (usually 512 or 1024).