MultipleShapeCorresponder_x64.exe $startIdx$ $stopIdx$ $fileExt$ $algorithmID$ $N$ $denseExtension$ $interpCoarseToDense$

$fileExt$	= input mesh file extension: obj, off, ply supported

$algorithmID$	= 1 to execute the optimal exponential algorithm (# shapes better be <9)
		= 2 to execute the approximation algorithm based on shortest paths (sequential algorithm as referred in the paper) 
		= 3 to execute the better approximation algorithm based on dynamic programming (our proposed solution in citation below)

e.g., MultipleShapeCorresponder_x64.exe 0 7 off 3 9 0 1
command above matches N=9 samples of meshes 0 through 7 using our original algorithm and interpolate the resulting coarse map to a denser one

e.g., MultipleShapeCorresponder_x64.exe 0 7 off 3 9 0 0
command above matches N=9 samples of meshes 0 through 7 using our original algorithm without any further interpolation

e.g., MultipleShapeCorresponder_x64.exe 0 7 off 3 80 1 0
command above matches N=80 samples of meshes 0 through 7 using our dense matching extension to our original algorithm; interpolation here does not make sense 'cos enumeration is already dense


so, put your input meshes in \input and make an \output folder ready for resutls to go


you can also compare initial maps by combinatorial matching (our original algo) vs. by spectral matching (dense extension) at coarse levels such as N=9 to compare the two initialization/graph construction


code is developed and executed on win7 yet is pretty portable except the #include <windows.h> for timing which is commented-out in the public release so shouldn't be a problem


citation:
Y. Sahillioglu and Y. Yemez, Multiple Shape Correspondence by Dynamic Programming, Computer Graphics Forum (Proc. PG), 2014.



--ysf
