Hi everyone. I know this repository is a bit old but I found the code to be really useful none the less.
I have found the following issues, though:
-
When loading a off or ply file generated from a digital elevation model with a straightforward triangulation, the pairs.json file in output is always empty. It seems to me that there's a trivial error: lines 112,113 and 114 of mesh.h can be fixed by changing vertices[i][1] to vertices[i][2]. If I do that I seem to obtain reasonable, non-empty results. Your README mentions that the filtration should be auto-generated from height values. It seems to me that the code works as intended only after applying this fix.
-
When I start from the same digital elevation model and I generate the scalar field as a separate file (which I was trying to do before finding the previous fix), something goes awry. The pairs.json output is not empty, but it contains just a single h0 pair, an obviously wrong result. The input is as follow: instead of having vertex coords (x,y,x) in the list of vertices, I generate coordinates (x,y,0.0) for all vertices and I put all the z-values in a separate .dat file. I then run computePers with two arguments: ./computePers mesh.off values.dat.
I can provide example files if needed.
Hi everyone. I know this repository is a bit old but I found the code to be really useful none the less.
I have found the following issues, though:
When loading a off or ply file generated from a digital elevation model with a straightforward triangulation, the pairs.json file in output is always empty. It seems to me that there's a trivial error: lines 112,113 and 114 of
mesh.hcan be fixed by changingvertices[i][1]tovertices[i][2]. If I do that I seem to obtain reasonable, non-empty results. Your README mentions that the filtration should be auto-generated from height values. It seems to me that the code works as intended only after applying this fix.When I start from the same digital elevation model and I generate the scalar field as a separate file (which I was trying to do before finding the previous fix), something goes awry. The pairs.json output is not empty, but it contains just a single h0 pair, an obviously wrong result. The input is as follow: instead of having vertex coords
(x,y,x)in the list of vertices, I generate coordinates(x,y,0.0)for all vertices and I put all the z-values in a separate.datfile. I then runcomputePerswith two arguments:./computePers mesh.off values.dat.I can provide example files if needed.