valette/acvd

Fast simplification of 3D surface meshes

acvd's Issues

Feature Request: Output quads instead of triangles.

Converting input to coarse quads would be very useful.

ACVD crashes after closing second window

Windows 10, x64 VS2015, Debug

Windows crashes when running ACVD with display on.

repro:

Run ACVD
close first window that pops up
close second window

What is the process for getting a remerged mesh with display enabled. I can't find any docs that describe the steps.

Setting the variable 'Display' to 0 runs without crashing.

Example code usage with data

Thanks for sharing this nice code. I'm new to this tool. I found there are many tools here:

1. ACVD
2. ACVDQ
3. AnisotropicRemeshing
4. AnisotropicRemeshingQ
5. ManifoldSimplification
6. VolumeAnalysis
7. etc

Can you add few simple tutorials/examples with data in Readme.md to show how to use your codes for each of them? I think hand model and fandisk in your paper are good cases.

I found this Python implementation of ACVD is a good example.

https://github.com/pyvista/pyacvd

Clustered mesh vertices are not equal to input number of clusters sometimes

Dear author,

Sometimes I found that the specified number of clusters is not equal to the number of vertices in the output mesh. Is it something that is inherent in this method or is it a bug?

Thank you!

Xingtong

This application has requested the Runtime to terminate it in an unusual way.

I try to use ACVD to simplify a 500M obj,this is the command:

ACVD.exe 1.obj 70000 0

and this is the output message:

load : 1.obj
*****************************************************************************
Mesh with 6977920 polygons, 6977922 points, 13955840 edges
The mesh is made of 0 triangles, 6977920 quads and 0 other polygons
0 non-manifold edges and 0 boundary edges
The mesh has 1 connected components
 Valences entropy: 0.0056823
0 disconnected vertices, 6.97792e+006 connected vertices
99.9996 percent of irregular vertices
*****************************************************************************
Input mesh: 6977922 vertices    and     6977920 faces

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

Bug fix in ACVDQ

I just found a small bug while comping ACVDQ in windows.

// Line 246 @ ACVDQ.cxx
bool* fixed = new bool[mesh->GetNumberOfPoints()];
//bool fixed[ mesh->GetNumberOfPoints() ]; //Dynamic array should be used here
fixedVertices = vtkIdList::New();
int id, n = 0;

for ( int i = 0; i < mesh->GetNumberOfPoints(); i++ )
    fixed[ i ] = false;

while( input >> id ) {

    n++;
    vtkIdType v1, v2, v3;
    mesh->GetFaceVertices( id, v1, v2, v3 );
    fixed[ v1 ] = fixed[ v2 ] = fixed[ v3 ] = true;

}

for ( int i = 0; i < mesh->GetNumberOfPoints(); i++ )
    if ( fixed[ i ] ) fixedVertices->InsertNextId( i );

input.close();
cout << "Added " << n << " constraints on triangles" << endl;

delete[] fixed; //Free dynamic error
}

Can you check this issue and update your code?

Horse mesh output is not manifold

Horse model here:
https://www.dropbox.com/s/je2eor0c1dwt4ec/horse.ply?dl=0

acvdq horse.ply 200 0 -m 1 -sc 1000

Output:

Even 'acvdq 500 0 -m 1 -sc 1000' has issues on the legs.

Triangle generated from simplification face in opposite direction from input.

In ACVDQ, the winding order from the triangles in the simplified mesh appears to be backwards.

Can't compile code

Have you guys kept your code up to date with the latest releases of VTK? I am getting errors when trying to compile. I noticed the code was released in 2008, so I assume that it hasn't been updated along with changing vtk classes. Let me know if I am correct in assuming this.

Differents results between MacOS et Windows 10

Hello !

I use MacOS X Mojave 10.14.4 and Windows 10 up to date.
In my case, i need to use ACVD for deploy a complex deformable sphere.
I search in this github repository but i did not find anything...
Has anyone ever encountered any differences?

With the same input...

Windows ouput :

Mesh with 3040 polygons, 1522 points, 4560 edges
Bounding Box: [94.6748, 63.8625, -75.726]
[101.292, 70.7261, -66.4477]
The mesh is made only of 3040 triangles
0 non-manifold edges and 0 boundary edges
The mesh has 1 connected components
Valences entropy: 1.64355
0 disconnected vertices, 1522 connected vertices
46.6491 percent of irregular vertices
Mesh geometry qualitity:
AngleMin=24.1465
AverageMinAngle=46.4177
Qmin=0.437835
Qav=0.827767
P30=0.372807

Mac ouput :

Mesh with 3040 polygons, 1522 points, 4560 edges
Bounding Box: [94.6693, 63.8688, -75.7243]
[101.3, 70.7324, -66.4492]
The mesh is made only of 3040 triangles
0 non-manifold edges and 0 boundary edges
The mesh has 1 connected components
Valences entropy: 1.66243
0 disconnected vertices, 1522 connected vertices
47.7004 percent of irregular vertices
Mesh geometry qualitity:
AngleMin=19.2219
AverageMinAngle=46.4299
Qmin=0.312611
Qav=0.827966
P30=0.307018

Regards

random_shuffle deprecated in c++17

In c++17, which is default standard for gcc 11, the random_shuffle function is deprecated : https://en.cppreference.com/w/cpp/algorithm/random_shuffle (this page also gives alternative implementation with random_device which is c++11 compliant)

Only one occurrence is found in the code :

ACVD/Common/vtkUniformClustering.h

Line 1195 in d3e1f6c

std::random_shuffle( Items, Items + NumberOfRemainingItems );

Note1 gcc still build it but MSVC does not.
Note2 'register' keyword is unused in c++17 and many warnings are thrown

ACVD fails using .obj format as input

I built ACVD on Windows 10 64 bit, VS2015, VTK 7.0.0.

Using .obj files as input does not correctly generate output ply files.

Bunny model does not produce high quality clustering

Here is the bunny model clustered with ACVD:
https://www.dropbox.com/s/x1wm0ltejrq61kw/bunnyClusters.jpg?dl=0

It does not match the bunny clustering quality on the ACVD page:
http://www.creatis.insa-lyon.fr/site/sites/default/files/resize/remote/89b80c070ee02e75a1627fe5e2a30065-224x226.gif

Command line options: bunny.obj 300 0

Are there options that options that I am missing?

vtkDiscreteRemeshing missing

When I create the solution with cmake it doesn't generate the vtkDiscreteRemeshing project, and when I compile I'm missing vtkDiscreteRemeshing.lib

Extra build dependencies needed

I just built this today, and I needed to install a few extra dependencies that weren't listed in order to get it to work.

I'm using Arch Linux (5.7.10-arch1-1) and these extra packages were needed:

• gdal
• unixodbc
• pdal
• pugixml

This comment helped a lot:
PointCloudLibrary/pcl#2616 (comment)

Thanks!

CMake Configure fails to create Visual Studio solution

Trying to create Visual Studio solution/projects, I get the following error msg even after I manually specified the location of VTK_DIR:

strncasecmp not found in ..\..\ACVD\DiscreteRemeshing\xmlParser.cpp(313)

Please patch this into xmlParser.cpp.

ifdef _MSC_VER

define strncasecmp _strnicmp

define strcasecmp _stricmp

endif

see: http://stackoverflow.com/questions/3694723/error-c3861-strcasecmp-identifier-not-found-in-visual-studio-2008

add regression tests

I think I was able to replace vtkSimpleCriticalSection.h with mutex it compiles but wondered if there is any regression test that we could make to check the code still works the same as before.

[bug report] self-intersection tri after remesh

Hi,

I used ACVDQ to remesh a simple cylinder stl but the output file has self-intersection tris (Tri4 and Tri11).

Tri4: 86 97 20
Tri11: 86 20 74

Here is the log and stl files are attached.

models.zip

ACVDQ.exe RemeshInput_CAVD.stl 1000 0.0 -m 0 -v 0 -b 1 -o F:\PyGeoMesh\Results\ -of RemeshMesh_ACVD.ply
load : F:\PyGeoMesh\Results\RemeshInput_CAVD.stl
*****************************************************************************
Mesh with 1890 polygons, 947 points, 2835 edges
Bounding Box: [320.5, 279.8, 339.5]
              [559.5, 519.504, 490.5] 
The mesh is made only of 1890 triangles
0 non-manifold edges and 0 boundary edges
The mesh has 1 connected components
 Valences entropy: 1.76791
0 disconnected vertices, 947 connected vertices
48.5744 percent of irregular vertices
Mesh geometry qualitity:
  AngleMin=27.9745
  AverageMinAngle=44.9631
  Qmin=0.46611
  Qav=0.806352
  P30=0.15873
*****************************************************************************
Force Manifold=0
Setting output verbose : 0
Setting boundary fixing to : 1
OutputDirectory: F:\PyGeoMesh\Results\
Output file name: RemeshMesh_ACVD.ply
Setting Metric Constraint to 0

Unconstrained initialization done
Setting Metric Constraint to 1
*****************************************************************************
Mesh with 1996 polygons, 1000 points, 2994 edges
Bounding Box: [320.433, 279.8, 339.433]
              [559.554, 519.597, 490.536] 
The mesh is made only of 1996 triangles
0 non-manifold edges and 0 boundary edges
The mesh has 1 connected components
 Valences entropy: 1.68123
0 disconnected vertices, 1000 connected vertices
49.9 percent of irregular vertices
Mesh geometry qualitity:
  AngleMin=14.2918
  AverageMinAngle=45.1805
  Qmin=0.23902
  Qav=0.809694
  P30=0.835003
*****************************************************************************
The remesh took :0.425 seconds.

VTK 7.x removes vtkstd

While trying to compile this project against the latest VTK (v7.1.0), I ran into a compile error with vtkstd being used. It appears that this has been deprecated and should probably be changed to the regular std namespace. Doing so in VolumeProcessing/vtkOOCMetaImageReader.cxx made the build pass again.

Cow model remesh leaves disconnected triangles

ACVD cow.obj 200 0

model is here
https://www.dropbox.com/s/mdm55swrd1245do/cow.obj?dl=0

ACVD cow.obj 200 0 -m 1

cause abnormal exit with "Not enough spare clusters! allocate more"

Unable to compile on MacOS 10.15 (Catalina) with VTK 9 and Cmake 3.18

Hello, I've come across a slew of compilation errors using the environment described above. Do you guys plan to support these? Example error below.

Thanks,
Omar

/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/../include/c++/v1/cmath:317:9: error: no member named 'signbit' in the global namespace using ::signbit;

VTK 9 Compilation Error

I am trying to build ACVD using latest VTK 9.0 library.

But when building against vtkSurfaceBase.h , the compiler aborts with errors:

vtkIdType n,*pts;
this->Polys->GetCell(this->Cells->GetCellLocation(face),n,pts);

The error message is:

“GetCellLocation” is not a member of vtkPolyData_detail::CellMap

How may I fix this?

Missing file when compiling

Hello,

I am trying to compile the latest version of ACVD with Visual Studio 2017. However, when I build the solution, I get a lot of the following:

fatal error LNK1181: cannot open input file 'vtkChartsCore.lib'
fatal error LNK1181: cannot open input file '..\..\bin\Release\vtkSurface.lib'
fatal error LNK1181: cannot open input file '..\..\bin\Release\vtkDiscreteRemeshing.lib'

What could be the cause?
Note: I am using VTK 7.1.1

Thanks!

MSVC compiler complains about int to vtkIdType conversion

Hi ACVD developers,

I am building ACVD using VTK 9 using MSVC (16.11.11) from Visual Studio 2019 for x64 architecture and the compilation failed due to the error "int to vtkIdType conversion" on line 119 in vtkVerticesProcessing.h.

It seems like easy fix. I just replace line 118 with
// int NumberOfEdges,*Edges,i;
vtkIdType NumberOfEdges,*Edges,i;

Then, the complication is succesful.

Error in the doc

Hello,
Just a small detail, but on the command line interface, the doc specify :
-d 0/1/2 : enables display (default : 0)
but the real default value is 1.

Should change the default value in DiscreteRemeshing/Examples/ACVD.cxx:60 and DiscreteRemeshing/Examples/ACVDQ.cxx:80, or change the doc in DiscreteRemeshing/Examples/ACVD.cxx:87 and DiscreteRemeshing/Examples/ACVDQ.cxx:103