Not able to read .ply voxel output.

[AzamatB]

 PoissonRecon --in bunny.points.ply --voxel bunny.voxel.ply --depth 10

My understanding is that the line above outputs bunny.voxel.ply file that contains 3D array (voxel grid), where each element is a value of the implicit function.

I tried reading bunny.voxel.ply with pcl library, but getting read error... Even though with the same code I am able to read the mesh file bunny.ply produced by

PoissonRecon --in bunny.points.ply --out bunny.ply --depth 10

So my question is how can I read that 3D array from bunny.voxel.ply?

[mkazhdan]

That's right. Though the first 4 bytes store the binary encoding of the resolution of the grid (2^d).

-- misha

On July 9, 2017 4:32:23 PM EDT, AzamatB notifications@github.com wrote:

PoissonRecon --in bunny.points.ply --voxel bunny.voxel.ply --depth 10

My understanding that the line above outputs 3D array (voxel grid), where each element is a value of the implicit function.

-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/mkazhdan/PoissonRecon/issues/28

[AzamatB]

So is there a way to read in this 3D array from bunny.voxel.ply?

The following code that uses pcl library gives error

#include <iostream>
#include <pcl/point_types.h>
#include <pcl/io/pcd_io.h>
#include <pcl/io/ply_io.h>

int main (int argc, char** argv) {
    pcl::PointCloud<pcl::PointXYZ>::Ptr cloud (new pcl::PointCloud<pcl::PointXYZ>);

    if (pcl::io::loadPLYFile<pcl::PointXYZ> ("bunny.voxel.ply", *cloud) == -1) //* load the file
    {
        PCL_ERROR ("Couldn't read file bunny.voxel.ply \n");
        return (-1);
    }
    std::cout << "Loaded "
              << cloud->width * cloud->height
              << " data points from bunny.voxel.ply"
              << std::endl;
    return (0);
}

[mkazhdan]

I'm not sure I follow the question here. Are you trying to read the voxel grid as a ply file? That won't work. It's written in raw binary, so that if r=2^d is the resolution, the file will have size (rrr+1)4 bytes. With the first 4 bytes used to store the binary representation of the integer r and the remaining rrr4 bytes storing the binary representation of rr*r single-precision floating point values​ of the implicit function.

On July 9, 2017 10:17:41 PM EDT, AzamatB notifications@github.com wrote:

#include <iostream>
#include <pcl/io/pcd_io.h>
#include <pcl/point_types.h>
#include <pcl/features/normal_3d.h>
#include <pcl/io/ply_io.h>

int main (int argc, char** argv) {
pcl::PointCloud<pcl::PointXYZ>::Ptr cloud (new
pcl::PointCloud<pcl::PointXYZ>);

if (pcl::io::loadPLYFile<pcl::PointXYZ> ("bunny.voxel.ply", *cloud) ==
-1) //* load the file
   {
       PCL_ERROR ("Couldn't read file test_pcd.pcd \n");
       return (-1);
   }
   std::cout << "Loaded "
             << cloud->width * cloud->height
       << " data points from test_pcd.pcd with the following fields: "
             << std::endl;
   return (0);
}

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/mkazhdan/PoissonRecon/issues/28#issuecomment-313985255

[AzamatB]

Ok, I'll try to explain. My goal is to get the 3D binary image of the object, i.e. voxelized representation instead of mesh representation. To do this I was planning to read 3D tensor of single-precision floating point values​ of the implicit function that as I understand is stored in the bunny.voxel.ply file produced by PoissonRecon --in bunny.points.ply --voxel bunny.voxel.ply --depth 10 and then work with it to get the binary 3D tensor from it. But I cannot retrieve (i.e. read into the memory) the 3D tensor of the implicit function values from bunny.voxel.ply.

[mkazhdan]

Assume you just want the 3D array of values, you can read it as:

FILE fp = fopen( "bunny.voxel" , "rb" ); int res; fread( &res , 1 , sizeof(int) , fp ); float voxel = new float[resresres]; fread( voxel , resresres , sizeof(float) , fp ); fclose( fp );

On July 9, 2017 11:50:00 PM EDT, AzamatB notifications@github.com wrote:

Ok, I'll try to explain. My goal is to get the 3D binary image of the object, i.e. voxelized representation instead of meshes. To do this I was planning to read 3D tensor of single-precision floating point values​ of the implicit function that as I understand is stored in the bunny.voxel.ply file produced by PoissonRecon --in bunny.points.ply --voxel bunny.voxel.ply --depth 10 and then work with it to get the binary 3D tensor from it. But I cannot retrieve (i.e. read into the memory) the 3D tensor of the implicit function values from bunny.voxel.ply.

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/mkazhdan/PoissonRecon/issues/28#issuecomment-313995352

[AzamatB]

Great! Thank you! One clarification: in this code voxel seems to be 1D array. So say if I want to access element, which in voxel grid (3D tensor), has indicies [x][y][z], how would I find this element's index in voxel?

[mkazhdan]

voxel(x,y,z)=voxel[resresz+res*y+x]

On July 10, 2017 3:43:28 PM EDT, AzamatB notifications@github.com wrote:

Great! Thank you! One clarification: in this code voxel seems to be 1D array. So say if I want to access element, which in voxel grid (3D tensor), has indicies [x][y][z]. how can I access it from voxel

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/mkazhdan/PoissonRecon/issues/28#issuecomment-314213493

[mikeroberts3000]

On a related note, does the volume always extend spatially from (0,0,0) to (1,1,1)? This question is related to https://github.com/mkazhdan/PoissonRecon/issues/21.

[mkazhdan]

Yep.

From: Mike Roberts [mailto:notifications@github.com] Sent: Tuesday, August 22, 2017 1:15 AM To: mkazhdan/PoissonRecon PoissonRecon@noreply.github.com Cc: mkazhdan misha@cs.jhu.edu; Comment comment@noreply.github.com Subject: Re: [mkazhdan/PoissonRecon] Not able to read .ply voxel output. (#28)

On a related note, does the volume always extend spatially from (0,0,0) to (1,1,1)? This question is related to #21 https://github.com/mkazhdan/PoissonRecon/issues/21 .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/mkazhdan/PoissonRecon/issues/28#issuecomment-323921236 , or mute the thread https://github.com/notifications/unsubscribe-auth/AKXc5RIWiq6DWU4vZufrdNOmiGjYwl-oks5samPEgaJpZM4OSLTS . https://github.com/notifications/beacon/AKXc5RSPm6zS6k_wf2G5zsh2HLOkoEqGks5samPEgaJpZM4OSLTS.gif

[wangmiaowei]

I write a python code to load this voxel: import numpy as np

def load_voxel_grid(file_name):
    with open(file_name, 'rb') as f:

        res = np.frombuffer(f.read(4), dtype=np.int32)[0]
        print(f"Resolution: {res}")

        resolution = res ** 3

        voxel_values = np.frombuffer(f.read(resolution * 4), dtype=np.float32)

    return res, voxel_values

if __name__ == '__main__':
    file_name = 'real_mesh.voxel' 
    resolution, voxel_values = load_voxel_grid(file_name)

    print("Voxel Values:", voxel_values[:10])

but the resolution is very large: Resolution: 822752071 any issue?

[mkazhdan]

I'm afraid I don't understand the question or it's relation to the Poisson Reconstruction code.

On October 13, 2024 6:43:20 AM EDT, Miaowei Wang @.***> wrote:

I write a python code to load this voxel: import numpy as np

def load_voxel_grid(file_name):
   with open(file_name, 'rb') as f:

       res = np.frombuffer(f.read(4), dtype=np.int32)[0]
       print(f"Resolution: {res}")

       resolution = res ** 3

       voxel_values = np.frombuffer(f.read(resolution * 4), dtype=np.float32)

   return res, voxel_values

if __name__ == '__main__':
   file_name = 'real_mesh.voxel' 
   resolution, voxel_values = load_voxel_grid(file_name)

   print("Voxel Values:", voxel_values[:10])

but the resolution is very large: Resolution: 822752071 any issue?

-- Reply to this email directly or view it on GitHub: https://github.com/mkazhdan/PoissonRecon/issues/28#issuecomment-2408925827 You are receiving this because you commented.

Message ID: @.***>

[mkazhdan]

Apologies. The grid format was changed a while back and the description in the readme was out of sync. I have modified it so that the description of the contents now matches what is output.

[wangmiaowei]

yes, there is an updation.