vtkImageData: No scalar field has been specified-assuming 1 component

Support
1

I am using vmtk to obtain cpr image. But vtkImageData reported a bug.

My code is:


curvedMPRImageFilter = vtkvmtk.vtkvmtkCurvedMPRImageFilter()
curvedMPRImageFilter.SetInputData(vtkImg)
curvedMPRImageFilter.SetCenterline(centerline)
curvedMPRImageFilter.SetParallelTransportNormalsArrayName('ParallelTransportNormals')
curvedMPRImageFilter.SetFrenetTangentArrayName('FSTangents')
curvedMPRImageFilter.SetInplaneOutputSpacing(1, 1)
curvedMPRImageFilter.SetInplaneOutputSize(300, 300)
curvedMPRImageFilter.SetReslicingBackgroundLevel(0)
curvedMPRImageFilter.Update()

A bug is reported for curvedMPRImageFilter.Update():

vtkImageData: No Scalar Field has been specified - assuming 1 component!

The vtkImg is converted from numpy data:

def numpyToVTK(data, multi_component=False, type='float'):

    flat_data_array = data.transpose(2,1,0).flatten()
    vtk_data = numpy_to_vtk(num_array=flat_data_array, deep=True, array_type=vtk.VTK_FLOAT)
    shape = data.shape

    img = vtk.vtkImageData()
    img.GetPointData().SetScalars(vtk_data)
    img.SetDimensions(shape[0], shape[1], shape[2])
    return img

I can not find a solution for this bug. Any suggestion is appreciated~~

2

Hello,

What is displayed when you run the following?

print(vtkImg.GetPointData())
3

The displayed information is:


Connected to pydev debugger (build 193.6911.25)
vtkPointData (0000026EFE343780)
  Debug: Off
  Modified Time: 2290
  Reference Count: 2
  Registered Events: 
    Registered Observers:
      vtkObserver (0000026E80B403E0)
        Event: 33
        EventName: ModifiedEvent
        Command: 0000026E80BC2CB0
        Priority: 0
        Tag: 1
  Number Of Arrays: 1
  Array 0 name = nullptr
  Number Of Components: 1
  Number Of Tuples: 52953088
  Copy Tuple Flags: ( 1 1 1 1 1 0 1 1 )
  Interpolate Flags: ( 1 1 1 1 1 0 0 1 )
  Pass Through Flags: ( 1 1 1 1 1 1 1 1 )
  Scalars: 
    Debug: Off
    Modified Time: 2278
    Reference Count: 1
    Registered Events: (none)
    Name: (none)
    Data type: float
    Size: 52953088
    MaxId: 52953087
    NumberOfComponents: 1
    Information: 0000000000000000
    Name: (none)
    Number Of Components: 1
    Number Of Tuples: 52953088
    Size: 52953088
    MaxId: 52953087
    LookupTable: (none)
  Vectors: (none)
  Normals: (none)
  TCoords: (none)
  Tensors: (none)
  GlobalIds: (none)
  PedigreeIds: (none)
  EdgeFlag: (none)
4

Hello,

If you use the SetActiveScalars function in the numpyToVTK to set the active scalars, will it still solve the problem?

def numpyToVTK(data, multi_component=False, type='float'):

    flat_data_array = data.transpose(2,1,0).flatten()
    vtk_data = numpy_to_vtk(num_array=flat_data_array, deep=True, array_type=vtk.VTK_FLOAT)
    vtk_data.SetName('foo')
    shape = data.shape

    img = vtk.vtkImageData()
    img.GetPointData().SetScalars(vtk_data)
    img.GetPointData().SetActiveScalars('foo')
    img.SetDimensions(shape[0], shape[1], shape[2])
    return img
5

Sorry, but the bug is still reported. The complete code to reproduce this bug is:

import numpy as np
from vmtk import vtkvmtk
import vtk
from vtk.util.numpy_support import vtk_to_numpy, numpy_to_vtk

def numpyToVTKpoint(line):

    centerline = vtk.vtkPolyData()
    points = vtk.vtkPoints()
    lineCell = vtk.vtkCellArray()

    centerline.SetPoints(points)
    centerline.SetLines(lineCell)

    # centerline.GetPoints().SetNumberOfPoints(2)
    # centerline.GetLines().Reset()
    # ptIds = [0, 1]
    # centerline.GetLines().InsertNextCell(2, ptIds)

    ptIds = np.arange(line.shape[0]).tolist()
    centerline.GetPoints().SetNumberOfPoints(line.shape[0])
    centerline.GetLines().Reset()
    centerline.GetLines().InsertNextCell(len(ptIds), ptIds)
    for i in range(line.shape[0]):
        centerline.GetPoints().SetPoint(i, line[i, :])

    fsTangents = vtk.vtkDoubleArray()
    fsTangents.SetName('FSTangents')
    fsTangents.SetNumberOfTuples(1)
    fsTangents.InsertNextValue(1.0)
    fsTangents.InsertNextValue(1.0)
    fsTangents.InsertNextValue(1.0)
    centerline.GetPointData().AddArray(fsTangents)
    centerline.Modified()

    parallelTransportNormals = vtk.vtkDoubleArray()
    parallelTransportNormals.SetName("ParallelTransportNormals")
    parallelTransportNormals.SetNumberOfTuples(1)
    parallelTransportNormals.InsertNextValue(1.0)
    parallelTransportNormals.InsertNextValue(1.0)
    parallelTransportNormals.InsertNextValue(1.0)
    centerline.GetPointData().AddArray(parallelTransportNormals)
    centerline.Modified()

    return centerline

def numpyToVTK(data):

    flat_data_array = data.transpose(2,1,0).flatten()
    vtk_data = numpy_to_vtk(num_array=flat_data_array, deep=True, array_type=vtk.VTK_FLOAT)
    vtk_data.SetName('foo')
    shape = data.shape

    img = vtk.vtkImageData()
    img.GetPointData().SetScalars(vtk_data)
    img.GetPointData().SetActiveScalars('foo')
    img.SetDimensions(shape[0], shape[1], shape[2])
    return img

def vtkToNumpy(data):
    temp = vtk_to_numpy(data.GetPointData().GetScalars())
    dims = data.GetDimensions()

    numpy_data = temp.reshape(dims[2], dims[1], dims[0])
    numpy_data = numpy_data.transpose(2,1,0)

    return numpy_data

image = np.zeros([256, 256, 256])
image[64:192, 64:192, :] = 1

x = np.ones(shape=[128])*128
y = np.ones(shape=[128])*128
z = np.arange(0, 128)+64
point = np.array([x, y, z]).T

centerline = numpyToVTKpoint(point)
vtkImg = numpyToVTK(image)

curvedMPRImageFilter = vtkvmtk.vtkvmtkCurvedMPRImageFilter()
curvedMPRImageFilter.SetInputData(vtkImg)
curvedMPRImageFilter.SetCenterline(centerline)
curvedMPRImageFilter.SetParallelTransportNormalsArrayName('ParallelTransportNormals')
curvedMPRImageFilter.SetFrenetTangentArrayName('FSTangents')
curvedMPRImageFilter.SetInplaneOutputSpacing(1, 1)
curvedMPRImageFilter.SetInplaneOutputSize(300, 300)
curvedMPRImageFilter.SetReslicingBackgroundLevel(0)
curvedMPRImageFilter.Update()
cprImg = curvedMPRImageFilter.GetOutput()
cprImg = vtkToNumpy(cprImg)
6

Hello,

This appears to be a bug in vmtk that was fixed on November 13, 2018.

However, it is not reflected in the latest release version 1.4.0 (March 20, 2018).

Are you using version 1.4.0?

7

Yes, my vmtk version is 1.4.0. Because I can only download 1.4.0 vmtk. There is on other version unless I compile the vmtk from the source code.

8

Hello, may I ask if your problem has been resolved?

“ERROR: In …/Common/DataModel/vtkImageData.cxx, line 1324
vtkImageData (0x2110c60): No Scalar Field has been specified - assuming 1 component!”

I also encountered this error when performing surface reconstruction on coronary arteries, which has been bothering me for a long time. Looking forward to your reply. Thank you very much