ExtractLabelPixelCurves.py

 import numpy as np
 
 def run(self,mvNode,labelNode, filename):
    """
    Extract pixel curves from a multivolume node
    prerequisite: labelNode and mvNode need to have the same dimensions, this is not checked.
    As long as the label node is created from the mvNode, that's no problem at all.
    """

    # get a numpy reference to the moving volume buffer
    mvimg=mvNode.GetImageData()
    npmvimg=vtk.util.numpy_support.vtk_to_numpy(mvimg.GetPointData().GetScalars())
    nFrames=mvNode.GetNumberOfFrames()

    # get a numpy reference to the label image buffer
    labelimg=labelNode.GetImageData()
    nplabelimg=vtk.util.numpy_support.vtk_to_numpy(labelimg.GetPointData().GetScalars())

    # get the times as a list:
    time=np.float(mvNode.GetAttribute('MultiVolume.FrameLabels',','))
    if len(time) != nFrames:
        return
        
    # we follow the multivolume explorer logic, but use  numpy, instead of loops
    # this dictionary holds all indices labeld with a given label value
    labelPixelCurves={}
    # get the labels that are defined in the map:
    labels=np.unique(nplabelimg)
    # 0 means no label, we remove this from the labels:
    labels=labels[labels!=0]

    for labelValue in labels:
        # get the indices of all pixels with the given label value
        labelindex=np.where(nplabelimg == labelValue)[0] # where returns a tuple, hence the [0]
        # and use these indices to get the pixel curves from the mvNode image
        pixelcurves=npmvimg[labelindex,:]
        # store the pixel curves for each label in the dictionary:
        labelPixelCurves[labelValue]=pixelcurves
        # you could save the pixelcurves simply as a textfile with np.savetxt:
        #try:
        #  np.savetxt(filename, pixelcurves
        #except IOError:
        #  print "Failed to write %s" % filename