Export volumetric datasets to Blender

It is straightforward to generate a binary file from volumetric data that can be used by Blender. The script is available in examples/Blender/Exporter. The following function will do:

def generate_blender_voxeldata(fname, data):
    """ Write a binary file readable by Blender

    Parameters
    ----------
    fname : String
        Filename to store the output, ending with .bin
    data : ndarray
        If the array is 3 dimensional, assume a structural image
        If the array is 4 dimensional, assume a functional image

    Note
    ----
    For instructions on how to render the structural or functional images,
    see http://brainblend.sourceforge.net/index.html
    """

    import struct
    import numpy as np

    if len(data.shape) == 4:
        frames = data.shape[3]
        time = True
    elif len(data.shape) == 3:
        frames = 1
        time = False
    else:
        raise Exception('Data array has no valid dimension!')

    xdim = data.shape[0]
    ydim = data.shape[1]
    zdim = data.shape[2]

    # write data to binary file
    binfile = open(fname, 'wb')
    x = struct.pack('i', xdim)
    y = struct.pack('i', ydim)
    z = struct.pack('i', zdim)
    f = struct.pack('i', frames)

    binfile.write(x)
    binfile.write(y)
    binfile.write(z)
    binfile.write(f)

    # convert to float32
    data = np.array(data, dtype = np.float32)

    if time:
        for i in range(frames):
            print 'write frame... ', i
            data[:,:,:,i].tofile(binfile)
    else:
        data.tofile(binfile)

    binfile.close()

You can use the generated file with structural or functional information as volumetric texture. Please refer to the BrainBlend documentation. You might to want to do some data preprocessing as well.

The Blender documentation about Volume Rendering has more information. A template Blender file importing_voxeldata_struct.blend is provided in the example folder.