simple itk is not behaving as expected while flip using affine transform

Question

I want to flip a volume using simpleitk.

vol = np.arange(1, 25).reshape((4, 6))
print(vol)
vol = sitk.GetImageFromArray(vol)

Result:

[[ 1  2  3  4  5  6]
 [ 7  8  9 10 11 12]
 [13 14 15 16 17 18]
 [19 20 21 22 23 24]]

according to the this torurial there are two ways of doing it.

using slicing
using an affine transform.

However, the result of the second approach is not correct because of the transformation center.

flip = sitk.AffineTransform(2) 
flip.SetMatrix([1.0,0,0,-1.0]) volume_shape= np.array(vol.GetSize()) 
center = (int(volume_shape[0] / 2), int(volume_shape[1] / 2))
#if I change the center to (3,1.5) it works 
flip.SetCenter(center)
interp_img = sitk.sitkLinear 
vol_resampled = sitk.Resample(vol, flip, interp_img,0.0)

print(sitk.GetArrayFromImage(vol_resampled))
print(sitk.GetArrayFromImage(vol[:, ::-1]))

Result:

[[ 0  0  0  0  0  0]
 [19 20 21 22 23 24]
 [13 14 15 16 17 18]
 [ 7  8  9 10 11 12]]

[[19 20 21 22 23 24]
 [13 14 15 16 17 18]
 [ 7  8  9 10 11 12]
 [ 1  2  3  4  5  6]]

I would appreciate it if someone could explain the reason.

I think it has to do with the rounding that you're doing when creating the 'center' in integers. Try not converting to ints, but just leaving them as floats. The rounding is shifting the pixel centers down one row in Y. — Dave Chen, Sep 14 '20 at 16:08
Thank you for your reply. I tried it but it didn't work and anyway 4/2 is 2 and 6/2 is 3 which cannot be rounded by int. — ehsan olyaee, Sep 17 '20 at 13:40

score 1 · Answer 1 · edited Jan 19 '22 at 00:12

The simple solution:

center = ((volume_shape[0] - 1) / 2, (volume_shape[1] - 1) / 2)

Here is the solution when spacing != 1.0:

spacing = vol.GetSpacing()
center = ((volume_shape[0] - 1) * spacing[0] / 2, (volume_shape[1] - 1) * spacing[1] / 2)

Reason: sitk.Resample only cares about the real pixel position. volume_shape = [6, 4] The pixels are located in [0, 1, 2, 3, 4, 5] for the first axis and [0, 1, 2, 3] for the second axis. You need to set the center as (2.5, 1.5) rather than (3, 2).

simple itk is not behaving as expected while flip using affine transform

1 Answers1