ANTsPyNet: Incorrect sform/qform in `croppedMni152`

Hi ANTs devs,

I was using preprocessing pipeline proposed by your paper for cortical thickness and brain segmentation. I want to obtain the MNI coordinate from the transforms (inside antspynet.preprocess_brain_image) Then I realize the sample image croppedMni152 has incorrect sform/qform (Image IJK voxel space to RAS, in this case MNI152).

>>> import nibabel
>>> import antspynet
>>> img = nibabel.load(antspynet.get_antsxnet_data("croppedMni152"))
>>> img.header.get_sform()
array([[0., 0., 0., 0.],
       [0., 0., 0., 0.],
       [0., 0., 0., 0.],
       [0., 0., 0., 1.]])
>>> img.header.get_qform()
array([[ 1.,  0.,  0., -0.],
       [ 0.,  1.,  0., -0.],
       [ 0.,  0.,  1.,  0.],
       [ 0.,  0.,  0.,  1.]])
>>> 

I wonder if I have an IJK point in croppedMni152, How can I obtain the MNI-152 template coordinate?