Hi all, I'm trying to port `interpolation.affine_transform` [1] to another langage and I have a question about your algorithm. I successfully ported order 1 (linear) interpolation from another lib, but I failed on order 3 (cubic). I tried reading the interpolation code (zoom shift) [2] but it's no simple task. So, I took vtk implementation [3] and ported it, but I can't get your results. I was surprised because it's a mathematical formula, it should give the same result wathever the algorithm. So, my question is: *is there something special with SciPy's cubic interpolation?* I don't get the feeling that the code fits the description in wikipedia [4]. Also, I saw the `spline_filter` function, I ported it and but I still can't get the same results, disabled or not. Nil Goyette [1] https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.ndimage.in... [2] https://github.com/scipy/scipy/blob/master/scipy/ndimage/src/ni_interpolatio... [3] https://gitlab.kitware.com/vtk/vtk/blob/master/Imaging/Core/vtkImageInterpol... [4] https://en.wikipedia.org/wiki/Tricubic_interpolation