spyrit.core.torch.neumann_boundary

spyrit.core.torch.neumann_boundary(img_shape)

Creates a finite difference matrix of shape \((h*w,h*w)\) for a 2D image of shape \((h,w)\). The boundary condition used is Neumann.

Args:

img_shape (tuple): The size of the image \((h,w)\).

Returns:

torch.tensor: The finite difference matrix.

Note

This function returns the same matrix as finite_diff_mat() with the Neumann boundary condition. Internal implementation is different and allows to process rectangular images.