The following diagram shows how the Wiener filter can be computed. As can be seen, the Wiener filter is the filter that has the objective function of minimizing the MSE between the restored and the original image:
As you can see from the preceding diagram, the Wiener filter reduces to an inverse filter when no noise is added to the input image (N=0). Here, K is a constant (that is, it does not depend on u, v), to be chosen according to prior knowledge about the noise.
The unsupervised_wiener() function from the scikit-image.restoration module was used to denoise the degraded image using deconvolution with a Wiener-Hunt approach, ...
