Algorithm 2. Overall Algorithm.

Input: Blurred image B, parameters $\mathsf{\theta }$, ${\mathsf{\alpha }}_s$ and the size of blur kernel;

Build an image pyramid {Bs} and all-zero kernel pyramid {ks} with level index {1, 2, …, n} according to the size of blur kernel;

1. Blind estimation of blur kernel

for = 1 to n do

Compute adaptive weight M(Bs) (Equation (4)).

for i = 1 to m (m iterations) do

Extract salient structure Is (Equation (2)).

Select reliable structure $\nabla S$ for kernel estimation (Equation (6))

Estimate blur kernel according to Algorithm 1.

Restore interim latent image L (Equation (14))

$\mathsf{\theta }\leftarrow \mathsf{\theta }/1.1$, ${\mathsf{\alpha }}_s\leftarrow {\mathsf{\alpha }}_s/1.1$

end for

Up-sample latent image: $L_{l+1}\leftarrow L_l\uparrow$.

Porject $k_l$ onto the constraints (Equation (8)) and up-sample blur kernel: ${\mathrm{k}}_{l+1}\leftarrow k_l\uparrow$.

end for

2. Image restoration using MR-based Wiener Filter.

-Recover I using k from B in full-scale resolution(Equation (15))

Output: Blur kernel k and latent sharp image I.