Algorithm 3. Removing distortion from image.

input: $x,U,V,\tau$

$extractsubregionofimagexusingslidingwindowsbasedontheentropyofeach$

$pixelas{f}_j,j=1,\dots ,d$

$foreach{f}_j:$

$if{f}_{j}satisfyentropycarcumestion$

calculate $\mathrm{\Delta }{\tau }_j$ using Algorithm 2

else

set $\mathrm{\Delta }{\tau }_j$ as outlier

end if

end for

using variance and average of all $\mathrm{\Delta }{\tau }_j$ to detect outlier ones

for each $f_j$:

if $\mathrm{\Delta }{\tau }_j$ is outlier

calculate $\mathrm{\Delta }{\tau }_j$ using $\mathrm{\Delta }\tau$ matrix of neighboring windows

end if

end for

${\tau }_j={\tau }_j+\mathrm{\Delta }{\tau }_j$

estimate MOG using ${\tau }_j$ and assign mean of MOG models to each windows

for each pixel index j in x

L = set of all jth pixel exist in all transformed windows

jth pixel of $x\circ \tau$ = median(L)

end for

return $x\circ \tau$