with the given julia code:

abstract type AbstractStencil end

struct AverageStencil $<$: AbstractStencil end

apply$\_$to$\_3$x$3($s::AverageStencil, A$33)=$ sum$($A$33)/$ length$($A$33)$

function stencil$\_$demo$($s::AbstractStencil$)$

count $=0$

plot$\_$next$($A$)=$ subplot$(1,4,$count$+=1),$ imshow$($A$[$:$,$:$,[1,1,1]])$

for noise $=[0,0\mathrm{.}3]$

A $=$ test$\_$image$(1,50,$ noise$)$

plot$\_$next$($A$)$

plot$\_$next$($apply$\_$stencil$($s$,$A$))$

end

end

function apply$\_$stencil$($s::AbstractStencil, A$)$

# Your code

B $=0*$A

for i $=2$:size$($A$,1)-1,$ j $=2$:size$($A$,2)-1$

B$[$i$,$j$]=$ apply$\_$to$\_3$x$3($s$,$A$)$

end

return B

end

In Julia code: define a new subtype EdgeStencil which applies the same operation as maxabsgradfilter in the lecture notes.

here is the function from the lecture notes

function image$\_$maxabsgradfilter$($A$)$

B $=0*$A

for i $=2$:size$($A$,1)-1,$ j $=2$:size$($A$,2)-1$

a $=$ A$[$i$-1$:i$+1,$ j$-1$:j$+1]$

B$[$i$,$j$]=$ max$($abs$($a$[3,2]-$ a$[1,2]),$ abs$($a$[2,3]-$ a$[2,1]))$

end

return B

end

then test with

stencil$\_$demo$($EdgeStencil$())$