# IMAGE PROCESSING

" Two roads diverged in a wood, and I,
I took the one less traveled by,
And that has made all the difference "-Robert Frost

### I can't sleep Illusion

In this illusion, I used slanted lines in two different directions  to highlight the foreground from the background.

First, the background is formed with slanting lines in one direction.

Steps to form background with slanting lines:
1.       Get the size of the foreground image. For instance, MXN =5X8
2.       Initialize a matrix of size 1XN=1X8.
3.       Generate values for the matrix.
Eg. [10  10 250 250 10 10 250 250]ie [BBWWBBWW] B-Black, W-White
4.       Initialize another matrix of size MXN. In the first row copy the random values.
5.       For the second row perform circshift. By circshift the pixel at the first column will be moved to the last column and the remaining pixels are shifted to left by one.Eg [10 250 250 10 10 250 250 10].
Now the colors will be [BWWBBWWB].
6.       Continue to perform circshift  for all the rows.  Now the slanting lines are formed.
10    10   250   250    10    10   250   250
10   250   250    10    10   250   250    10
250   250    10    10   250   250    10    10
250    10    10   250   250    10    10   250
10    10   250   250    10    10   250   250

MATLAB CODE:

m=size(TImg,1);
n=size(TImg,2);

if(mod(n,2)~=0)
n=n-mod(n,2);
end
if(mod(n,4)~=0)
n=n-mod(n,4);
end
BImg=uint8(zeros([1 n 3]));
flag=0;
%Values for first column (bbwwbbwwbbww)
for i=1:2:n
if(flag==1)
BImg(1,i:i+1,:)=10;
flag=0;
else
BImg(1,i:i+1,:)=240;
flag=1;
end
end
%Initialization
A=uint8(zeros([m size(BImg,2) 3]));
Arot=uint8(zeros([m size(BImg,2) 3]));
SImg=uint8(zeros([m size(BImg,2) 3]));
FImg=uint8(zeros([m size(BImg,2) 3]));
A(1,:,:)=BImg(1,:,:);

%Slanting lines
for i=2:m
BImg=circshift(BImg,[1,-1]);
A(i,:,:)=BImg(1,:,:);
end

Steps to perform on foreground Image:
1.       Rotate the background image from left to right.
2.       Convert the foreground image into binary image and negate it. Then perform image multiplication with the rotated background image.

MATLAB CODE:
%Flip the image from left to right
Arot(:,:,1)=fliplr(A(:,:,1));
Arot(:,:,2)=fliplr(A(:,:,2));
Arot(:,:,3)=fliplr(A(:,:,3));
%Foreground
BImg=uint8(~im2bw(TImg));
BImg=imresize(BImg,[size(Arot,1) size(Arot,2)]);
SImg(:,:,1)=BImg.*Arot(:,:,1);
SImg(:,:,2)=BImg.*Arot(:,:,2);
SImg(:,:,3)=BImg.*Arot(:,:,3);
SImg(SImg==0)=1;

1.       Multiply the background image with the binary foreground image to obtain the mask.
MATLAB CODE:
BImg1=uint8(im2bw(TImg));
BImg1=imresize(BImg1,[size(Arot,1) size(Arot,2)]);
FImg(:,:,1)=BImg1.*A(:,:,1);
FImg(:,:,2)=BImg1.*A(:,:,2);
FImg(:,:,3)=BImg1.*A(:,:,3);
FImg(FImg==0)=1;

Combine Foreground Image and the background Image mask.

MATLAB CODE:
Zig=FImg.*SImg;
figure,imshow(Zig);

GUESS THE IMAGE:

Check this link for more optical Illusion Images: Images from Blog

Like "IMAGE PROCESSING" page