MATLAB : Aligning RGB Channels

Align the 3 images from earlier project to get a clearer color image. This project needs to be done using SUM of SQUARED DIFFERENCES APPROACH.

please help me with the code....

%Read the image
img = imread('course1image.jpg');
B=img(1:340,1:400);
G=img(341:682,1:400);
R=img(683:1023,1:400);
b=im2double(B);
g=im2double(G);
r=im2double(R);
shiftr=circshift(r,[-10,10]);
shiftb=circshift(b,[-5,5]);
 ColorImg_aligned=cat(3,shiftr,g,shiftb);

[DGM]

See also:

https://www.mathworks.com/matlabcentral/answers/542507-color-imaging-rgb-channels

https://www.mathworks.com/matlabcentral/answers/629558-color-imaging-rgb-channels

https://www.mathworks.com/matlabcentral/answers/558349-color-imaging-rgb-channel

Answer 1

[Thyagharajan K K]

I have solved the problem which passes first 5 tests. To solve the last part, I need the following clarifications.

What is the meaning of this instruction already set in the problem? should we move 51x51 green block on red channel to a maximum of 10 positions left and 10 position right to find matching

shiftr=circshift(r,[-10,10]);

if 51x51 g pixel block finds a match in r should we extend the size of the channels. The problem is not clear. It is difficult to understand the aim and methodology expected to follow. In the video r and b channels are moved completely on g channel but in the problem 51x51 pixel block is introduced for matching. The problem can be defined with more clarity.

%Read the image
img = imread('course1image.jpg');
B=img(1:341,1:400);
G=img(342:682,1:400);
R=img(683:1023,1:400);
b=double(B);
g=double(G);
r=double(R);

%take a block of size 50x50 pixels
g1 = g(146:196, 175:225);
r1 = r(146:196, 175:225);
b1 = b(146:196, 175:225);
ref_img_region = g1;

shiftr=circshift(r,[-10,10]);
shiftb=circshift(b,[-5,5]);

ColorImg_aligned=cat(3,uint8(shiftr),uint8(g),uint8(shiftb));

[Deema Alyones]

I have a question on one of the parts in the code you wrote. What does the numbers inside the parentheses represent and how can I find these values from the image?

B=img(1:341,1:400);
G=img(342:682,1:400);
R=img(683:1023,1:400);

and

g1 = g(146:196, 175:225);
r1 = r(146:196, 175:225);
b1 = b(146:196, 175:225);

Answer 2

[Medical Imaging]

I hope the below soultion helps:

%Read the image
img = imread('course1image.jpg');
B=img(1:341,1:400);
G=img(342:682,1:400);
R=img(683:1023,1:400);
b=double(B);
g=double(G);
r=double(R);

b1=double(b(146:196,175:225));
g1=double(g(146:196,175:225));
r1=double(r(146:196,175:225));

ref_img_region = double(g1);
ref_img_region = double(ref_img_region);

error = inf;
for i = -10:10
    for j = -10:10
        shiftr1=circshift(r1,[i,j]);
        temp1 = sum(sum((double(g1) - double(shiftr1)) .^ 2));
        if temp1 < error
            error = temp1;
            shiftr_row = i;
            shiftr_col = j;
        end
    end
end

error = inf;
for i = -10:10
    for j = -10:10
        shiftb2=circshift(b1,[i,j]);
        temp2 = sum(sum((double(g1) - double(shiftb2)) .^ 2));
        if temp2 < error
            error = temp2;
            shiftb_row = i;
            shiftb_col = j;
        end
    end
end

shiftr=circshift(r,[shiftr_row,shiftr_col]);
shiftb=circshift(b,[shiftb_row,shiftb_col]);
outpict=cat(3,uint8(shiftr),uint8(g),uint8(shiftb));

imshow(ColorImg_aligned)

Answer 3

[Chit La Pyae Myo Hein]

%Read the image
img = imread('course1image.jpg');
[r,c] = size(img);

B = img(1:r/3,:);
G = img((r/3)+1:(2*r/3),:);
R = img((2*r/3)+1:r,:);

ref_img_region = G;
[rg,cg] = size(ref_img_region);
ref_img_region = ref_img_region(ceil((rg-50)/2) :ceil((rg-50)/2) + 50,ceil((cg-50)/2) :ceil((cg-50)/2) + 50);
%disp(size(ref_img_region));
ref_img_region = double(ref_img_region);

% Naive way
% ColorImg_aligned = cat(3,R,G,B);
% imshow(ColorImg_aligned);

% SSD way
nR = align(G,R);
nB = align(G,B);
ColorImg_aligned = cat(3,nR,G,nB);
imshow(ColorImg_aligned);

function aligned = align(green,red)
    [red_row,red_col] = size(red);
    [green_row,green_col] = size(green);

    % checking SSD for cropped part of the images for faster calculation 
    cropped_red = red(ceil((red_row-50)/2) : ceil((red_row-50)/2) + 50,ceil((red_col-50)/2) :ceil((red_col-50)/2) + 50);
    cropped_green = green(ceil((green_row-50)/2) : ceil((green_row-50)/2) + 50,ceil((green_col-50)/2) :ceil((green_col-50)/2) + 50);

    MiN = 9999999999;
    r_index = 0;
    r_dim = 1;
    
    for i = -10:10
        for j = -10:10
            ssd = SSD(cropped_green,circshift(cropped_red,[i,j]));
            if ssd < MiN
                MiN = ssd;
                r_index = i;
                r_dim = j;
            end
        end
    end
    aligned = circshift(red,[r_index,r_dim]);
end       

function ssd = SSD(a1,a2)
    x = double(a1)-double(a2);
    ssd = sum(x(:).^2);
end