color image processing (3*240960 double)

Hi, I have doubt, I am attaching code that my senior worked on to create illusion in gray levels. Now I am trying to make that work in color iimages. So, the one thing I noticed while debugging his code is 3*240960 double. He used only 3*240960 double form to find median from his axis and did the rectification required to create illusion only in this 3*240960 double. After doing all that he finally inversed the 3*240960 with the matrix to create a displayable RGB image. But why is that done like that? I was confused with his codes... for a long time. Because I did in my way to convert the image in desired color space and i was working on the RGB image. Which never gave my desired resullt. So I want to know why 3*240960 double is the way worked for him?

Please kindly help.

The below code is my code, that I did with the help of matlab answers section. The attached code is his code.

RGB = imread('originalimg.png');  
MyImrgb = reshape(im2double(RGB), [],3);
Imrgb = MyImrgb.^2.2; %gamma correction
[x, y, z] = size(Imrgb);
%% transform my picture in a DKL color space
% Define a DKL2RGB mat based on the classic calibration technic
% this P matrix defines DKL in RGB color space
                     %ld      % rg      %yv              
 ldrgyv2rgbMat =  [1.0000  1.0000              0.236748566577269   %R
                   1.0000 -0.299338211934457  -0.235643322285071    %G
                   1.0000  0.0137437185685517  1];  % B
MyImrgbCol = reshape(Imrgb * ldrgyv2rgbMat, size(RGB));
 MyImldrgyvCol = reshape(reshape(MyImrgbCol, [],3) *inv(ldrgyv2rgbMat),size(MyImrgbCol));
 figure(1),imshow(MyImldrgyvCol)
 figure (2);
plot3(MyImldrgyvCol(2,1:25:end),MyImldrgyvCol(3,1:25:end),MyImldrgyvCol(1,1:25:end),'r+'); hold on; % displaying the image in the 3 axis where the z axis is kept constant value displaced only on x and y axis 
plot3([0 0], [-1 1], [0 0], 'k-','LineWidth', 2);
plot3([-1 1], [0 0], [0 0], 'k-','LineWidth', 2);
plot3([0 0], [0 0], [-1 1], 'k-','LineWidth', 2);

xlabel('LM'); ylabel('S'); zlabel('LD');
axis([-1 1 -1 1 -1 1])
vecx = -1 : 0.50 : 1; vecy = -1 : 0.50 : 1; vecz = -1 : 0.50 : 1;
grid on;
set(gca, ...
    'XTick'       , vecx        , ...
    'YTick'       , vecy        , ...
    'ZTick'       , vecz       );
LMaxes1 = (MyImldrgyvCol(2,:));
MedLM = median(LMaxes1);
figure(3),imhist(LMaxes1)
figure(4),imhist(MyImldrgyvCol(2,:,:))

Thanks

[Image Analyst]

I'm not sure of the question. So he reshaped to get a list of RGB values, then multiplied by a 3x3 matrix to get values in the new color space, and then reshaped into an RGB image again. So? And it looks like you did the same thing. Your code worked when I ran it, but you didn't say why what it produced was not the "desired result".

You say you want to know "why 3*240960 double is the way worked for him"? Well an Nx3 times a 3x3 will give another Nx3 matrix which can be converted into an RGB image. Can't you ask him questions about his code?

And it just looks like an image where the dark values are darker than the original. What is the illusion supposed to be?

[Malini Bakthavatchalam]

Nope, he is not in the lab anymore, all I have is only the previous works...

My question is When i am trying to look at the median value for particular axis say x axis i have to use this function like . But when we run his code LMaxes1 = median(MyImldrgyvCol(2,:)); my MyImldrgyvCol has the value 502*480*3. But when I run his code ... he works with MyImldrgyvCol which is a 3*240960. so after doing all his works in (3*240960) form, only at the end he is transfering to RGB image. but I am not able to do the same way. Also what is the meaning of

RGB = imread('originalimg.png');

MyImrgb = reshape(im2double(RGB), [],3);

Imrgb = MyImrgb.^2.2; %gamma correction

[x, y, z] = size(Imrgb);

%% transform my picture in a DKL color space

% Define a DKL2RGB mat based on the classic calibration technic

% this P matrix defines DKL in RGB color space

%ld % rg %yv

ldrgyv2rgbMat = [1.0000 1.0000 0.236748566577269 %R

1.0000 -0.299338211934457 -0.235643322285071 %G

1.0000 0.0137437185685517 1]; % B

%MyImrgbCol = reshape(Imrgb * ldrgyv2rgbMat, size(RGB));

%MyImldrgyvCol = reshape(reshape(MyImrgbCol, [],3) *inv(ldrgyv2rgbMat),size(MyImrgbCol));

MyImrgbCol = reshape(MyImrgb, [x*y, z]);

% define each RGB pixe lin DKL (inverse matrix Q or P with \)

MyImldrgyvCol = ldrgyv2rgbMat\(MyImrgbCol'*2 - 1);

figure

plot3(MyImldrgyvCol(2,1:25:end),MyImldrgyvCol(3,1:25:end),MyImldrgyvCol(1,1:25:end),'r+'); hold on; % displaying the image in the 3 axis where the z axis is kept constant value displaced only on x and y axis

plot3([0 0], [-1 1], [0 0], 'k-','LineWidth', 2);

plot3([-1 1], [0 0], [0 0], 'k-','LineWidth', 2);

plot3([0 0], [0 0], [-1 1], 'k-','LineWidth', 2);

xlabel('LM'); ylabel('S'); zlabel('LD');

axis([-1 1 -1 1 -1 1])

vecx = -1 : 0.50 : 1; vecy = -1 : 0.50 : 1; vecz = -1 : 0.50 : 1;

grid on;

set(gca, ...

'XTick' , vecx , ...

'YTick' , vecy , ...

'ZTick' , vecz );

LMaxes1 = (MyImldrgyvCol(2,:));

MedLM = median(LMaxes1);

figure(3),imhist(LMaxes1)

figure(4),imhist(MyImldrgyvCol(2,:,:))

DKL = reshape(MyImldrgyvCol, [x,y,z])

imshow(DKL)

I get this error

Error using \

Matrix dimensions must agree.

Error in trycolordkl (line 19)

MyImldrgyvCol = ldrgyv2rgbMat\(MyImrgbCol'*2 - 1);

[Malini Bakthavatchalam]

@Image Analyst , sir I am using the same code then why am I getting the same error ? could you explain me please.