How can I use one plane of RGB as a mask for other planes and generate new image?

Original = imread('Image.tif');
[x,y,n]=size(Original);
% %-------------------------------Red Plane
Red(:,:,2)=zeros(x,y);
Red(:,:,3)=zeros(x,y);
Red(:,:,1)= Original(:,:,1);
%--------------------------------- Green Plane
Green(:,:,3)=zeros(x,y);
Green(:,:,1)=zeros(x,y);
Green(:,:,2)= Original(:,:,2);
% -- - - - - - - - - -------------- Blue Plane
Blue(:,:,2)=zeros(x,y);
Blue(:,:,1)=zeros(x,y);
Blue(:,:,3)= Original(:,:,3);
%-------------------------converts the green planes to Black and White to make mask
GreenBW= im2bw(Original(:,:,2));

OriginalMasked(:,:,1) = GreenBW.*Red(:,:,1);
OriginalMasked(:,:,2) = Green(:,:,2);
OriginalMasked(:,:,3) = GreenBW.*Blue(:,:,3);
imshow(OriginalMasked);

[Walter Roberson]

What you have looks okay.

[Kamran Moradi]

The numbers are the same as you can see in in the command picture, but the result is different, that region is yellow but it is white in masked one. The variable format has been changed also from unit8 to double. Still confusion...

Answer 1

[Walter Roberson]

When you are creating Red, Green, Blue, copy over Original first so that you get the correct datatype for the destination. Currently you are copying it last.

[Kamran Moradi]

this is what I get when I use: Original = imread('Image.tif');

GreenBW= repmat(Original(:,:,2),[1,1,3]);

maskedRgbImage = bsxfun(@times, Original, cast(GreenBW, 'like', Original));

Can you explain more your last answer. I am beginner in image processing. Thanks

[Walter Roberson]

In the code you posted, you used

Red(:,:,2)=zeros(x,y);
Red(:,:,3)=zeros(x,y);
Red(:,:,1)= Original(:,:,1);

If you change that to

Red(:,:,1)= Original(:,:,1);
Red(:,:,2)=zeros(x,y);
Red(:,:,3)=zeros(x,y);

then the array will be created with the same data type as Original instead of being created with the default double() data type.

[Kamran Moradi]

Thank you for your great advises. It Worked.

Original = imread('Image.tif');

[x,y]=size(Original);

% %-------------------------------Red Plane

Red(:,:,1)= Original(:,:,1);

Red(:,:,2)=0;

Red(:,:,3)=0;

% %--------------------------------- Green Plane

Green(:,:,2)= Original(:,:,2);

Green(:,:,3)=0;

% % -- - - - - - - - - -------------- Blue Plane

Blue(:,:,3)= Original(:,:,3);

GreenBW= im2bw(Original(:,:,2));

OriginalMasked = bsxfun(@times, Original, cast(GreenBW, 'like', Original));

imshow(OriginalMasked);

[Image Analyst]

Be careful about using non-standard definitions for x and y like you did. That usually eventually leads to trouble. Most people would do

[y, x] = size(TwoDMatrix);

Of course you didn't do that, and you did something even more risky or dangerous - you used it on an image read in from imread(). Why is it dangerous to use

[x, y] = size(theImage);

instead of

[rows, columns, numberOfColorChannels] = size(theImage);

(besides just swapping x and y)? Well, read this: http://blogs.mathworks.com/steve/2011/03/22/too-much-information-about-the-size-function/

If you don't use a third argument on an image array, and the image array is a color image, the second argument ("x" or "y" as you called it) will actually be the number of columns multiplied by the number of color channels.

Answer 2

[Image Analyst]

An alternate way to mask an RGB image by a logical 2D mask image, promoted by Sean, is this:

% Mask the image using bsxfun() function
maskedRgbImage = bsxfun(@times, rgbImage, cast(mask, 'like', rgbImage));

[Kamran Moradi]

Thank you for the answer, but the result is the same as above. The numbers are the same as you can see in in the command picture, but the result is different, that region is yellow but it is white in masked one. Still confusion... This command is faster though and does not have variable type problem

Answer 3

[Negesse Tadesse]

does it help for image compresson?

[Image Analyst]

Does what help with image compression? Masking? It will make the images smaller because much of the image is replaced by a constant number.