how to convert grey image to RGB

Hi all,

i have 60 slice grey image. how want to convert to RGB image?

as attached is my grey image (grey.jpg). rgb.jpg is from internet what i want like.

this is my coding, but got error.

P = zeros(103, 103, 60);
for K = 1 : 60
    K_file=30+10*K;
    petname = sprintf('I%d.dcm', K_file);
    P(:,:,K) = dicomread(petname);
         scale = 130/103 ;
 Pi(:,:,K) = imresize(P(:,:,K),scale) ; % where P is your 103*103 3D matrix
 
end
rgbImage = gray2rgb(Pi);
imshow3D(rgbImage)

this is my function

function [Image]=gray2rgb(Image)
%Gives a grayscale image an extra dimension
%in order to use color within it
[m n]=size(Image);
rgb=zeros(m,n,3);
rgb(:,:,1)=Image;
rgb(:,:,2)=rgb(:,:,1);
rgb(:,:,3)=rgb(:,:,1);
Image=rgb/255;
end

BUT GOT ERROR.

Unable to perform assignment because the size of the left side is 130-by-7800 and the size of
the right side is 130-by-130-by-60.
Error in gray2rgb (line 6)
rgb(:,:,1)=Image;
Error in sliderspect1 (line 12)
rgbImage = gray2rgb(Pi);

I THINK MY function gray2rgb wrong.

can some one help me?

Answer 1

[Jan]

Your Pi is a [130 x 130 x 60] matrix, but your function gray2rgb() expects a 2D matrix as input. It is not clear how you want to convert the 60 layers of the image data to RGB channels.

Maybe you want:

[r,c,m]  = size(Pi);
rgbImage = repmat(reshape(Pi / 255, [r, c, 1, m]), [1, 1, 3, 1]);

[mohd akmal masud]

Hi Jan,

i have change my coding accordingly your suggestion. But my image still grey as picture attached.

clc
clear all
P = zeros(103, 103, 60);
for K = 1 : 60
    K_file=30+10*K;
    petname = sprintf('I%d.dcm', K_file);
    P(:,:,K) = dicomread(petname);
         scale = 130/103 ;
 Pi(:,:,K) = imresize(P(:,:,K),scale) ; % where P is your 103*103 3D matrix
 
end

rgbImage = gray2rgb(Pi);
imshow3D(rgbImage);

below is my function i changed.

function [Pi]=gray2rgb(Pi)
%Gives a grayscale image an extra dimension
%in order to use color within it
[r,c,m]  = size(Pi);
rgbImage = repmat(reshape(Pi / 255, [r, c, 1, m]), [1, 1, 3, 1]);

end

CAN HELP ME?

[Jan]

Yes, of course the image is grey, because the original images contain the grey level information only. If the value of each pixel of the input is used for all 3 RGB channels, the output looks grey also.

Which information do you want to use to define the colors?

Answer 2

[Image Analyst]

Try

function rgbImage = gray2rgb(grayImage)
cmap = jet(256); % Or whatever one you want.
rgbImage = ind2rgb(grayImage, cmap); % Convert gray scale image into an RGB image.

[mohd akmal masud]

Hi Image Analyst,

i have try, but error,

Index in position 1 is invalid. Array indices must be positive integers or logical values.

Error in matlab.images.internal.ind2rgb (line 24)
rgb(1:height, 1:width, 1) = reshape(cm(indexedImage, 1), [height width]);

Error in ind2rgb (line 17)
    r = matlab.images.internal.ind2rgb(a, cm);

Error in gray2rgb (line 5)
rgbImage = ind2rgb(grayImage, cmap); % Convert gray scale image into an RGB image.

Error in sliderspect1 (line 13)
rgbImage = gray2rgb(Pi);

[mohd akmal masud]

this is my coding

clc
clear all
P = zeros(103, 103, 60);
for K = 1 : 60
    K_file=30+10*K;
    petname = sprintf('I%d.dcm', K_file);
    P(:,:,K) = dicomread(petname);
         scale = 130/103 ;
 Pi(:,:,K) = imresize(P(:,:,K),scale) ; % where P is your 103*103 3D matrix
 
end

rgbImage = gray2rgb(Pi);
imshow3D(rgbImage);

this is my function

function rgbImage = gray2rgb(Pi)
cmap = jet(256); % Or whatever one you want.
rgbImage = ind2rgb(Pi, cmap); % Convert gray scale image into an RGB image.

[Image Analyst]

To convert a gray scale image to an RGB image, it must have one channel. You have 60 channels, like either it's a hyperspectral image with 60 color channels, or a volumetric image, like from CT or MRI, with 60 slices. Neither of those can be converted into an RGB image. You either need three channels, or one channel, not 60.

[Carla Bariot]

Hello Image Analyst,

I have a question : how did you decide to take cmap = jet(256) ?

I try this code on a image with a lot of green I convert it in to grayscale image et after applied gray2rdg function an my image is red instead of green.

[DGM]

If you deleted all the color information in your image, what makes you expect to get it back?

The selection of jet() or any other color map is arbitrary.

[Image Analyst]

Carla, like I said, you can pick whatever colormap you want. I just arbitrarily picked jet because it works and gives an image with a lot of vivid colors. Of course you won't get back your original image with an arbitrary color map.

If you want, you can convert to an indexed image with rgb2ind() and that will return a colormap that you can use to get somewhat close to your original image though not exact since you'll only have a few colors (like 256) instead of all of them (up to 16.7 million for an 8 bit image). And if you view the indexed image as a gray scale image, it won't look anything like you expect it to because the image values are indexes (rows) into a color map rather than intensity (reflectance or luminance) values of the scene.

[john]

how do i convert grayscale image to same color image but it should be an rbg image?

I tried the above function which u gave. My input was a grayimage with just 0 and 255 values when i used above function and runned the program it got converted to blue and brown. But i want the same original image but it should be an rgb so how do i get it?

[Image Analyst]

@john if you have an image with just 0 and 255, and use ind2rgb() then you will end up with an RGB color image with only two colors - those colors being cmap(1,:) and cmap(255, :), which I guess are ble and brown for the colormap you used. There is no way to get back the original true color RGB image from only a gray scale version of it. as you can imagine, there are nearly an infinite number of color images that could be made from it. You'll just have to save your original RGB image if you want it later in your program.

Answer 3

[Jamal Nasir]

there is no traditional method to convert gray image to color image

this need using machine learning for prediction the pixel value depend on the region in

similar sense

[Image Analyst]

He has a 60 channel hyperspectral image, as you can see from his code. One can use one of the strategies in the attached paper to convert multiple spectral channels into an RGB image.