I do not know any built-in function in MATLAB which provides CBIR methods.

For additional help please write which color spaces and color descriptors you want to use.

This is a very simple example in the case of an RGB image (the array "a" will have three dimensions, the third dimension being the color space)

% GETTING COLOR COMPONENTS OF AN IMAGE

a=imread('file.bmp')

R= a(:,:,1); %red component

G=a(:,:,2); %green component

B=a(:,:,3); %blue component

% QUANTIFYING A PARTICULAR COLOR FEATURE:

% Let's say for example that you are interested in the relative red contens of an image

relative_red_index=sum(R(:)) / (sum(R(:))+sum(G(:))+sum(B(:))

Then every image in a database could be labeled with its quantifyed index, and a particular query could look for pictures that have similar indices. In the line of what Szabolcs Sergyan mentioned, you should first establish which color features you would like to quantify for classifying the images.

Some links for inspiration:

(CBIR) http://www.ijest.info/docs/IJEST11-03-12-156.pdf

(COLOR DESCRIPTORS): http://staff.science.uva.nl/~gevers/pub/GeversPAMI10.pdf

Thanx for reply, I want to use RGB space with color histogram as the descriptor. waiting for ur reply.

You can use color moment