How to divide image and calculating quantized mean values of hsv and lab color channel.

 I want to divide an image into 8* 16 overlapping blocks, 
blocks are overlapping of 50% in horizontal and vertical. 
then from each block, I want to compute quantized mean values 
of HSV and lab color channel.

Answer 1

[Image Analyst]

Easily accomplished with blockproc(). See my attached demos for about 6 or 8 different ways to call blockproc(). Find the one you want and adapt it to your situation. Post your adapted code back here if you still need help.

[s malik]

I have read your tutorial and It is quite helpful. Specifically I have changed the code as under.

Img=imread('v1.tif');% specifically the size of the image is 128*48

hsvImg=rgb2hsv(Img);% as I want the mean HSV values

% meanFilterFunction = @(theBlockStructure) mean2(theBlockStructure.data(:)) ;

blockSize = [8 16];

overlap=[4 8];

blockyImage = blockproc(hsvImg, blockSize, meanFilterFunction,'BorderSize', overlap,'TrimBorder', 0);

[rows, columns] = size(blockyImage);

please tell me whether overlapping is correct as i need 50% blocks overlapping.