# i've divided the image into non-overlapping blocks and i want to calculate the mean of every block...

6 views (last 30 days)
angel on 14 Feb 2013
after calculating the mean i;ve to store it in array corresponding to their block number
plz tell me how to do?
Mean=[];
for row = 1 : blockSizeR : rows
for col = 1 : blockSizeC : columns
row1 = row;
row2 = row1 + blockSizeR - 1;
row2 = min(rows, row2);
col1 = col;
col2 = col1 + blockSizeC - 1;
col2 = min(columns, col2);
oneBlock = grayImage(row1:row2, col1:col2);
m=mean(mean(oneBlock));
M=m;
Mean=[Mean;M];
disp(Mean);
end
end
the above code shows only the mean value... but i want that it also shows the block number corresponding to the mean value...
plz rectify or modify the above code in order to do so.

Image Analyst on 14 Feb 2013
Not sure how you want to define block number but how about:
blockNumber = 1;
for row = 1 : blockSizeR : rows
for col = 1 : blockSizeC : columns
row1 = row;
row2 = row1 + blockSizeR - 1;
row2 = min(rows, row2);
col1 = col;
col2 = col1 + blockSizeC - 1;
col2 = min(columns, col2);
oneBlock = grayImage(row1:row2, col1:col2);
Mean(blockNumber) = mean2(oneBlock);
fprintf('The mean for block #%d = \$f', ...
blockNumber, Mean(blockNumber));
blockNumber = blockNumber + 1;
end
end

angel on 15 Feb 2013
can u code it sir for me plz.... u mean
S=sort(Mean(blockNumber);
fprintf('\nblock #%d = %f',blockNumber,S);
if not, correct it plz
Walter Roberson on 15 Feb 2013
[S, blockNumber] = sort(Mean);
fprintf('\nBlock #%d = %f', [blockNumber(:), S(:)].' );
angel on 15 Feb 2013
thank u so much sir :)
ur help means a lot to me.

Abbas Cheddad on 14 Feb 2013
Hi there,
You could do the following:
%%%%%%%%%%%%%
fun = @(block_struct) mean2(block_struct.data);
I2 = blockproc(A,[8 8],fun);
Block_Mean(:,1)=reshape(I2,[1 size(I2,1)*size(I2,2)]);
Block_Mean(:,2)=1:length(Block_Mean);
%%%%%%%%%%%%%%
Where "A" denotes your image and [8 8] is the size of the block of your choice. This will result in "Block_Mean" which holds the mean intensity and the corresponding block number.

angel on 14 Feb 2013
@ Abbas: ya sir but block_proc returns single image after processing all the blocks..i've given the code above i have modified it as the image anaylst told me... i m trying to applying the sort function but not stricking to me how to apply
Abbas Cheddad on 14 Feb 2013
Block_proc returns a single image with the same size as "A" if you use the documented code as in:
fun = @(block_struct) ...
mean2(block_struct.data) * ones(size(block_struct.data));
But in the code I have given you, each pixel in "I2" corresponds to the mean of a single block.
Anyhow, since what you have in hand works for you then happy days!

Jan on 14 Feb 2013
The fast C-MEX function FEX: BlockMean performs this operation.
R = BlockMean(img, blockSizeR, blockSizeC);

angel on 14 Feb 2013
this shows error:
Undefined function 'BlockMean' for input arguments of type 'uint8'. input image is 256 x 256
sir plz help me in sorting the mean values which i've calculated...
plz go through the above code that i've mentioned and tell me..
Walter Roberson on 15 Feb 2013
You need to download and install that code in order to have a BlockMean routine.
angel on 15 Feb 2013
ok sir :)
regards Angel