Shrinking image by averaging

I want to create a function: shrinkbyaveraging(imagematrix) that returns an image that is half as large in both height and width. Each pixel in the returned image should be the average of 4 pixels from the input image. Shrinkbyaveraging only needs to work for images having an even number of rows/columns. However, I do not want to use imresize.

function shrink = shrinkbyaveraging(imagematrix) [row,colm] = size(imagematrix); x = row/2; y = colm/2;

but im not sure what the question is asking since i cannot use imresize. :/

[KALYAN ACHARJYA]

What you have done so far, please share?

[Image Analyst]

If you do not want to use imresize(), then I guess you do not want to use interp2() either, and do not want to use anyone else's function or code. So, how can we help? Point you to Wikipedia's page on bilinear interpolation??? And to this page on how to format your code?

[Rik]

[Rik]

Also some deleted comments :

OP responding to Kalyan:

function shrink = shrinkbyaveraging(imagematrix) [row,colm] = size(imagematrix); x = row/2; y = colm/2;

but im not sure what the question is asking since i cannot use imresize. :/

OP responding to Image Analyst:

I am taking a first year introductory computing science course and I dont think it is that hard. But, I am stuck on it.

Answer 1

[Michael Hawks]

Try this:

Nrow = 12; Ncol = 15; 
input = rand(Nrow,Ncol); 
shifts = [0 0;  0 1; 1 0; 0 -1]; 
A = zeros(Nrow,Ncol,4); 
for ii = 1:4;
    out(:,:,ii) = circshift( input, shifts(ii,:) ); 
end
out = sum( out, 3); 
out = out(2:2:end,2:2:end);

In the for loop, we're building up an mxnx4 array of shifted copies of the input array. Then finding the mean over the neighborhood is as easy as averaging along dimension 3. Then the last line downsamples the array to the size you want.

Answer 2

[Sean de Wolski]

This works

img = repmat(1:6,4,1)
shrink = @(x)reshape(sum(reshape(sum(reshape(x,2,1,[])),size(x,1)/2,2,[]),2)./4,size(x,1)/2,[])
shrink(img)