I am trying to apply DCT compression on an image 1024 *1024 but I want to divide this image into 8*8 sub blocks. How can I divide this image to 8*8 sub blocks? can anybody help me please

 image_folder= 'C:\Octave\Octave-4.2.1\share\octave\packages\images' 
filename=dir(fullfile(image_folder,'*.jpg'))
n = 4
f=fullfile(image_folder,filename(n).name)
my_image=imread(f)
figure()
imshow(my_image)

Answer 1

[David Goodmanson]

Hi John,

Here is one approach. For 8x8 blocks, you basically have a 256x256 matrix where each 'element' is an 8x8 block. Then you can get the contents of the block in row i (out of 256 rows) and column j (out of 256 columns) with

Q = rand(1024,1024);       % original matrix  
R = zeros(size(Q)); 
bi = @(i) 8*(i-1)+(1:8);   % block index function

i = 7; j = 77;             % for example
q = Q(bi(i),bi(j))         % 8x8 block

Suppose you change q to qnew by some process. Then you can use for loops i,j and fill in R (or the original Q) with

R(bi(i),bi(j)) = qnew;

To make the bi call as short as possible, bi was hardwired for 8x8. For a general block of size mxn,

bi = @(n,i) n*(i-1)+(1:n);   % block index function

q = Q(bi(m,i),bi(n,j))
R(bi(m,i),bi(n,j)) = qnew;

Other approaches are possible, such as stacking all the blocks into an 8x8x65536 array. but getting the blocks in and out is less easy.