how to divide the image into overlapping blocks?

37 ビュー (過去 30 日間)
sehreen 2014 年 2 月 11 日
コメント済み: Walter Roberson 2017 年 9 月 18 日
i need to identify the duplicate regions in an image for that i have to divide the image into overlapping blocks of fixed size lets suppose 16*16 pixels and then dct is performed on each block.
Division of blocks should be performed in such a way that if the block is a square of size b × b then the square is slid by one pixel along the image from the upper left corner right and down to the lower right please tell me how can i perform such kind of division of an image.


Walter Roberson
Walter Roberson 2014 年 2 月 11 日
See mat2cell() to divide the array into cells. Start by letting the CurrentImage be the original image. Then repeat
T = CurrentImage;
%throw away partial blocks
T(end-mod(size(T,1),16)+1:end, :) = [];
T(:, end-mod(size(T,2),16)+1:end) = [];
now break out of loop if T is empty
Now you can mat2cell to extract 16 x 16 blocks. Store them. Then,
CurrentImage = CurrentImage(2:end,2:end);
and start loop over again. You would be extracting the blocks one over and one down from the first set of blocks. Next iteration you would be going with one over and one down from that, and so on.
  1 件のコメント
sehreen 2014 年 2 月 13 日
thank u very much for your respose
[row col]=size(img);
for i=1:row-7
for j=1:col-7
sir i am using the above code to divide the image into overlapping blocks of size 8*8 . Now i need to perform
  1. DCT on each block
  2. Quantize the DCT coefficients of each block and then store each quantized block in a row of a matrix.can you please tell me how could i perform it.


その他の回答 (1 件)

Anand 2014 年 2 月 11 日
You could use the blockproc function. You would need to specify a 'BorderSize' of [1 1] and use a BlockSize of [16 16].
  4 件のコメント


Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by