Need help on my DCT and Quantization code for Image Compression.

Hello guys, I would really appreciate it if anyone could point out the mistakes in my code. I am trying to encode and decode an image by reading it in, performing DCT, Quantization then dequantizing it and performing inverse DCT. After running this code, the output Image, I2 is kind of pixellated. I have no idea how to fix it. The output should be somewhat similar to the original image but slightly blurred as it has undergone compression. Please help! My code is as follows :-

I = imread('cameraman.tif');
 I = im2double(I);
T = dctmtx(8); % dct matrix

%Performing DCT on blocks of 8 by 8
dct = @(block_struct) T * block_struct.data * T';
B = blockproc(I,[8 8],dct);

B = ceil(B); 

% A Standard Quantization Matrix
q_mtx =     [16 11 10 16 24 40 51 61; 
            12 12 14 19 26 58 60 55;
            14 13 16 24 40 57 69 56; 
            14 17 22 29 51 87 80 62;
            18 22 37 56 68 109 103 77;
            24 35 55 64 81 104 113 92;
            49 64 78 87 103 121 120 101;
            72 92 95 98 112 100 103 99];

   %PErforming Quantization by Dividing with q_mtx on blocks of 8 by 8
   c = @(block_struct) (block_struct.data) ./ q_mtx;        
   B2 = blockproc(B,[8 8],c);
  %    B2 = ceil(B2)

%Performing Inverse Quantization By Multiplying with q_mtx on Blocks of 8
%by 8
B3 = blockproc(B2,[8 8],@(block_struct) q_mtx .* block_struct.data);

%Performing Inverse DCT on Blocks of 8 by 8
invdct = @(block_struct) T' * block_struct.data * T;
% B3 = ceil(B3);
I2 = blockproc(B3,[8 8],invdct);
imshow(I), figure, imshow(I2)

[matlabgeek]

Have you fixed this problem?

Answer 1

[Hiroshi Nakamura]

Just in case you're wondering after running the code, I2 looks like this :-

Please help!

[B.k Sumedha]

I think this link might help. http://in.mathworks.com/help/wavelet/ref/wcompress.html

Answer 2

[B.k Sumedha]

I = imread('cameraman.tif');

I = im2double(I);

T = dctmtx(8);

B = blkproc(I,[8 8],'P1*x*P2',T,T');

mask = [1   1   1   1   0   0   0   0

          1   1   1   0   0   0   0   0

          1   1   0   0   0   0   0   0

          1   0   0   0   0   0   0   0

          0   0   0   0   0   0   0   0

          0   0   0   0   0   0   0   0

          0   0   0   0   0   0   0   0

          0   0   0   0   0   0   0   0];

B2 = blkproc(B,[8 8],'P1.*x',mask);

I2 = blkproc(B2,[8 8],'P1*x*P2',T',T);

imshow(I), figure, imshow(I2)

[Hiroshi Nakamura]

Thanks but after looking up Quantization Matrix on Wiki, I should divide it by some Quantization Matrix other than what you have shown me. Thanks anyway!

[B.k Sumedha]

If the answer is accepted then pleae click on accepted answer:)

Answer 3

[Prathyusha K]

I = imread('cameraman.tif');

I = double(I);

T = dctmtx(8); % dct matrix

dct = @(block_struct) T * block_struct.data * T';

B = blockproc(I,[8 8],dct);

q_mtx = [16 11 10 16 24 40 51 61; 12 12 14 19 26 58 60 55; 14 13 16 24 40 57 69 56; 14 17 22 29 51 87 80 62; 18 22 37 56 68 109 103 77; 24 35 55 64 81 104 113 92; 49 64 78 87 103 121 120 101; 72 92 95 98 112 100 103 99];

c = @(block_struct)(block_struct.data) ./ q_mtx;

B2 = blockproc(B,[8 8],c);

B2 = round(B2);

B3 = blockproc(B2,[8 8],@(block_struct) q_mtx .* block_struct.data);

invdct = @(block_struct) round(T' * block_struct.data * T);

I2 = blockproc(B3,[8 8],invdct);

imagesc(I); colormap gray; figure, imagesc(I2), colormap gray;