Stegnaography using DCT

this is the code we are using to implement steganography based on DCT....

I = imread('lena_std.jpg');

I = im2double(I(:,:,1));

I2=imread('test.bmp');

subplot(2,2,1),imshow(I);

subplot(2,2,2),imshow(I2);

I2=de2bi(I2(:,:,1));

x=reshape(I2',1,38400);

T = dctmtx(8);

dct = @(x)T * x * T';

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

k=0;

for i=1:256

      for j=1:256

          if B(i,j)>0

              b=dec2bin(typecast(single(B(i,j)),'uint32'),32);

              b(1,7:8)=x(1,k*2+1:k*2+2);

              k=k+1;

              c=double(typecast(uint32(bin2dec(b)),'single') );

              B(i,j)=c;

          end

      end

end

subplot(2,2,3),imshow(B);

invdct = @(x)T' * x * T;

B3= blkproc(B,[2 2],invdct);

subplot(2,2,4),imshow(B3);

could somebody tell me the problem in this code with respect to the binary and decimal conversion....

[kbae]

could you give me the extraction code for this embedding method?thank you

[Ajmal Mohammed V M]

Could u pls share the extraction code for the above problem

Answer 1

[Walter Roberson]

You will find that de2bi() does not return the characters '0' and '1' that are required for bin2dec().

You will also find that the bit ordering is different between de2bi() and dec2bin(). Compare

dec2bin(8,5)
de2bi(8,5)

[Pranoy]

ya...so how do i convert my payload image to binary to embed it in the cover image??

[Walter Roberson]

Mostly, be consistent about whether you use de2bi() or dec2bin()

Answer 2

[NILANJ]

I2=de2bi(I2(:,:,1)); can any one tell me plz that in this function why (:,:,1) ???

[Walter Roberson]

First bit plane -- corresponding to Red in an RGB image.

Answer 3

[Maulana Wahid]

could you please explain your DCT code step by step?

Answer 4

[Peter Eze]

use transpose(dec2bin(ImageMessage(i,j))) to create the binary string for each pixel of the message to be embedded

Answer 5

[juhi patel]

de2bi()is used for you want your data in matrix form.&dec2bin() is used if you want output in string format.you are mixing both function in this code.using de2bi() will solve your problem since image is in matrix form.

[Walter Roberson]

I almost always use dec2bin() and subtract '0' to get matrix form. Then I do not have to worry about the fact that de2bi() is part of a communications toolbox rather than part of basic MATLAB.

Answer 6

[Fady Samann]

First, you are not doing quantization to remove the samller DCT coefficients after that you can change LSB by masking it and and OR it with the message bit. Use bitwise operations, because dec2bin and bin2dec treat the binary as a string of 0's and 1's.

https://en.wikipedia.org/wiki/Quantization_(image_processing)#Quantization_matrices

the process is just like JPG compression but in the middle of quantization and dequantization, you can change the LSB.

https://www.mathworks.com/matlabcentral/answers/217004-need-help-on-my-dct-and-quantization-code-for-image-compression

Second, when you pick the right DCT coefficient to change its LSB, the coefficient must be greater than one because of changing the LSB of one to zero (from the message bits) will set the coefficient to zero then during the decoding process this coefficient will be skipped because of the if condition (B(i,j)>0).

Note: round the number only after the quantization stage in the encoding and decoding of the message.