dct

% Script file:dctblock.m
%
% Purpose:an algorithm  to  achieve the embeding watermarking  to  a image by
% dct bloking  and  hvs, then  detect the watermarking by
% comparability.
%
%
% Record of revisions:
%  Date  Programmer Description of change
%  ====   ==========   ===================== 
%  11 / 19 / 03  Mao Li Original code
% Define variables:
% a0  -- Original gray image matrix
% a1  -- Embeded gray image matrix
% da0  -- the dct original gray image matrix
% ca0  -- the column vector of a0
% cda0  -- the column vector of da0
% cda1  -- the embeded column vector of cda0
% da1  -- the image matrix of cda1
% r  -- image matrix row
% c  -- image matrix column
% k  -- the number of blocks
% i  -- the subscript of a matrix
% SNR  -- the similarity degree

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% input original data ,block break  and  convert a block into a column vector %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear;
clc;
a0 = imread( ' c:oat.png'); 
 %a0 = rgb2gray(a0);
[r,c] = size(a0);
k = (r * c / 64 );
da0 = blkproc(a0,[ 8 , 8 ], ' dct2'); 
 ca0 = im2col(a0,[ 8 , 8 ], ' distinct'); 
 cda0 = im2col(da0,[ 8 , 8 ], ' distinct'); 
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% produce the watermarking signal (random sequence) %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
randn( ' state',110); 
 w0 = randn( 1 , 5120 );
w0 = reshape(w0, 5 , 1024 );

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% changing submultiple %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
alpha = 0.02 ;

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% embed the watermarking into the low frequences coefficients %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
cda1 = cda0;
 for  i = 1 :k
cda1( 2 ,i) = cda0( 2 ,i) + alpha * w0( 1 ,i);
cda1( 3 ,i) = cda0( 3 ,i) + alpha * w0( 2 ,i);
cda1( 9 ,i) = cda0( 9 ,i) + alpha * w0( 3 ,i);
cda1( 10 ,i) = cda0( 10 ,i) + alpha * w0( 4 ,i);
cda1( 17 ,i) = cda0( 17 ,i) + alpha * w0( 5 ,i);
 end 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% reconstruct the watermarking %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
da1 = col2im(cda1,[ 8 , 8 ],[r,c,], ' distinct'); 
 a1 = blkproc(da1,[ 8 , 8 ], ' idct2'); 
 figure;
subplot( 1 , 2 , 1 ),imshow(a0,[]),title( ' the original image'); 
 subplot( 1 , 2 , 2 ),imshow(a1,[]),title( ' the embeded image'); 
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% detect the watemarking %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
dca0 = blkproc(a0,[ 8 , 8 ], ' dct2'); 
 dca1 = blkproc(a1,[ 8 , 8 ], ' dct2'); 
 cdca0 = im2col(dca0,[ 8 , 8 ], ' distinct'); 
 cdca1 = im2col(dca1,[ 8 , 8 ], ' distinct'); 
 for  i = 1 :k
w1( 1 ,i) = (cdca1( 2 ,i) - cdca0( 2 ,i)) / alpha;
w1( 2 ,i) = (cdca1( 3 ,i) - cdca0( 3 ,i)) / alpha;
w1( 3 ,i) = (cdca1( 9 ,i) - cdca0( 9 ,i)) / alpha;
w1( 4 ,i) = (cdca1( 10 ,i) - cdca0( 10 ,i)) / alpha;
w1( 5 ,i) = (cdca1( 17 ,i) - cdca0( 17 ,i)) / alpha;
 end 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% calculate the similarity degree %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
SNR1 = sum(sum(w0. * w1)) / sqrt(sum(sum(w1. ^ 2 )))


  



 
% Script file:dct_high.m
%
% Purpose:an algorithm  to  achieve the embeding watermarking  to  a image by
% dct bloking  and  hvs, then  detect the watermarking by
% comparability.
%
%
% Record of revisions:
%  Date  Programmer Description of change
%  ====   ==========   ===================== 
%  11 / 25 / 03  Mao Li Original code
% Define variables:
% a0  -- Original gray image matrix
% a1  -- Embeded gray image matrix
% da0  -- the dct original gray image matrix
% ca0  -- the column vector of a0
% cda0  -- the column vector of da0
% cda1  -- the embeded column vector of cda0
% da1  -- the image matrix of cda1
% r  -- image matrix row
% c  -- image matrix column
% k  -- the number of blocks
% i  -- the subscript of a matrix
% SNR  -- the similarity degree

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% input original data ,block break  and  convert a block into a column vector %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%clear;
%clc;
a0 = imread( ' c:oat.png'); 
 %a0 = rgb2gray(a0);
[r,c] = size(a0);
k = (r * c / 64 );
da0 = blkproc(a0,[ 8 , 8 ], ' dct2'); 
 ca0 = im2col(a0,[ 8 , 8 ], ' distinct'); 
 cda0 = im2col(da0,[ 8 , 8 ], ' distinct'); 
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% produce the watermarking signal (random sequence) %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
randn( ' state',110); 
 w0 = randn( 1 , 5120 );
w0 = reshape(w0, 5 , 1024 );

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% changing submultiple %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
alpha = 0.02 ;

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% embed the watermarking into the low frequences coefficients %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
cda1 = cda0;
 for  i = 1 :k
cda1( 48 ,i) = cda0( 48 ,i) + alpha * w0( 1 ,i);
cda1( 55 ,i) = cda0( 55 ,i) + alpha * w0( 2 ,i);
cda1( 56 ,i) = cda0( 56 ,i) + alpha * w0( 3 ,i);
cda1( 62 ,i) = cda0( 62 ,i) + alpha * w0( 4 ,i);
cda1( 63 ,i) = cda0( 63 ,i) + alpha * w0( 5 ,i);
 end 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% reconstruct the watermarking %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
da1 = col2im(cda1,[ 8 , 8 ],[r,c,], ' distinct'); 
 a1 = blkproc(da1,[ 8 , 8 ], ' idct2'); 
 figure;
subplot( 1 , 2 , 1 ),imshow(a0,[]),title( ' the original image'); 
 subplot( 1 , 2 , 2 ),imshow(a1,[]),title( ' the embeded image'); 
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% detect the watemarking %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
dca0 = blkproc(a0,[ 8 , 8 ], ' dct2'); 
 dca1 = blkproc(a1,[ 8 , 8 ], ' dct2'); 
 cdca0 = im2col(dca0,[ 8 , 8 ], ' distinct'); 
 cdca1 = im2col(dca1,[ 8 , 8 ], ' distinct'); 
 for  i = 1 :k
w1( 1 ,i) = (cdca1( 48 ,i) - cdca0( 48 ,i)) / alpha;
w1( 2 ,i) = (cdca1( 55 ,i) - cdca0( 55 ,i)) / alpha;
w1( 3 ,i) = (cdca1( 56 ,i) - cdca0( 56 ,i)) / alpha;
w1( 4 ,i) = (cdca1( 62 ,i) - cdca0( 62 ,i)) / alpha;
w1( 5 ,i) = (cdca1( 63 ,i) - cdca0( 63 ,i)) / alpha;
 end 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% calculate the similarity degree %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
SNR = sum(sum(w0. * w1)) / sqrt(sum(sum(w1. ^ 2 )))