Hide the message or image inside an image

                                Steganography is a technique that hides secret information inside another document or image and it is extracted by different person.

In this post, we will see how to hide text message or image inside an image without tampering the original or cover image.  Learn about bit plane slicing to know more about the background,

EXAMPLE 1: Hide Secret Message inside an image

DECRYPTION:

It is straight-forward reverse process. The secret message is obtained by extracting the least significant bits from the encrypted matrix

MATLAB CODE:

%ENCRYPTION

clear all

%READ THE INPUT IMAGE

A = imread('cameraman.tif');

%

Output = A;

%MESSAGE TO BE HIDDEN INSIDE THE IMAGE

Str = 'Image Processing! Let us learn together. Visit angeljohnsy.blogspot.com';

%NUMBER OF CHARACTERS IN THE MESSAGE

numLetters = numel(Str);

%REPRESENT THE LENGTH OF THE MESSAGE IN BINARY FORMAT

Bi_Total = de2bi(numLetters,8);

Bi_Msg = de2bi(double(Str),8)';

Encrypt_msg = Bi_Msg(:)';

%FIND THE LENGTH OF THE MESSAGE IN THE BINARY FORMAT

num_En_msg = numel(Encrypt_msg);

%DEFINE THE INTERVAL BETWEEN EACH VALUE

dt=floor(numel(A)/num_En_msg);

%BINARY VALUE OF THE INTERVAL

Bi_dt = de2bi(dt,8);

% FIRST 8 BITS - INTERVAL

% SECOND 8 BITS - TOTAL LENGTH OF THE BINARY VALUES

Info_data = [Bi_dt,Bi_Total];

Output(1:16) = bitset(A(1:16),1,Info_data);

%REPLACE THE LEAST SIGNIFICANT BIT IN THE IMAGE

Start_pt =17;

End_pt = (num_En_msg*dt)+16;

Output(Start_pt:dt:End_pt) = bitset(A(Start_pt:dt:End_pt),1,Encrypt_msg);

figure,subplot(121),imshow(A);title('Original Image');

subplot(122),imshow(Output);title('Image with the hidden message');

%SHOW THE DIFFERENCE BETWEEN THE ORIGINAL AND THE ENCRYPTED IMAGE

figure,imagesc(double(A)-double(Output));colormap(jet);colorbar;

%WRITE THE ENCRYPTED IMAGE IN THE PNG FORMAT

imwrite(Output,'Encrypt1.png');

display('Message is hidden inside the image!');

EXPLANATION:

A message is hidden inside the image by converting the message into binary format and placing it in the Least Significant Bit of the image pixels. The encrypted image is saved as ‘PNG’ image file in order to have lossless compression. With the lossless compression technique, the pixel values will not be altered so the information embedded in the LSB will not be lost.

The maximum absolute difference in value between the original and the encrypted image is one. This is obvious due to the fact that we are dealing only with the LSB bit of the image.

We can place the bits of the secret message in a specific pattern or with regular intervals instead of replacing the pixels adjacent to each other. The image shown below indicates that the bits are placed with specific interval values.

The number of characters in the secret message should be less than 10% of the cover image. This is due to the fact that for each character we need 8 bits and to place these 8 bits we need 8 pixels.

MATLAB CODE:

%DECRYPT THE MESSAGE

clear all

clc

%READ THE ENCRYPTED IMAGE

A = imread('Encrypt1.png');

%FIRST 8 BITS REPRESENT THE INTERVAL

Dt = bitget(A(1:8),1);

Dt = double(bi2de(Dt));

%LENGTH OF THE TEXT MESSAGE

Len = bitget(A(9:16),1);

Len = double(bi2de(Len));

%FETCH THE LEAST SIGNIFICANT BIT FROM THE ENCRYPTD IMAGE

Start_pt = 17;

End_pt = (Len*Dt*8)+16;

Decrypt_msg = bitget(A(Start_pt:Dt:End_pt),1);

%CONVERT THE IMAGE INTO DECIMAL

Decrypt_msg = reshape(Decrypt_msg,8,[])';

Dec_Char   = bi2de(Decrypt_msg)';

display('Text Message inside the image:');

display(char(Dec_Char));

EXPLANATION:

Decryption is the reverse process. First the interval between the pixels is fetched, followed by the length of the message.  Based on the length of the message, the bits are extracted and the final message is obtained.

EXAMPLE 2: Hide Secret image inside an image