Distortionless Data Hiding Based on Integer Wavelet Transform

Lossless data hiding using Integer Wavelet transform by Guorong Xuan. (Reversible Watermarking)

Description

The code implements the following paper:

Guorong Xuan, Jiang Zhu, Jidong Chen, Shi Y.Q., Zhicheng Ni, Wei Su, "Distortionless data hiding based on integer wavelet transform" IET Electronics Letters, Volume: 38, Issue 25, page 1646- 1648, December 2002. (ISSN: 0013-5194)

The watermark embedding and extraction modules have been implemented separately.

The watermark embedding module embeds the watermark using bitplane compression in the integer wavelet transform domain.

The watermark extraction module extracts the previously embedded watermark and restores the original image.
 
A logo image is used as a watermark but can be replaced with any thing.

• MATLAB release: MATLAB 7 (R14)

[Pon Selvalingam]

 

[lavanya ananthan]

Watermark can't extracted from distorted watermark (eg distorted by speckle noise) any help?

[Asad Ali]

In order to extract watermark after some distortion has taken place you have to use some error correcting codes such as BCH encoding. The watermark will first be BCH encoded and then embedded into the image. So at a later time while extracting after addition of noise the error correcting codes will compensate for the errors and recover the correct watermark. Check the matlab function "bchenc".

[lavanya ananthan]

After adding noise to whole image,i am getting following errors
>The symbol counts parameter must be a vector of positive finite integers.
>CH5 = arithdeco(seq11,count1,size(CH,1)*size(CH,2));
 
help?