The Application Of The Wavelet Transform In The Digital Watermarking Technique

Digital watermarking is a technique for labelling digital media through hiding secret information in media and the embedded information can neither be removed nor decoded without the required secret key.The main work of this paper is to propose a novel multiscale-edge-based digital watermarking method in the wavelet transform modulus maxima domain. We employ the nearly perfect reconstruction and singularity characterization proper-ties of the wavelet transform modulus maxima representation (WTMM) to obtain a sparse image representation, where the significant coefficients represent the perceptually important structural edges of the image. The proposed method gives a framework for multiscale-edge-based watermarking schemes, which utilized the wavelet transform modulus maxima to construct the multirescale frequency components and perceptually significant multiscale structural edges as embedd-ing positions.In the proposed method, the embedding scheme combines the dyadic wavelet transform coefficients with the spatial information, and uses both the modulus maxima value and the Lipschitz regularity as the criterion to select the embedding positions, which conform to the Human Visual System model, so desirable characteristics of the watermark could be exhibited. The experimental results show that the proposed algorithm provides extra robustness against JPEG compression and image processing attacks.