An image watermarking scheme based on moment normalization and SVD in contourlet domain

作者：Mo Jia^*; Ma Zhaofeng; Huang Qinlong; Yang Yixian; Niu Xinxin

来源：Journal of Computational Information Systems, 2012, 8(11): 4777-4788.

• Contourlet transform
• Contourlets
• Copyright protections
• DC components
• Digital image watermarking
• Digital information
• Discrete cosine transformation
• Host images
• Image normalization
• Low frequency
• Moment normalization
• Nonoverlapping
• Singular values
• Standard images
• Sub-blocks
• Subband images
• Watermark embedding
• Watermarking schemes

Featured by its duplicability, storage and dissemination convinience, digital information has been an increasingly salient issue of copyright protection. In view of the digital watermarking as a major technology for information security, the paper aims to suggest a watermarking scheme in contourlet transform domain based on moment normalization and matrix singular value decompostion. The scheme is composed of preprocessing, watermark embedding and extracting algorithms. In the first place, the host image is normalized into the one and only standard image to get the low-frequency subband image through Contourlet decompostion. Secondly, the low-frequency subband image is divided to obtain the nonoverlapping subblocks that are further decomposed via discrete cosine transformation (DCT), of which the DC component frames the DC matrix. Finally, the resulted DC matrix is embedded into by the preprocessed watermark through singular value decomposition (SVD). Watermark extracting is an inversion to its embedding. The experiment demonstrates the scheme enjoys satisfying effectiveness in invisibility ,robustness and security.

• 出版日期2012
• 单位电子科技大学; 北京大学