Research Of Robust Watermaking Algorithm Based On Correlation Sequence

With the development of network and information technology, digital information such as images, audios and videos are spread rapidly through the internet. These kinds of information spreading, on one hand, promote the development o f digital information, but on the other hand, appear1ot of the phenomenon that abuse intellectual property rights of others. In order to prevent illegal copyrighted multimedia information dissemination and abuse, digital watermarking technology is becoming one of the main technical in copyright protection area. Digital watermark technology embeds a meaningful logo as watermark information into the digital carriers and protects the multimedia that has copyright information without prejudice to the premise of the value of carriers. At present, in the watermark information logos, correlation sequences become main digital identifications. Correlation sequences are sequences that have high auto-correlation and cross-correlation. In watermark system, use correlation sequences as copyright identification, not only do not affect the quality of the digital carriers, but also can distinguish several different levels of copyright owner. This dissertation proposed a robustness watermark algorithm by comparing the auto-correlation and cross-correlation of each sequences and selected the highest-correlation sequences for the watermark identity.Firstly, we constructed the same length of each sequences according each sequences construction method. Furthermore, using the correlation detection method compared their correlation properties. From the results we can know the complete complementary code has the highest correlation in the current correlation sequences. In our watermark embedding algorithm, we embedded a group of complete complementary code as watermark information.Secondly, in this dissertation compared auto-correlation and cross-correlation that from different construction functions of complete complementary code, and compared two kinds of complete complementary code such as "generate tree" construction function and unitary matrix construction function. At last, the proposed algorithm embedded the complete complementary code that has highest correlation properties with the proposed watermark embedding algorithm. In the proposed watermark embedding process, we verified the copyright and certificated the ownership respectively by different kinds of embedding schemes. Also, in the extracting process used complete complementary code’s correlation function to verified copyright. Furthermore, the proposed algorithm extracted ownership identity with correlation function of correlation sequences.The experiment result represents that proposed watermark algorithm has high robustness under the attacks of add-noise, rotation, compression and scaling. Hereafter, in experiment use the PSNR detects the imperceptibility. Experiment results shows that the value of PSNR is higher than40dB, which satisfy the request of imperceptibility of watermark system. At last, in order to verify the efficient of proposed algorithm, in the watermark algorithm extract the ownership identity with different group of complete complementary codes. Experiment result represents that the proposed algorithm can extract the ownership identity from the carriers correctly, which prove the efficiency of proposed algorithm.