Research Of Digital Audio Watermarking Algorithm In Wavelet Domain Based On DC-DM

Along with the rapid development of network, communication and computer technology, the production, dissemination and acquisition of digital multimedia data become more simple and quick. Simultaneously, the illegal use, tampering, forgery and other issues also emerge in endlessly. Such behaviors not only violate the owner’s personal interest, but also pose a serious threat to information security. Faced with these problems, solutions are widely sought, and then the digital watermarking technology is proposed. As a new copyright protection technology, the authenticity and certification of multimedia products are realized by digital watermarking technology. Its implementation is to embed specific information into the digital products like audio, video and images by a particular way. When certification is required, specific information can be extracted. Those specific information are referred to as watermarks, watermarks can be product-related information, or just random numbers that not contain any meaning, the embedding will not affect the use of digital products.Since Digital watermarking technology has been put forward, it has grabbed extensive attention from domestic and overseas scholars. Now it has become a hot research hotspot in the field of information security and information hiding. At present, the study and implementation of digital watermarking mainly concentrate in image and video watermarking domain. Because human auditory perception system is more sensitive than visual perception system, it’s difficult to embed watermarks to audio signal, and related technology is not mature yet, so the study of digital audio watermarking technology is still in the stage of theoretical discussion.Digital watermarking technology has been taken as the research content in this thesis. Through the research of audio signal theory and digital watermarking technology, and the summary of algorithms, a robust digital watermarking algorithm that can be used to copyright protection has been proposed. The advantages of wavelet transform in audio processing and the stability of mean value have been well utilized in this algorithm; a binary image is embedded into the mean of coefficients in low frequency domain. The DC-DM algorithm is chose as the quantization method, it can obtain the tradeoff between robust, perceptual and bit error rate. Before embedding, Arnold transformation and Rossler chaotic encryption have been applied to watermark image, the security of watermark is improved. This algorithm is a blind watermarking algorithm, no original audio signal is needed in extraction process.Extensive and in-depth simulations and analysis are carried out in this thesis. Simulations on different types of audio data show that the algorithm is compatible and applicable. Perceptual evaluations of results indicate that the algorithm is robust and inaudible.