Research Of Digital Audio Watermarking With Resynchronization Based On Wavelet Transform

With the extensive application and rapid development of the computer networking, Internet has become an important way to transmit information in the time of knowledge. But at the same time, pirating becomes more and more. As a new technology of copyright protection, audio watermarking has been used to protect the copyright of multimedia, such as images, digital audio and video and so on, and becomes a very active research field in Computer Science. It's a technology of information hiding by processing the audio signal and embedding watermarking information into the digital audio, and make sure that people can't feel the change of the audio.First, some groundwork and important topics about digital watermarking are discussed, and the strong points and the shortages of different methods are analyzed. Then the general algorithm of watermarking based on wavelet transform domain is discussed. And we realized one of the algorithm which is blind. Through experiment we evaluate its performances and find its shortage, i.e. it can't resist assaults of desynchronization. Based on the above mentioned research, an algorithm of watermarking with synchronization is proposed, which solutes the problem that most algorithms of watermarking cant resist assaults of desynchronization. And the algorithm has good performance of transparence and robustness, especially its performance of resisting assaults of desynchronization is heightened.The contribution of this thesis is following:1. An algorithm of audio watermarking based on DWT is realized. And the watermark is an image. This algorithm is a blind one. The watermark is embedded in DWT domain, and before embedding the watermark is encrypted to enforce the security. This algorithm is simple and general. Experiments show that the watermarking is imperceptible and rather intensive to synchronization structure.2. An method that can realize resynchronization of watermark is proposed, which takes use of frame synchronization technology and coding technology,. The frame synchronization codes are embedded into audio as part of the watermark. To heighten the effenciency of resynchronization, the synchronization code is embedded in time domain of the audio. And the number of each segment is embedded before it. So we realize the synchronization of both through synchronization codes and the numbers.3. An algorithm of digital audio watermarking with resynchronization is put forward and realized. The watermark is an image. The technique of pseudo-random permutation is used during watermark embedding process. Correct coding is appliedin order to lower the error rate of detection. Take use of the synchronization code, the position of each segment can be found. Watermark is embedded into the approximate coefficients of 3 level of DWT. Experiments show that the algorithm has good transparency and robustness against noises, filtering, resample, MP3coding, and random cropping.