Study On The Digital Audio Robust Watermarking Scheme

With the rapidly development of the digital technique in recent years, digital image, audio and video have become more popular and application, and increase the efficiency of the information broadcasting. The development of computer network technique predigests the exchange and transmission of the information more. But, with the side effect, there is despiteful individual or organization being copy, tamper and broadcasting the copyright content, without admission of the work owner. Therefore, how to carry out copyright protection and ensure the security of the information has become one of the urgent problems to be solved in digital communication.In this paper, a robust audio watermark technique is discussed. Firstly, the author introduces the meaning of the task, the research background, and the related concepts and characteristics. One typical method is reviewed as well. Secondly, the author brings forward two digital audio watermark methods based on Blind Source Separation and Singular Value Decomposition. In the first method, based on Independent Component Analysis arithmetic, the embedding watermark and the DCT coefficients of host signal are mixed in virtue of the ICA instantaneous mixing model, and obtains two observation signals, at the same time, one of them is kept as the secret. In the watermark extraction process, use the ICA algorithm to separate the estimating value of the watermark signal and the host signal, finally extracts the original watermark image. Because of the overfull secret, to reduce its information, an improved edition of method is discussed later in this paper. In the improved method, one mixed source is performed on SVD, and the singular value is kept. In the watermark extraction process, one observation signal is came back by the crossing matrix which is composes of the singular value, accordingly achieving the aim of reducing secret. Because of the out-of-order separating signals process of the ICA arithmetic, it is necessary to separate the watermark signal and the estimate value of host signal firstly, and then give a judgement that which is the estimating value of the host signal. In this way, there is more needless of work adding. Therefore, Independent Component Analysis with reference signal arithmetic is inducted in the second method, and a digital watermark method base on ICAr is brought forward. This method can separate the estimating value of the watermark signal separately by using the information provided by the reference signal, the shortcomings of separating the source signals using the ICA arithmetic, such as disorderliness and randomicity can be...