The Audio Watermarking Method Research Based On New Multi-scale Transforms

With the rapid development of information technology, people can get digital information more conveniently. But it brought a series of network information security problems at the same time, for example, the illegal copying and intentional tampering of multimedia information. So the importance and urgency of digital media intellectual property right protection is more and more obvious nowadays. Digital watermarking technology arises at the historic moment. This paper revolves "the multimedia information security technology research based on novel multi-scale transforms digital watermarking", image and audio copyright protection technology were studied. This paper mainly has the following several aspects of research work:1. Based on synchronization code, undecimated discrete wavelet transform(UDWT) and empirical mode decomposition(EMD), we propose a robust blind digital audio watermarking technology in mixed-domain. To locate the hidden watermark accurately, synchronization code is used to resist cropping and shifting attacks. Wavelet transform is used in nonlinear and nonstationary audio signal for its well time-frequency analysis features. But, it has weak partial analysis ability and selecting different wavelet basis functions may produce different results. Considering the above factors, empirical mode decomposition is presented. It is a novel method for processing audio signals adaptively. Experimental results show that the proposed scheme is inaudible and robust against various attacks.2. A non-subsampled shearlet transform(NSST) and histogram based blind audio watermarking algorithm is proposed.To resist filtering, adding noise and equalization, NSST is introduced from image areas to audio areas successfully. To obtain the low frequency, NSST is used to this audio. In addition, in view of resisting the desynchronization attacks, the audio histogram is introduced. We make use of the embedding strategy to embed the watermarks to the low frequency subband. Experimental results show us that the scheme can resist all kinds of common attacks and desynchronization attacks. It is robust and inaudible.