Research on digital audio zero watermarking algorithm based on transform domain
|School||Jiangxi University of Technology|
|Course||Communication and Information System|
|Keywords||digital audio watermarking discrete wavelet transform discrete cosine transform zero-watermarking adaptive|
With the rapid development of information technology and audio compression technology,various forms of multimedia works can be created, stored and transferred through a variety ofcommunication networks easily. However, digital media is also facing with severe securityproblems. Digital audio watermarking technology solved this problem well, it can protectcopyright and authenticate content of digital works, and thus it has become a hot topic in thecurrent academic research ifeld.Analysis based on the discrete cosine transform, discrete wavelet transform, the short-termenergy, zero crossing rate and the mean, variance, combined with the human auditory system(HAS) and singular value decomposition (SVD)，two transform domain-based digital audiowatermarking algorithms were designed in this paper. The main work of this paper was asfollows:(1)Introduced application background, current research status, basic models and algorithmsclassiifcation of audio watermarking technology, and the above knowledge could provide atheoretical basis of the design work of algorithm in this paper.(2)The local features of audio signals in time domain can against attacks strongly, as well asthe stability features of the singular value, a digital audio zero-watermarking algorithm in thewavelet domain is presented. The algorithm uses ifrstly chaotic sequence to encrypt thewatermarking image, then chooses the most suitable for the audio segment constructedzero-watermarking according to the two local features based on the short-term energy andzero-crossing rate of audio signals in time domain. The selected audio segments are carried ondiscrete wavelet transform, extracts low frequency components of the wavelet domain forsingular value decomposition. The zero-watermarking is constructed judging the parity of largestsingular value of the low frequency coeiffcients. The experimental results show that the algorithmhas high security and good imperceptibility, and it can against attacks such as noise interference,low pass ifltering, quantization, resampling and compression and so on.(3)The mean of the audio signal is very stability under the conventional signal processing.Based on this characteristic, an adaptive double audio watermarking algorithm in the hybridtransform domain is presented. The algorithm carried on DWT and DCT for each segment oforiginal audio signal, analyzed the mean and variance of the Low frequency coefifcients andmiddle-low frequency, zero-watermarking was structured and semi-fragile watermarking wasembedded adaptively. The experimental results show that the imperceptibility of the algorithm isvery good,it has good robustness against many kinds of attacks, and it is able to locate themalicious tampering.