Researches On Robust Digital Audio Watermarking Algorithms

Along with the overall progress of science and the internet technology, the digitalmultimedia technique has fast developed, it becomes more and more convenient to copy,modify and transmit digital works, such as audio, image and video. The progress of thetechnology brings convenience for people, but causes the security problem of the digitalworks at the same time, such as unauthorized tamper and edit.So it is more and more importantto protect digital copyright. As an efficient tool forprotecting the intellectual property rights, digital audio watermarking techniquehasveryimportant application prospect and obtains fast development. However, there are stillsomekey issues waiting to be resolved in synchronization attack on time domain.In this paper, we pay more attention to robust audio watermarking algorithms, whichcan effective resist synchronization attack. The main contributions are as follows:(1) Introduce the background, concept, development process and status of digitalwatermarking, summarizes several characteristics of the digital watermark and the keyproblem to be resolved, then proposes the classification of digital watermarking system.(2) Introduce the digital audio watermarking model on the basis of communicationmodel, the application fields, the main attack methods, and analyze watermark qualityevaluation standard, summarize all kinds of existent audio watermarking algorithms andexisted problems, including the time domain and transform domain algorithm.(3) Present a mean quantization audio watermarking algorithm based on wavelettransformation. We first introduce the basic theory of quantization, then present the mainthought of this algorithm, that is, selects low frequency coefficients of wavelet basis usingmean quantization method. And we describe the details of the watermark embedding andextracting processes. Experiment results show that the algorithm presented in this paper isrobust to many attacks, such as MP3compression, reguantization, resampling, and randomcropping.(4) Design a robust audio watermarking algorithm based on wavelet transformation andlog coordinate transformation to solve synchronization attacks. And we analyze how to usespread spectrum technology and m-sequence to improve the security of embeddedinformation. We take log coordinates transformation and Fourier transformation for lowfrequency coefficients of wavelet, then embed watermarks to those coefficients. Finally,experimental results show that the audio watermarking algorithm we proposed couldresistnot only common signal processing attacks, but also synchronization attacks such asrandom cropping, time scale modification and pitch shifting.