Digital Watermarking Technology For Protecting The Copyright Of Audio Aggregation And Authenticating The Integrity Of Audio Aggregation

Recently, the exchange of multimedia data, such as image, audio, video and so on, becomesmore frequently on the Internet. It is easy to obtain these information from the network, but someproblems such as the security of information have become increasingly severe. Currently, digitalwatermarking technology becomes a research focus in information security field because it is aneffective means to solve the problem of information security. Digital watermarking technologyoften takes image, audio, video and text as embedding carrier. All along, digital audiowatermarking technology takes single audio work as embedding carrier. That is to say, currently,the applications such as copyright protection, content integrity authentication are all for singleaudio work. However, we find that the audio works often are published in the form of artist albumsor CD. Therefore, the audio watermarking technology based on audio aggregation needs furtherstudy. In this paper, from the perspective of audio aggregation, we study on the audio aggregationwatermarking technology, the main content is shown as follows.An algorithm for audio aggregation based on singular value decomposition (SVD) which canrealize copyright protection has been presented. As an audio aggregation contains several differentaudio works, which means the characteristics value of each audio work is different. Combined withthe stability of the matrix’s singular value, zero-watermark sequence is generated for each audio byusing the audio work’s singular values. Finally, a random matrix is generated to mix the zero-watermark sequences, which can improve the security of the algorithm greatly.Two content integrity authentication schemes for audio aggregation based on key sharingscheme have been proposed. The first scheme is an audio aggregation zero-watermark algorithmbased on (k, n/k=n).Binary watermark image is distributed into n shares by (k, n) threshold scheme.Then each share is embedded in the n audio works in the aggregation. In the watermark distributedprocess, in order to obtain better vulnerability, k is set as n. To embed watermark by generating keywithout changing audio work’s any features, combined with zero-watermarking technology, usingeach audio work’s nature to encrypt the key matrix. Therefore, we needn’t considering the order ofthe audio works in the audio aggregation when extracting the watermark. The second scheme is acontent integrity authentication scheme for audio aggregation algorithm based on vector-sharingscheme. Firstly, watermark distribution. We analyze the principle of vector-sharing scheme deeply,then n shares are generated from the visible binary watermark by using vector-sharing scheme. n is the number of audio works in the audio aggregation. Secondly, we should select the position forembedding watermark. The experimental results show that the lower the energy of wavelet packetnodes, the more sensitive to various common signal-processing operations. We selected the nodecoefficient of the lowest energy from corresponding audio works to embed. Finally, usingquantization index modulation (QIM) to complete the embedding process and the watermark canbe extracted without the original audio aggregation.A multipurpose audio aggregation watermarking algorithm based on multistage vectorquantization has been proposed. Digital watermarks for different purposes have been embedded indifferent stages of multistage vector quantization. Firstly, robust watermark is embedded in the firstquantization stage using a novel scheme. In the scheme, the input vectors are divided into twoclusters according to the parity of codeword index. Information ‘0’ and ‘1’ are assigned to the twoclusters respectively. In the second quantization stage, the output indexes are put into Huffmancoding after obtaining their probability distribution in the index table. Then the classical method ofcryptography is applied to complete zero-watermark embedding. The experimental results showthe robust watermarking with strong anti-attack capability and the fragile watermarking with strongfragility can be used for protecting audio aggregation’s copyright and authenticating the integrity ofaudio aggregation respectively.