Design And Implementation Of Digital Audio Watermarking Algorithm Based On Chaotic Encryption

With the rapid development of today’s network technology and the increasing maturity of multimedia technology, a lot of digital audio and video products are transmitted in the open network represented by the Internet, which makes the piracy problem of the digital products increasingly obvious. Therefore, how to effectively protect the copyright of digital products under the network environment is becoming more and more urgent, and digital watermarking technology comes into being in such a background. This technology protects the security of data through hiding some secret information in digital products. Meanwhile, the research on audio watermarking technology is widely investigated both in the domestic and overseas but there are some drawbacks in the encryption. Nowadays, chaotic systems have become the research hotspot in the modern cryptography because of its huge advantages such as unique inherent randomness, sensitivity to initial conditions and unpredictability and so on. On the basis of introducing chaos theory into this thesis, we carry out the new research and exploration about the watermark embedding and detection problems.In this thesis, we choose the binary image as digital watermark information and propose a watermark processing method based on the combination of wavelet domain transformation and DCT domain transformation so as to improve the robustness and transparency of the embedded audio. Compared with the traditional single domain algorithms, the algorithm of embedding the watermark information into the DC component of DCT transformation which is the approximation of wavelet transformation has better robustness. Meanwhile, considering the unbalance between the transparency and robustness in the traditional hybrid domain algorithms, we propose and implement a zero watermark processing method based on the combination of transform domain. In this method, we utilize the DC coefficient’s characteristic that its polarity is not easy to be changed and then construct a "zero watermark" based on the coefficient’s sign, which enables that the algorithm has higher transparency and stronger robustness and has a reasonable balance between them.Finally, we assess the security, transparency and robustness of these two algorithms, respectively. We do the attack tests on the audio after being embedded with watermark, including adding Gaussian noise, resample, low pass filtering, noise reduction and loss compression etc. The results show that these two algorithms have better robustness and transparency compared with traditional methods and can achieve the blind extraction, which is much more applicable and can effectively ensure the security of digital audio under the network environment.