Content-based audio zero-watermarking algorithm against TSM

With the development of the Internet, authentication and copyright protection of digital products becomes an important issue. In this paper, a zero-watermarking algorithm based on audio content is proposed, which can stand against synchronization attacks. The audio signal is divided into several segments according to the size of watermarking image. After that, SVD transform is performed for each matrix. And then all those largest singular values of each matrix are classified and encoded based on K-means algorithm. The final zero-watermarking sequences can be obtained by processing these codes and watermarking image. The experiment results show that the proposed algorithm has great robustness against common attacks. Especially, in the case of many synchronization attacks such as TSM and jittering, it also has outstanding performance.