Efficiently Self-synchronized Audio Watermarking against Re-sampling Attack

This paper presents a novel self-synchronized audio watermarking algorithm especially robust to re-sampling attack based on the frequency centroid. Thanks to the steady property of the audio centroid, each sub-band which carries the centroid of one audio frame is performed with DCT, and the scrambled watermark bits are embedded into DC component. The experimental results show that the proposed audio watermarking scheme is not only inaudible, but also robust against various common signal processing, especially the re-sampling attack. Compared with the existing feature-based watermarking schemes, the proposed scheme can be applicable to the various types of audio signals. Furthermore, the synchronization is implemented without using additional synchronization information.