An audio zero-watermark scheme based on energy comparing

摘要

Zero-watermark technique, embedding watermark without modifying carriers, has been broadly applied for copyright protection of images. However, there is little research on audio zero-watermark. This paper proposes an audio zero-watermark scheme based on energy relationship between adjacent audio sections. Taking use of discrete wavelet transformation (DWT), it gets power approximations, or energies, of audio segments. Then, it extracts the audio profile, i.e. the zero-watermark, according to the relative size of energies of consecutive fragments. The experimental results demonstrate that the proposed scheme is robust against general malicious attacks including noise addition, resampling, low-pass filtering, etc., and this approach effectively solves the contradiction between inaudibility and robustness.