关键词: 小波包分析, 双水印, 版权保护, 内容认证

Abstract: Aiming at solving the problems of previous dual watermarking algorithm, such as poor robustness and inaccurate positioning, a wavelet packet analysis-based digital audio dual watermarking scheme was presented. In this scheme, zero-watermarking and semi-fragile watermarking were embedded into the low-frequency coefficients and the intermediate low-frequency coefficients of wavelet packet decomposition for the original audio signal. The robustness of improved zero-watermarking algorithm has been greatly increased; therefore, the copyright protection of audio signals can be completed better; binary watermark image was embedded in the intermediate low-frequency coefficients by quantization different from the previous algorithm requiring to convert two-dimensional image to one-dimensional sequence, which did not affect zero-watermarking extraction. The algorithm has good robustness in resisting a series of conventional treatment and amplitude scaling, can not only examine the malicious tampering location of the watermark image, but also locate the tampering position of original vector audio, so it achieves the content authentication of the audio signal truly.

Key words: wavelet packet analysis, dual watermark, copyright protection, content authentication