基于恒定统计特征的数字音频水印技术

Technique of Audio Digital Watermarking Based on Statistical Invariableness

在音频水印技术中,满足听觉透明性和低计算复杂度要求的同时,要保证系统的高鲁棒性是音频水印技术中的难点。针对这个难点,本文使用带有边信息的通信模型对水印系统进行描述,通过对该模型中的统计域参数进行修改来嵌入水印,并实现了水印的盲提取。试验表明,这种方法在听觉透明性和鲁棒性方面有令人满意的效果。

To insure the system＇s high robustness is the difficulty of the audio watermark technology, with satisfying the hearing transparency and low computation complexity. This paper presents a communication model used to describe the watermark system, which imbeds the watermark through modifying the statistical parameters, and blind extraction is carried out. Experimental results prove perfect performance of this approach in hearing transparent and the robustness.