基于MIDI音乐的数字音频半脆弱性水印算法的研究

提出了一种基于MIDI音乐的音频半脆弱性水印算法，将待嵌入水印转化为MIDI音乐，并对水印进行了主成份分析，可在具有较小冗余的载体中进行嵌入；尔后，用混沌加密及重复编码，来提高水印的鲁棒性；同时采用三种检测方法，提高对水印提取的判断正确率。实验结果表明，算法对于一般性的攻击（剪切、回声等）具有较好的鲁棒性，可实现较好的认证功能。

A Digital Audio Semi- Fragile Watermarking Algorithm Based on MIDI Music

Proposed A digital audio Semi -fragile watermarking algorithm is proposed based on MIDI mu- sic. The watermark is convered into MIDI music, and the principal component is analyzed, and then it is em- bedded in the carrier which have less redundancy; Later, the chaotic encryption and repeated encoding are used to improve it＇s robustness ; simutaneously, three test methods are used to improve the perception accuracy of the watermark detection. Experimental results show that the algorithm have robustness to the general attacks （cutting, echoing, etc. ）, and better authentication can be realized.