基于盲源分离的多重音频数据库水印算法

摘要

By introducing speech, a multiple audio watermarking algorithm for database based on blind source separation is presented.The watermarking,a mixture of the speech signals of multiple copyright owners is inserted into database. After multiple watermarking is extracted, the speech signals of copyright owners will be separated by blind source separation algorithm based on maximum signal noise ratio. The speeches are matched by speech recognition technology based on DWT algorithm in order to prove that speeches are owned by multiple copyright owners.%将语音特征引入数据库水印,提出一种基于盲源分离的多重音频数据库水印算法.把语音信号混合成水印信息后嵌入到关系数据库中.多重水印提取后,采用基于最大信噪比的盲源分离算法分离语音信息,应用动态时间规整算法的语音识别技术,将语音信息进行模板匹配确定版权归属.