【摘要】 说话人识别在自动身份鉴别方面具有重要的现实意义。文章在分析了说话人识别实现的可行性基础上，提出采用长时线性预测倒谱（ＬＰＣＣＥＰ）系数和矢量量化（ＶＱ）模型的不依赖于文本的说话人识别方法，并成功地应用于一个数据库查询系统中。该方法训练时间短，识别响应实时，数据存储量少。实验表明，在１００名说话人集合内，识别率达９８％。更多还原

【Abstract】 Speaker identification has significant applications in automatic individual recognition area.In this paper, the possibility of accomplishment of speaker identification is analyzed,and then a method for text-independent speaker identification is presented,which is based on vector quantitaion (VQ) and long-time LPCCEP parameters. This method is used in a multimedia data inquiring system,and it has some advantages:such as training time shorter,response real time and data memory smaller.The tested results show that the correct recognition rate in 100 persons set is up to 98%.更多还原