...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Cybernetics and information technologies: CIT >A Speaker Recognition Method Based on Personal IdentificationVoice and Trapezoidal Fuzzy Similarity
【24h】

A Speaker Recognition Method Based on Personal IdentificationVoice and Trapezoidal Fuzzy Similarity

机译：基于个人识别语音和梯形模糊相似度的说话人识别方法

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

A text-dependent speaker recognition method is proposed usingtrapezoidal fuzzy similarity function to measure the similarity of voice featuresbetween a test user and the registered speaker who has nearest distance. Thetrapezoidal fuzzy similarity function is constructed based on three-time datarecorded during enrolment process as personal identification voice (PIV) andstatistical data of an individual recorded many times in a long time period to coverthe intra-variation. A set of acoustic voice features is also introduced to presentsome general speaker and text dependent characteristics that are effective formodeling PIV, thus allowing to capture the inter- variation from one speaker toanother. The experimental results on 24 speakers recorded in four different sessionsshow that, without false acceptation, the proposed system can decrease 30.05% offalse rejection cases, compared to the traditional nearest neighbor approach. Thefocus of this work is on applications which require fast processing and few burdensfor users.
机译：提出了一种基于文本的说话人识别方法，该方法采用梯形模糊相似度函数来测量测试用户与距离最近的注册说话人之间语音特征的相似度。梯形模糊相似度函数是基于在注册过程中记录的三次数据（如个人身份语音（PIV））和长时间记录的个人统计数据而构建的，以覆盖内部变异。还引入了一组声学语音功能来呈现一些有效的通用说话者和与文本相关的特性，这些特性对于PIV建模非常有效，因此可以捕获一个说话者与另一个说话者之间的相互变化。在四个不同的会话中记录的24位发言人的实验结果表明，与传统的最近邻居方法相比，所提出的系统在没有错误接受的情况下，可以减少30.05％的错误拒绝案例。这项工作的重点是需要快速处理且用户负担很少的应用程序。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号