• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>2015 Global Conference on Communication Technologies >Effects of fuzzy parameter on text dependent speaker verification under uncontrolled noisy environment
【24h】

Effects of fuzzy parameter on text dependent speaker verification under uncontrolled noisy environment

机译：嘈杂环境下模糊参数对文本相关说话人验证的影响

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

Statistical pattern recognition has been considered to be one of the most successful approaches in the recent advancement of speech and speaker recognition. Out of all the approaches Hidden Markov Models, Gaussian mixture models and Vector Quantization has been considered to be one of the most successful techniques in regards to the performance of the speaker recognition systems. However the performance of these techniques degrades when subjected to limited data condition and noisy environment. Fuzzy approaches with their variable fuzzy parameters may reduce the degradation. This paper attempts to highlight the effect of learning parameter of objective function while implementing Fuzzy Vector Quantization on Text Dependent Speaker Verification under limited data condition and also under practical noisy environment. The entire set of experiments were performed between learning parameter m=1.1 to m=2 and the system accuracy was observed in each case. The experimental results performed on telephonic database suggests better results for learning parameter m=1.37 where the maximum accuracy of the system reaches 84.81%. However the performance of the system also depends on codebook size. Our research focuses the effectiveness of the variation of learning parameter in speaker verification performance and robustness of the system.
机译：在语音和说话者识别的最新发展中，统计模式识别被认为是最成功的方法之一。在所有方法中，隐马尔可夫模型，高斯混合模型和矢量量化被认为是关于说话人识别系统性能的最成功技术之一。但是，在受限的数据条件和嘈杂的环境下，这些技术的性能会下降。具有可变模糊参数的模糊方法可以减少降级。本文试图强调目标函数学习参数在有限数据条件下以及实际嘈杂环境下实现模糊矢量量化对文本相关说话人验证的影响。在学习参数m = 1.1到m = 2之间进行了整套实验，并观察了每种情况下的系统精度。在电话数据库上执行的实验结果表明，学习参数m = 1.37可获得更好的结果，其中系统的最大精度达到84.81％。但是，系统的性能还取决于密码本的大小。我们的研究集中于学习参数变化对说话人验证性能和系统健壮性的有效性。

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