首页> 外文会议>Conference on Speech Technology and Human-Computer Dialogue >Score fusion methods for text-independent speaker verification applications
【24h】

Score fusion methods for text-independent speaker verification applications

机译:关于文本独立扬声器验证应用程序的分数融合方法

获取原文

摘要

Speaker verification methods are various and use different types of features, but each system alone do not perform satisfactory results. This paper makes a comparison of different features and methods for score fusion for an independent speaker verification application. Several types of spectral features are used as speaker data. The scores obtained with these types of features were fusioned with combination methods (as: mean, sum, max, min, weighted sum) and classification methods (as: SVM, linear discriminant). The experiments were performed on a laboratory registered database for Romanian language and demonstrate that fusion methods outperformed the baseline GMM-UBM method.
机译:扬声器验证方法各种和使用不同类型的功能,但单独的每个系统都不表现出令人满意的结果。 本文对独立扬声器验证应用进行了分数融合的不同特征和方法进行了比较。 几种类型的光谱特征用作扬声器数据。 通过这些类型的特征获得的分数归于组合方法(AS:平均值,总和,最大,MIN,加权和)和分类方法(如下:SVM,线性判别)。 该实验是对罗马尼语的实验室注册数据库进行的,并证明了融合方法优于基线GMM-UBM方法。

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