首页> 外文会议>International Workshop on Acoustic Echo and Noise Control >(Outrageously) simple statistical models for blind source separation (BSS) or The unreasonable effectiveness of statistical Independence - (PPT)
【24h】

(Outrageously) simple statistical models for blind source separation (BSS) or The unreasonable effectiveness of statistical Independence - (PPT)

机译:(乌龟)盲源分离(BSS)的简单统计模型或统计独立性的不合理效果 - (PPT)

获取原文

摘要

A natural extension is to model the signal in the time-frequency plane or in the wavelet domain. The implicit model then is that each source has a different pattern distribution of energy in time-frequency domains or in time-scale domains. The size of the time-frequency tt time-scale domains should be chosen to have a reasonable number of Gabor atoms in each domain so that the energy can be estimated in each of them. Note: it is also possible to use the likelihood to select the partitioning of the time-frequency plane which gives the best contrast between the energy distribution of the sources (adaptive partitioning).
机译:自然扩展是在时频平面或小波域中模拟信号。然后,隐式模型是每个源在时频域中或在时间尺度域中具有不同的能量模式分布。应选择时频TT时间尺度域的大小在每个结构域中具有合理数量的Gabor原子,使得能量可以在每个结构中估计。注意:还可以使用可能性来选择时频平面的分区,这在源的能量分布(自适应分区)之间提供了最佳对比度。

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