-
公开(公告)号:US07890325B2
公开(公告)日:2011-02-15
申请号:US11376803
申请日:2006-03-16
申请人: Peng Liu , Ye Tian , Jian-Lai Zhou , Frank Kao-Ping K. Soong
发明人: Peng Liu , Ye Tian , Jian-Lai Zhou , Frank Kao-Ping K. Soong
CPC分类号: G10L15/08 , G10L15/02 , G10L15/187 , G10L15/193
摘要: Speech recognition such as command and control speech recognition generally use a context free grammar to constrain the decoding process. Word or subword background model are constructed to repopulate dynamic hypothesis space, especially when word spareness is at issue. The background models can be later used in speech recognition. During speech recognition, background and conventional context free grammar decoding are used to measure confidence. The discussion above is merely provided for general background information and is not intended to be used as an aid in determining the scope of the claimed subject matter.
摘要翻译: 诸如命令和控制语音识别之类的语音识别通常使用无上下文的语法来限制解码过程。 构建词或子词背景模型,以重新构建动态假设空间,特别是在词语空间问题时。 背景模型可以稍后用于语音识别。 在语音识别期间,使用背景和常规上下文无关语法解码来测量置信度。 上面的讨论仅用于一般背景信息,并不旨在用于帮助确定所要求保护的主题的范围。
-
公开(公告)号:US20080059184A1
公开(公告)日:2008-03-06
申请号:US11507859
申请日:2006-08-22
申请人: Frank Kao-Ping K. Soong , Jian-Lai Zhou , Peng Liu
发明人: Frank Kao-Ping K. Soong , Jian-Lai Zhou , Peng Liu
IPC分类号: G10L15/14
CPC分类号: G10L15/142
摘要: Measurement of Kullback-Leibler Divergence (KLD) between hidden Markov models (HMM) of acoustic units utilizes an unscented transform to approximate KLD between Gaussian mixtures. Dynamic programming equalizes the number of states between HMMs having a different number of states, while the total KLD of the HMMs is obtained by summing individual KLDs calculated by state pair by state pair comparisons.
摘要翻译: 声学单元的隐马尔可夫模型(HMM)之间的Kullback-Leibler发散(KLD)的测量利用无差异变换来近似高斯混合之间的KLD。 动态规划使具有不同数量状态的HMM之间的状态数量相等,而HMM的总KLD是通过将通过状态对比较的状态对计算的各个KLD求和来获得的。
-
公开(公告)号:US08234116B2
公开(公告)日:2012-07-31
申请号:US11507859
申请日:2006-08-22
申请人: Peng Liu , Frank Kao-Ping K. Soong , Jian-Lai Zhou
发明人: Peng Liu , Frank Kao-Ping K. Soong , Jian-Lai Zhou
CPC分类号: G10L15/142
摘要: Measurement of Kullback-Leibler Divergence (KLD) between hidden Markov models (HMM) of acoustic units utilizes an unscented transform to approximate KLD between Gaussian mixtures. Dynamic programming equalizes the number of states between HMMs having a different number of states, while the total KLD of the HMMs is obtained by summing individual KLDs calculated by state pair by state pair comparisons.
摘要翻译: 声学单元的隐马尔可夫模型(HMM)之间的Kullback-Leibler发散(KLD)的测量利用无差异变换来近似高斯混合之间的KLD。 动态规划使具有不同数量状态的HMM之间的状态数量相等,而HMM的总KLD是通过将通过状态对比较的状态对计算的各个KLD求和来获得的。
-
公开(公告)号:US07848917B2
公开(公告)日:2010-12-07
申请号:US11394809
申请日:2006-03-30
申请人: Frank Kao-Ping K. Soong , Jian-Lai Zhou , Peng Liu
发明人: Frank Kao-Ping K. Soong , Jian-Lai Zhou , Peng Liu
IPC分类号: G06F17/27
CPC分类号: G06F17/27
摘要: Multiple input modalities are selectively used by a user or process to prune a word graph. Pruning initiates rescoring in order to generate a new word graph with a revised best path.
摘要翻译: 用户或进程选择性地使用多种输入模式来修剪单词图形。 修剪开始拯救,以生成一个修改最佳路径的新字图。
-
公开(公告)号:US20080215318A1
公开(公告)日:2008-09-04
申请号:US11680827
申请日:2007-03-01
申请人: Zhengyou Zhang , Yuan Kong , Chao Huang , Frank Kao-Ping K. Soong
发明人: Zhengyou Zhang , Yuan Kong , Chao Huang , Frank Kao-Ping K. Soong
IPC分类号: G10L15/00
摘要: Recognition of events can be performed by accessing an audio signal having static and dynamic features. A value for the audio signal can be calculated by utilizing different weights for the static and dynamic features such that a frame of the audio signal can be associated with a particular event. A filter can also be used to aid in determining the event for the frame.
摘要翻译: 可以通过访问具有静态和动态特征的音频信号来执行对事件的识别。 可以通过利用静态和动态特征的不同权重来计算音频信号的值,使得音频信号的帧可以与特定事件相关联。 也可以使用滤波器来帮助确定帧的事件。
-
-
-
-