专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明授权

US5946653A Speaker independent speech recognition system and method 失效
标题翻译：演讲者独立的语音识别系统和方法
公开(公告)号：US5946653A
公开(公告)日：1999-08-31
申请号：US942211
申请日：1997-10-01
申请人： William Michael Campbell , John Eric Kleider , Charles Conway Broun , Carl Steven Gifford , Khaled Assaleh
发明人： William Michael Campbell , John Eric Kleider , Charles Conway Broun , Carl Steven Gifford , Khaled Assaleh
IPC分类号： G10L9/00
CPC分类号： G10L15/10 , G10L15/02 , G10L2015/0631
摘要： An improved method of training a SISRS uses less processing and memory resources by operating on vectors instead of matrices which represent spoken commands. Memory requirements are linearly proportional to the number of spoken commands for storing each command model. A spoken command is identified from the set of spoken commands by a command recognition procedure (200). The command recognition procedure (200) includes sampling the speaker's speech, deriving cepstral coefficients and delta-cepstral coefficients, and performing a polynomial expansion on cepstral coefficients. The identified spoken command is selected using the dot product of the command model data and the average command structure representing the unidentified spoken command.
摘要翻译：训练SISRS的改进方法通过对矢量进行操作而不是代表口头命令的矩阵来使用较少的处理和存储资源。存储器要求与用于存储每个命令模型的口头命令的数量成线性比例。通过命令识别程序（200）从该组口令命令中识别口语命令。命令识别程序（200）包括对扬声器的语音进行采样，导出倒谱系数和增量倒谱系数，并对倒谱系数进行多项式扩展。使用命令模型数据的点积和代表未识别的口语命令的平均命令结构来选择所识别的口语命令。

2. 发明授权

US06192353B1 Multiresolutional classifier with training system and method 失效
标题翻译：具有训练系统和方法的多元分类器
公开(公告)号：US06192353B1
公开(公告)日：2001-02-20
申请号：US09020953
申请日：1998-02-09
申请人： Khaled Assaleh , William Michael Campbell , John Eric Kleider
发明人： Khaled Assaleh , William Michael Campbell , John Eric Kleider
IPC分类号： G06F1518
CPC分类号： G06N3/08 , G06K9/6807
摘要： An improved method and system for training and classifying using a low complexity and high accuracy multiresolutional polynomial classifier (412) is presented. A method of training an multiresolutional polynomial classifier which reduces the complexity of existing classifiers allows models representing subgroups of classes to easily be created. The models which represent subgroups of classes are applied to an unidentified input to produce a coarse classification of the unidentified input using a low order classifier. Once a coarse classification of the unidentified input is performed, a more detailed classification is performed using another low complexity classifier.
摘要翻译：提出了一种改进的使用低复杂度和高精度多分辨多项式分类器进行训练和分类的方法和系统（412）。训练多分类多项式分类器的方法可以降低现有分类器的复杂度，可以容易地创建表示类的子组的模型。表示子类别的模型被应用于未识别的输入以使用低阶分类器来产生未识别输入的粗略分类。一旦执行了未识别输入的粗略分类，则使用另一个低复杂度分类器来执行更详细的分类。

3. 发明授权

US5930748A Speaker identification system and method 失效
标题翻译：扬声器识别系统和方法
公开(公告)号：US5930748A
公开(公告)日：1999-07-27
申请号：US893755
申请日：1997-07-11
申请人： John Eric Kleider , Khaled Assaleh
发明人： John Eric Kleider , Khaled Assaleh
IPC分类号： G10L15/06 , G10L17/00 , G10L3/02
CPC分类号： G10L15/075 , G10L17/04
摘要： A speaker identification system (10) employs a supervised training process (100) that uses row action projection (RAP) to generate speaker model data for a set of speakers. The training process employing RAP uses less memory and processing resources by operating on a single row of a matrix at a time. Memory requirements are linearly proportional to number of speakers for storing each speakers information. A speaker is identified from the set of speakers by sampling the speaker's speech (202), deriving cepstral coefficients (208), and performing a polynomial expansion (212) on cepstral coefficients. The identified speaker (228) is selected using the product of the speaker model data (213) and the polynomial expanded coefficients from the speech sample.
摘要翻译：扬声器识别系统（10）采用使用行动投影（RAP）来产生一组扬声器的扬声器模型数据的监督训练过程（100）。使用RAP的培训过程一次通过在矩阵的单行上运行来减少存储器和处理资源。存储器要求与用于存储每个扬声器信息的扬声器数成正比。通过对扬声器的语音（202）进行采样，导出倒谱系数（208）和对倒谱系数执行多项式展开（212），从该组扬声器识别扬声器。使用扬声器模型数据（213）和来自语音样本的多项式展开系数的乘积来选择所识别的扬声器（228）。

4. 发明授权

US06243695B1 Access control system and method therefor 失效
标题翻译：具有防欺诈功能的电信设备
公开(公告)号：US06243695B1
公开(公告)日：2001-06-05
申请号：US09045361
申请日：1998-03-18
申请人： Khaled Assaleh , William Michael Campbell
发明人： Khaled Assaleh , William Michael Campbell
IPC分类号： G06F1518
CPC分类号： G06K9/68
摘要： A TCS (200) and procedure (400) for identifying an unidentified class as a class of a group of classes includes a new tree-structured classifier (208) and training processor (204). Unidentified feature vectors representing an unidentified class are combined with predetermined models to compute a score for each of the unidentified feature vectors. Based on the scores for each of the unidentified feature vectors, an association is made with the predetermined models to identify the unidentified class. Predetermined models are created using a training procedure (300) for predetermined feature vectors associated therewith. A procedure (400) for identifying an unidentified class as a class of a group of classes is useful when determining access privileges to a device or system.
摘要翻译：用于将未识别的类识别为类的一类的TCS（200）和过程（400）包括新的树结构分类器（208）和训练处理器（204）。表示未识别类别的不明特征向量与预定模型组合以计算每个未识别特征向量的得分。基于每个未识别特征向量的分数，使用预定模型进行关联以识别未识别的类。使用用于与其相关联的预定特征向量的训练过程（300）创建预定模型。当确定对设备或系统的访问权限时，用于将未识别的类识别为一组类的类的过程（400）是有用的。

5. 发明授权

US08175876B2 System and method for an endpoint detection of speech for improved speech recognition in noisy environments 失效
标题翻译：用于在嘈杂环境中改善语音识别的语音终端检测的系统和方法
公开(公告)号：US08175876B2
公开(公告)日：2012-05-08
申请号：US12459168
申请日：2009-06-25
申请人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
发明人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
IPC分类号： G10L17/00
CPC分类号： G10L25/87
摘要： According to a disclosed embodiment, an endpointer determines the background energy of a first portion of a speech signal, and a cepstral computing module extracts one or more features of the first portion. The endpointer calculates an average distance of the first portion based on the features. Subsequently, an energy computing module measures the energy of a second portion of the speech signal, and the cepstral computing module extracts one or more features of the second portion. Based on the features of the second portion, the endpointer calculates a distance of the second portion. Thereafter, the endpointer contrasts the energy of the second portion with the background energy of the first portion, and compares the distance of the second portion with the distance of the first portion. The second portion of the speech signal is classified by the endpointer as speech or non-speech based on the contrast and the comparison.
摘要翻译：根据所公开的实施例，终端指针确定语音信号的第一部分的背景能量，倒谱计算模块提取第一部分的一个或多个特征。 endpointer基于特征计算第一部分的平均距离。随后，能量计算模块测量语音信号的第二部分的能量，并且倒谱计算模块提取第二部分的一个或多个特征。基于第二部分的特征，终点计算器计算第二部分的距离。此后，endpointer将第二部分的能量与第一部分的背景能量进行对比，并将第二部分的距离与第一部分的距离进行比较。基于对比度和比较，语音信号的第二部分被endpointer分类为语音或非语音。

6. 发明授权

US07474915B2 Separating mixed signals containing a distorted signal 有权
标题翻译：分离包含失真信号的混合信号
公开(公告)号：US07474915B2
公开(公告)日：2009-01-06
申请号：US11189677
申请日：2005-07-26
申请人： Khaled Assaleh , Hasan Al-Nashash
发明人： Khaled Assaleh , Hasan Al-Nashash
IPC分类号： A61B5/04
CPC分类号： A61B5/0444 , A61B5/4362 , A61B5/726
摘要： This invention provides a novel technique separating two mixed signals. The method is particularly useful for fetal electrocardiogram estimation. A maternal electrocardiogram signal recorded at the thorax area is nonlinearly mapped onto an electrocardiogram signal recorded at the abdomen using polynomial networks. The fetal electrocardiogram component is then estimated by subtracting the nonlinearly mapped maternal electrocardiogram signal from the abdominal electrocardiogram signal.
摘要翻译：本发明提供一种分离两个混合信号的新技术。该方法对胎儿心电图估计特别有用。记录在胸部区域的母体心电图信号被非线性映射到使用多项式网络记录在腹部的心电图信号。然后通过从腹部心电图信号中减去非线性映射的母亲心电图信号来估计胎儿心电图分量。

7. 发明申请

US20120191455A1 System and Method for an Endpoint Detection of Speech for Improved Speech Recognition in Noisy Environments 审中-公开
标题翻译：用于在嘈杂环境中改进语音识别的语音端点检测的系统和方法
公开(公告)号：US20120191455A1
公开(公告)日：2012-07-26
申请号：US13438715
申请日：2012-04-03
申请人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
发明人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
IPC分类号： G10L15/00
CPC分类号： G10L25/87
摘要： According to a disclosed embodiment, an endpointer determines the background energy of a first portion of a speech signal, and a cepstral computing module extracts one or more features of the first portion. The endpointer calculates an average distance of the first portion based on the features. Subsequently, an energy computing module measures the energy of a second portion of the speech signal, and the cepstral computing module extracts one or more features of the second portion. Based on the features of the second portion, the endpointer calculates a distance of the second portion. Thereafter, the endpointer contrasts the energy of the second portion with the background energy of the first portion, and compares the distance of the second portion with the distance of the first portion. The second portion of the speech signal is classified by the endpointer as speech or non-speech based on the contrast and the comparison.
摘要翻译：根据所公开的实施例，终端指针确定语音信号的第一部分的背景能量，倒谱计算模块提取第一部分的一个或多个特征。 endpointer基于特征计算第一部分的平均距离。随后，能量计算模块测量语音信号的第二部分的能量，并且倒谱计算模块提取第二部分的一个或多个特征。基于第二部分的特征，终点计算器计算第二部分的距离。此后，endpointer将第二部分的能量与第一部分的背景能量进行对比，并将第二部分的距离与第一部分的距离进行比较。基于对比度和比较，语音信号的第二部分被endpointer分类为语音或非语音。

8. 发明授权

US06691089B1 User configurable levels of security for a speaker verification system 有权
标题翻译：扬声器验证系统的用户可配置的安全级别
公开(公告)号：US06691089B1
公开(公告)日：2004-02-10
申请号：US09409942
申请日：1999-09-30
申请人： Huan-yu Su , Khaled Assaleh
发明人： Huan-yu Su , Khaled Assaleh
IPC分类号： G10L1506
CPC分类号： G06F21/32 , G06F2221/2113 , G10L17/24
摘要： A text-prompted speaker verification system that can be configured by users based on a desired level of security. A user is prompted for a multiple-digit (or multiple-word) password. The number of digits or words used for each password is defined by the system in accordance with a user set preferred level of security. The level of training required by the system is defined by the user in accordance with a preferred level of security. The set of words used to generate passwords can also be user configurable based upon the desired level of security. The level of security associated with the frequency of false accept errors verses false reject errors is user configurable for each particular application.
摘要翻译：文本提示的说话人验证系统，可以由用户根据所需的安全级别进行配置。系统提示用户输入多位数字（或多字）密码。用于每个密码的数字或字数由系统根据用户设置的优选安全级别来定义。系统所需的训练水平由用户根据优选的安全级别来定义。用于生成密码的一组单词也可以根据所需的安全级别进行用户可配置。与错误接收错误频率相关联的安全级别与错误拒绝错误的级别是用户可针对每个特定应用程序配置的。

9. 发明授权

US06556969B1 Low complexity speaker verification using simplified hidden markov models with universal cohort models and automatic score thresholding 有权
标题翻译：使用具有通用队列模型的简化隐马尔可夫模型和自动分数阈值来进行低复杂度的扬声器验证
公开(公告)号：US06556969B1
公开(公告)日：2003-04-29
申请号：US09408453
申请日：1999-09-30
申请人： Khaled Assaleh , Ayman Asadi
发明人： Khaled Assaleh , Ayman Asadi
IPC分类号： G10L1506
CPC分类号： G10L17/00 , G10L17/04 , G10L25/24
摘要： A low complexity speaker verification system that employs universal cohort models an automatic score thresholding. The universal cohort models are generated using a simplified cohort model generating scheme. In certain embodiments of the invention, a simplified hidden Markov modeling (HMM) scheme is used to generate the cohort models. In addition, the low complexity speaker verification system is trained by various users of the low complexity speaker verification system. The total number of users of the low complexity speaker verification system may be modified over time as required by the specific application, and the universal cohort models may be updated accordingly to accommodate the new users. The present invention employs a combination of universal cohort modeling and thresholding to ensure high performance. In addition, given the simplified generation of the cohort models and training of the low complexity speaker verification system, substantially reduced processing resources and memory are amenable for high performance of the low complexity speaker verification system. In certain embodiments of the invention, the invention is an integrated speaker training and speaker verification system that performs both training and speaker verification.
摘要翻译：一种低复杂度的扬声器验证系统，其采用通用队列模型自动分数阈值。使用简化队列模型生成方案生成通用队列模型。在本发明的某些实施例中，使用简化的隐马尔可夫模型（HMM）方案来生成队列模型。此外，低复杂性扬声器验证系统由低复杂度扬声器验证系统的各种用户训练。低复杂度扬声器验证系统的用户总数可以根据具体应用的要求随时间进行修改，并且可以相应地更新通用队列模型以适应新用户。本发明采用通用队列建模和阈值化的组合来确保高性能。另外，考虑到队列模型的简化生成和低复杂度的扬声器验证系统的训练，大大减少的处理资源和存储器适用于低复杂性扬声器验证系统的高性能。在本发明的某些实施例中，本发明是一个执行训练和说话者验证的综合讲话者训练和说话者验证系统。

10. 发明申请

US20100030559A1 System and method for an endpoint detection of speech for improved speech recognition in noisy environments 失效
标题翻译：用于在嘈杂环境中改善语音识别的语音终端检测的系统和方法
公开(公告)号：US20100030559A1
公开(公告)日：2010-02-04
申请号：US12459168
申请日：2009-06-25
申请人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
发明人： Sahar E. Bou-Ghazale , Ayman O. Asadi , Khaled Assaleh
IPC分类号： G10L17/00 , G10L15/00
CPC分类号： G10L25/87
摘要： According to a disclosed embodiment, an endpointer determines the background energy of a first portion of a speech signal, and a cepstral computing module extracts one or more features of the first portion. The endpointer calculates an average distance of the first portion based on the features. Subsequently, an energy computing module measures the energy of a second portion of the speech signal, and the cepstral computing module extracts one or more features of the second portion. Based on the features of the second portion, the endpointer calculates a distance of the second portion. Thereafter, the endpointer contrasts the energy of the second portion with the background energy of the first portion, and compares the distance of the second portion with the distance of the first portion. The second portion of the speech signal is classified by the endpointer as speech or non-speech based on the contrast and the comparison.
摘要翻译：根据所公开的实施例，终端指针确定语音信号的第一部分的背景能量，倒谱计算模块提取第一部分的一个或多个特征。 endpointer基于特征计算第一部分的平均距离。随后，能量计算模块测量语音信号的第二部分的能量，并且倒谱计算模块提取第二部分的一个或多个特征。基于第二部分的特征，终点计算器计算第二部分的距离。此后，endpointer将第二部分的能量与第一部分的背景能量进行对比，并将第二部分的距离与第一部分的距离进行比较。基于对比度和比较，语音信号的第二部分被endpointer分类为语音或非语音。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式