专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

71. 发明公开

EP0727768A1 Method of and apparatus for reducing noise in speech signal 有权
标题翻译： Verfahren und Vorrichtung zur Verminderung von Rauschen bei Sprachsignalen
公开(公告)号：EP0727768A1
公开(公告)日：1996-08-21
申请号：EP96301058.2
申请日：1996-02-16
申请人： SONY CORPORATION
发明人： Chan, Joseph , Nishiguchi, Masayuki
IPC分类号： G10L5/06 , G10L9/10 , G10L7/08 , G10L9/06
CPC分类号： G10L21/0208 , G10L25/09 , G10L25/93
摘要： A method and an apparatus for reducing the noise in a speech signal capable of suppressing the noise in the input signal and simplifying the processing. The apparatus includes a fast Fourier transform unit 3 for transforming the input speech signal into a frequency-domain signal, and an Hn value calculation unit 7 for controlling filter characteristics for filtering employed for removing the noise from the input speech signal. The apparatus also includes a spectrum correction unit 10 for reducing the input speech signal by the filtering conforming to the filter characteristics produced by the Hn value calculation unit 7. The Hn value calculation unit 7 calculates the Hn value responsive to a value derived from the frame-based maximum SN ratio of the input signal spectrum obtained by the fast Fourier transform unit 3 and an estimated noise level and controls the processing for removing the noise in the spectrum correction unit 10 responsive to the Hn value.
摘要翻译：一种用于降低能够抑制输入信号中的噪声并简化处理的语音信号中的噪声的方法和装置。该装置包括用于将输入语音信号变换为频域信号的快速傅里叶变换单元3和用于控制用于从输入语音信号中去除噪声的滤波滤波器特性的Hn值计算单元7。该装置还包括频谱校正单元10，用于通过符合由Hn值计算单元7产生的滤波特性的滤波来减少输入语音信号.Hn值计算单元7响应于从帧导出的值计算Hn值由快速傅立叶变换单元3获得的输入信号频谱的最大SN比和估计噪声电平，并根据Hn值控制频谱校正单元10中的噪声去除处理。

72. 发明公开

EP0726561A2 Voice-recognition device 失效
标题翻译：语音识别设备
公开(公告)号：EP0726561A2
公开(公告)日：1996-08-14
申请号：EP96300055.9
申请日：1996-01-03
申请人： Toyota Jidosha Kabushiki Kaisha
发明人： Aoshima, Shigeki
IPC分类号： G10L7/08 , G10L5/06
CPC分类号： G10L15/02 , G10L15/10 , G10L15/12 , G10L25/24
摘要： A sound processor (12) calculates first through third parameters according to an LPC cepstrum, a primary delta cepstrum and a secondary delta cepstrum. The first parameter catches a static characteristic, the second parameter catches a dynamic characteristic with time, and the third parameter catches a locally dynamic characteristic with time. A word dictionary (14) stores first through third parameters for a standard pattern. Hence, a DP matching unit (16) recognizes a voice based on a distance between an input voice of the three parameters and the standard pattern.
摘要翻译：声音处理器（12）根据LPC倒频谱，初级三角倒频谱和次级三角倒频谱计算第一至第三参数。第一个参数捕获一个静态特性，第二个参数随时间捕获一个动态特性，第三个参数捕获一个随时间变化的局部动态特性。单词词典（14）存储标准模式的第一至第三参数。因此，DP匹配单元（16）基于三个参数的输入语音与标准模式之间的距离来识别语音。

73. 发明公开

EP0724251A1 Speech adaptation device 失效
标题翻译： Sprachanpassungsgerät
公开(公告)号：EP0724251A1
公开(公告)日：1996-07-31
申请号：EP96101048.5
申请日：1996-01-25
申请人： NEC CORPORATION
发明人： Takagi, Keizaburo, c/o NEC Corp.
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06 , G10L9/18
CPC分类号： G10L15/065 , G10L15/02 , G10L2015/088
摘要： A speech adaptation device comprises a vocabulary independent reference pattern memory for memorizing a plurality of vocabulary independent reference patterns having one or more categories. Each category has one or more acoustic units, and has such a connection relation of the acoustic units that allows reception of any sequence of the acoustic units appearing in the input speech. A preliminary matching unit is for use in making time-alignment between the time series of the feature vectors of the input speech obtained from the analysis unit and the vocabulary independent reference pattern to obtain mean vectors for individual categories of the input speech and the vocabulary independent reference pattern from the aligned portion for the individual categories of the feature vectors of the input speech and the vocabulary independent reference pattern. An adaptation unit is for use in making correction of at least one of the time series of the feature vectors of the input speech and the vocabulary independent reference pattern by using the mean vectors for each category calculated by the preliminary matching unit.
摘要翻译：语音适配装置包括用于存储具有一个或多个类别的多个词汇独立参考模式的词汇独立参考模式存储器。每个类别具有一个或多个声学单元，并且具有允许接收出现在输入语音中的声学单元的任何序列的声学单元的这种连接关系。初步匹配单元用于在从分析单元获得的输入语音的特征向量的时间序列与词汇独立参考模式之间进行时间对准，以获得输入语音的各个类别和词汇的独立的平均向量来自对齐部分的参考模式，用于输入语音的特征向量的各个类别和词汇独立参考模式。适应单元用于通过使用由初步匹配单元计算的每个类别的平均向量来校正输入语音的特征向量的时间序列和词汇独立参考模式中的至少一个。

74. 发明公开

EP0677835A3 Verfahren zum Ermitteln einer Folge von Wörtern 失效
标题翻译：一种用于确定单词的一个序列的方法。
公开(公告)号：EP0677835A3
公开(公告)日：1996-04-17
申请号：EP95200872.0
申请日：1995-04-06
申请人： Philips Patentverwaltung GmbH , Philips Electronics N.V.
发明人： Aubert, Louis Xavier, Dr. , Ney, Hermann, Dr.
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06
CPC分类号： G10L15/08 , G10L15/12 , G10L15/193 , G10L2015/085
摘要： Bei der Erkennung zusammenhängend gesprochener Sprache werden mit Hilfe der dynamischen Programmierung viele Hypothesen im Suchraum erzeugt. Wenn in einem Wort verschiedene Hypothesen zu verschiedenen Vorgängerwörtern gestartet werden und zum selben Endpunkt laufen, werden an diesem Endpunkt die Daten der Hypothesen getrennt als Wortergebnisse gespeichert. Aus diesen Wortergebnissen wird nun erfindungsgemäß ein Wortgitter gebildet, in dem weitere Maßnahmen wie die Berücksichtigung eines Sprachmodells durchgeführt werden. Dabei wird die Anzahl der möglichen Pfade in diesem Wortgitter verringert, indem für jedes Wort nur das optimale Vorgängerwort bzw. die optimale Vortgängerwortkette bei Berücksichtigung des Sprachmodells beibehalten wird. Aus der Verfolgung der übrig bleibenden Pfade nach rückwärts kann eine einzige Wortfolge als günstigste Folge ermittelt und ausgegeben werden.

75. 发明公开

EP0706171A1 Speech recognition method and apparatus 失效
标题翻译： Einrichtung und Verfahren zur Spracherkennung
公开(公告)号：EP0706171A1
公开(公告)日：1996-04-10
申请号：EP95306890.5
申请日：1995-09-29
申请人： CANON KABUSHIKI KAISHA
发明人： Komori, Yasuhiro, c/o Canon Kabushiki Kaisha , Ohora, Yasunori, c/o Canon Kabushiki Kaisha , Yamada, Masayuki, c/o Canon Kabushiki Kaisha
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06 , G10L9/18
CPC分类号： G10L15/144 , G10L15/187
摘要： A speech recognition method uses continuous mixture Hidden Markov Models (HMM) for probability processing including a first type of HMM having a small number of mixtures and a second type of HMM having a larger number of mixtures. First output probabilities are formed for inputted speech using the small number of mixtures type HMM and second output probabilities are formed for the input speech using the large number of mixtures type HMM for selected states corresponding to the highest output probabilities of the first type HMM. The input speech is recognized from both the first and second output probabilities.
摘要翻译：语音识别方法使用连续混合隐马尔可夫模型（HMM）进行概率处理，包括具有少量混合物的第一类型HMM和具有较大数目混合物的第二类型HMM。使用少数混合型HMM形成用于输入语音的第一输出概率，并且使用对应于第一类型HMM的最高输出概率的选定状态的大量混合型HMM形成用于输入语音的第二输出概率。从第一和第二输出概率识别输入语音。

76. 发明公开

EP0685835A1 Speech recognition based on HMMs 失效
标题翻译： Gr on on on on“HMMs”。
公开(公告)号：EP0685835A1
公开(公告)日：1995-12-06
申请号：EP95107651.2
申请日：1995-05-19
申请人： TECNOMEN OY
发明人： Ranta, Jari
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06 , G10L9/18
CPC分类号： G10L15/144 , G10L15/02 , G10L25/12 , G10L25/18
摘要： A speech recognition method that combines HMMs and vector quantization to model the speech signal and adds spectral derivative information in the speech parameters is presented. Each state of a HMM is modelled by two different VQ-codebooks. One is trained by using the spectral parameters and the second is trained by using the spectral derivative parameters.
摘要翻译：提出了一种组合HMM和矢量量化以对语音信号进行建模并在语音参数中添加频谱导数信息的语音识别方法。 HMM的每个状态由两个不同的VQ码本建模。通过使用光谱参数训练一个，并且通过使用光谱衍生参数来训练第二个。

77. 发明公开

EP0680035A1 Erroneous input processing method and apparatus in an information processing system using composite input 失效
标题翻译：用于与复合输入容易出错的信息处理系统的处理方法和装置。
公开(公告)号：EP0680035A1
公开(公告)日：1995-11-02
申请号：EP95105941.9
申请日：1995-04-20
申请人： HITACHI, LTD.
发明人： Ando, Haru , Kikuchi, Hideaki , Hataoka, Nobuo , Matsuda, Yasumasa , Oheda, Shigeto , Hasegawa,Tsukasa
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06
CPC分类号： G06K9/033 , G06F3/167 , G10L15/22 , G10L15/26 , G10L2015/223 , G10L2015/226 , G10L2015/228
摘要： A user inputs voice through a voice recognition program (13), a microphone (8) and an A/D converter (7) while pointing by use of a pointing gesture, touch pen or the like with reference to an image displayed on a display unit (4). For the result of recognition of the inputted voice, a processing or display indicated by a candidate having the first rank of reliability of recognition is performed and an indication showing a plurality of candidates having the second rank and the lower ranks than that is displayed in a menu form on a display screen (21). In the case where there is an error (that is, in the case where the processing or display indicated by the candidate having the first rank is not a processing intended by the user or the user makes an erroneous input), the error is corrected in such a manner that a correct input candidate is selected by a finger, pen or the like from the displayed menu of candidates having the second rank and the lower ranks than that and a processing operation or display associated with the selected candidate is performed again. Information being redundant or duplicative as compared with the time of processing of the candidate having the first rank is held in a system as it is, thereby reducing a step in which the selection by the user from the candidates of recognition must be made at least one or more times or a labor in which the input by the user must be made again. The result of the processing desired by the user can be obtained simply by inputting only the correct input candidate again, thereby providing an interactive system which is natural and easy to use.
摘要翻译：用户输入语音通过语音识别程序（13），麦克风（8）和A / D转换器（7），同时通过使用指点手势的指向，触摸笔或参照在显示器上显示的图像等单元（4）。用于识别所输入的语音的结果，处理或显示由具有识别的可靠性的第一秩的候选被执行并且在指示表示具有第二秩候选的多元性表示，并且比下部行列被显示在的显示屏幕（21）上的菜单形式。在存在错误的情况下（即，在处理或显示由具有第一秩候选指示的情况下，不能由用户意图的处理或用户对错误的输入），在错误被纠正求的方式做了正确的输入候选被手指，笔或从具有所述第二秩候选的显示的菜单等，并比下行列和操作或与所选择的候选关联的显示再一次执行的处理选择。信息是多余的或重复的，与具有所述第一秩的候选的处理时相比，在一个系统保持原样，从而减小通过从识别的候选用户的选择必须使至少一个步骤次以上或由用户输入必须再次做了一个实验。用户所期望的处理结果可以简单地通过再次输入只有正确输入候选，从而提供的所有这一切是自然的，易于使用的交互系统获得。

78. 发明公开

EP0642118A1 Automatic system for guided acquistion of telephone line speech signals 失效
标题翻译： Automatisches SystemfürgeführtenZugriff auf Sprachsignale von Telefonlinien。
公开(公告)号：EP0642118A1
公开(公告)日：1995-03-08
申请号：EP94113018.9
申请日：1994-08-20
申请人： ALCATEL ITALIA S.p.A.
发明人： De Santis, Gerardo , Riccio, Antonello , Rigosi, Francesca
IPC分类号： G10L9/06 , G10L9/08 , G10L5/06 , G10L7/08
CPC分类号： H04M3/5158 , H04M2201/40 , H04M2203/2016
摘要： The invention relates to an automatic system for guided acquisition of speech signals from a telephone line.
By automating the acquisition process of speech signals under control of a unique intelligent control unit, the work of the telephone operator is made very easier.
摘要翻译：本发明涉及一种用于从电话线路引导获取语音信号的自动系统。通过在独特的智能控制单元的控制下使语音信号的采集过程自动化，电话操作员的工作变得更加容易。

79. 发明公开

EP0629997A1 Voice communication method and apparatus 失效
标题翻译： Verfahren und Vorrichtung zur Sprachkommunikation。
公开(公告)号：EP0629997A1
公开(公告)日：1994-12-21
申请号：EP94304344.8
申请日：1994-06-15
申请人： CANON KABUSHIKI KAISHA
发明人： Yamada, Masayuki, c/o Canon Kabushiki Kaisha , Ohora, Yasunori, c/o Canon Kabushiki Kaisha , Komori, Yasuhiro, c/o Canon Kabushiki Kaisha
IPC分类号： G10L5/06 , G10L7/08 , G10L9/06
CPC分类号： G10L15/22 , G10L15/063 , G10L15/1815 , G10L2015/228
摘要： The invention relates to voice communication method and apparatus and it is an object that when the acceptable vocabulary or grammar is dynamically changed in accordance with a communication state, a troublesomeness for allowing the user to again utter a speech which does not correspond to the predicted vocabulary and which cannot be acceptable is eliminated. For the above purpose, the speech which cannot be perceived is handled as an unknown word, a question to induce an answer such that the unknown word becomes included the acceptable vocabulary is performed, and the unknown word portion is reevaluated at a time point when the unknown word becomes included the acceptable vocabulary due to the induced answer. Thus, the user does not need to again utter the same answer.
摘要翻译：本发明涉及语音通信方法和装置，其目的是当根据通信状态动态地改变可接受的词汇或语法时，允许用户再次发出不符合预测词汇的语音的麻烦并且不能接受的是消除。为了上述目的，不能被感知的语音被作为一个未知的单词来处理，这是一个引起答案的问题，使得未知词被包括在可接受的词汇中，并且在未知词部分被重新评估时未知单词由于引起答案而被包括在可接受的词汇中。因此，用户不需要再次发出相同的答案。

80. 发明公开

EP0626674A1 A method and apparatus for speech encoding, speech decoding and speech post processing 失效
标题翻译：对于语音编码和语音解码和Sprachnachverarbeitung方法和装置。
公开(公告)号：EP0626674A1
公开(公告)日：1994-11-30
申请号：EP94106988.2
申请日：1994-05-04
申请人： MITSUBISHI DENKI KABUSHIKI KAISHA
发明人： Ishii, Jun
IPC分类号： G10L9/14 , G10L9/18 , G10L5/06 , G10L7/08
CPC分类号： G10L19/06 , G10L19/002 , G10L19/0212 , G10L19/24 , G10L19/26
摘要： A speech analysis means and a window locating means are implemented in a speech coding apparatus. The speech coding apparatus encodes input speech per analysis frame defined having a fixed length and is offset at fixed interval. The speech analysis means extracts frequency spectrum characteristic parameters of the input speech taken within an analysis window. The location of the analysis window is specified by the window locating means. The window locating means selects the location of the analysis window which is used in extracting the frequency spectrum characteristic parameters at the speech analysis means. In this case, depending upon the characteristic parameter of the input speech within and near the frame concerned, the window locating means selects the location of the analysis window within the range which is not to be exceeding the range of the frame concerned.
摘要翻译：一种语音分析装置和窗口定位装置被实现在一语音编码装置。该语音编码装置进行编码的输入语音每具有固定长度的定义的分析帧，并在固定的时间间隔偏移。语音分析装置内窗的分析采取的输入语音的提取物频谱特性的参数。分析窗口的位置由窗口定位装置指定。定位装置的窗口选择了分析窗口中的所有其在提取在语音分析装置的频谱特性参数使用的位置。在这种情况下，取决于该输入语音的内和关注帧附近的特征参数，窗口定位装置选择所述分析窗口的位置的范围内的所有这是不被超过帧关注的范围内。

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