利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

1017 条记录 (耗时 0.022 秒)

    SPEAKER CHARACTERIZATION THROUGH SPEECH ANALYSIS
    95.
    发明公开
    SPEAKER CHARACTERIZATION THROUGH SPEECH ANALYSIS 有权
    SPRECHERCHARAKTERISIERUNG DURCH SPRACHANALSESE

    公开(公告)号：EP2304718A2

    公开(公告)日：2011-04-06

    申请号：EP09766323.1

    申请日：2009-06-17

    申请人： Voicesense Ltd.

    发明人： DEGANI, Yoav ,  ZAMIR, Yishai

    IPC分类号： G10L17/00 ,  G10L15/06 ,  G10L15/18

    CPC分类号： G10L15/06 ,  G10L15/1807 ,  G10L17/26

    摘要： A computer implemented method, data processing system, apparatus and computer program product for determining current behavioral, psychological and speech styles characteristics of a speaker in a given situation and context, through analysis of current speech utterances of the speaker. The analysis calculates different prosodic parameters of the speech utterances, consisting of unique secondary derivatives of the primary pitch and amplitude speech parameters, and compares these parameters with pre-obtained reference speech data, indicative of various behavioral, psychological and speech styles characteristics. The method includes the formation of the classification speech parameters reference database, as well as the analysis of the speaker's speech utterances in order to determine the current behavioral, psychological and speech styles characteristics of the speaker in the given situation.

    摘要翻译： 一种计算机实现的方法，数据处理系统，装置和计算机程序产品，用于通过分析扬声器的当前语音话语来确定给定情况和语境中的扬声器的当前行为，心理和言语风格特征。 该分析计算出语音话语的不同韵律参数，由主音阶和幅度语音参数的唯一次级导数组成，并将这些参数与预先获得的参考语音数据进行比较，指示各种行为，心理和言语风格特征。 该方法包括形成分类语音参数参考数据库，以及对演讲者言语言语音的分析，以便在给定情况下确定演讲者的当前行为，心理和言语风格特征。

    AUDIO RECOGNITION DEVICE, AUDIO RECOGNITION METHOD, AND ELECTRONIC DEVICE
    96.
    发明公开
    AUDIO RECOGNITION DEVICE, AUDIO RECOGNITION METHOD, AND ELECTRONIC DEVICE 有权
    SPRACHERKENNUNGSSYSTEM UNDERVERHREN。

    公开(公告)号：EP2293289A1

    公开(公告)日：2011-03-09

    申请号：EP09758194.6

    申请日：2009-05-11

    申请人： Raytron, Inc.

    发明人： YOSHIDA, Mitsuji ,  HYODO, Kazutaka

    IPC分类号： G10L15/08 ,  G10L15/06 ,  G10L15/10

    CPC分类号： G10L15/08 ,  G10L15/10 ,  G10L15/20

    摘要： A speech characteristic-amount calculation circuit 31 calculates an amount of speech characteristics of each phrase in input speech. An estimation process likelihood calculation circuit 33 compares the calculated speech characteristic amount of a phrase with speech pattern sequence information of a plurality of phrases stored in a storage unit 34 to select a plurality of candidates having from a higher likelihood value to a lower likelihood value for the phrases. A recognition filtering device 4 determines whether to reject or not reject the extracted candidates based on the likelihood difference ratio between the difference in likelihood values between the first candidate and the second candidate and the difference in likelihood values between the second candidate and the third candidate.

    摘要翻译： 语音特征量计算电路31计算输入语音中的每个短语的语音特征量。 估计处理似然度计算电路33将所计算出的语音特征量与存储在存储单元34中的多个短语的语音模式序列信息进行比较，以选择具有较高似然值的多个候选者到较低似然值， 短语 识别过滤装置4基于第一候选者和第二候选者之间的似然值的差异和第二候选者与第三候选者之间的似然值的差异之间的似然差比确定是否拒绝或不拒绝所提取的候选。

    Method for automated training of a plurality of artificial neural networks
    98.
    发明公开
    Method for automated training of a plurality of artificial neural networks 有权
    Verfahren zum automatisierten培训einer Vielzahlkünstlicher神经元Netzwerke

    公开(公告)号：EP2221805A1

    公开(公告)日：2010-08-25

    申请号：EP09002464.7

    申请日：2009-02-20

    申请人： Harman Becker Automotive Systems GmbH

    发明人： Vasquez, Daniel ,  Guillermo, Aradilla ,  Gruhn, Rainer

    IPC分类号： G10L15/06 ,  G10L15/16

    CPC分类号： G10L15/16 ,  G06N3/0454 ,  G10L15/063

    摘要： The invention provides a method for automated training of a plurality of artificial neural networks for phoneme recognition using training data, wherein the training data comprises speech signals subdivided into frames, each frame associated with a phoneme label, wherein the phoneme label indicates a phoneme associated with the frame, the method comprising the steps of:
    providing a sequence of frames from the training data, wherein the number of frames in the sequence of frames is at least equal to the number of artificial neural networks,
    assigning to each of the artificial neural networks a different subsequence of the provided sequence, wherein each subsequence comprises a predetermined number of frames,
    determining a common phoneme label for the sequence of frames based on the phoneme labels of one or more frames of one or more subsequences of the provided sequence, and
    training each artificial neural network using the common phoneme label.

    摘要翻译： 本发明提供一种用于使用训练数据进行音素识别的多个人造神经网络的自动训练的方法，其中训练数据包括细分为帧的语音信号，每个帧与音素标签相关联，其中，音素标签指示与 所述方法包括以下步骤：从所述训练数据提供帧序列，其中所述帧序列中的帧数至少等于所述人造神经网络的数量，分配给每个所述人造神经网络 提供的序列的不同子序列，其中每个子序列包括预定数量的帧，基于所提供序列的一个或多个子序列的一个或多个帧的音素标签确定帧序列的公共音素标签，以及训练 每个人造神经网络使用普通的音素标签。

    Speaker recognition
    99.
    发明公开
    Speaker recognition 有权
    Sprechererkennung

    公开(公告)号：EP2216775A1

    公开(公告)日：2010-08-11

    申请号：EP09001624.7

    申请日：2009-02-05

    申请人： Harman Becker Automotive Systems GmbH

    发明人： Herbig, Tobias ,  Gerl, Franz, Dr.

    IPC分类号： G10L17/00 ,  G10L15/06

    CPC分类号： G10L15/07 ,  G10L17/06

    摘要： A method of recognizing a speaker of an utterance (602) in a speech recognition system, comprising
    - comparing the utterance (602) to a plurality of speaker models (604) for different speakers;
    - determining a likelihood score (606) for each speaker model, the likelihood score (606) indicating how well the speaker model corresponds to the utterance; and
    - for each speaker model (604), determining a probability (609) that the utterance (602) originates from the speaker corresponding to the speaker model (604),
    wherein the determination of the probability (609) for a speaker model (604) is based on the likelihood scores (606) for the speaker models and takes a prior knowledge (607) about the speaker model into account.

    摘要翻译： 一种在语音识别系统中识别说话者的方法（602），包括：将所述话语（602）与不同说话者的多个说话者模型（604）进行比较; - 确定每个说话者模型的可能性得分（606），所述可能性得分（606）指示说话者模型对应于话语的程度; 以及 - 对于每个说话者模型（604），确定所述话语（602）源自与所述说话者模型（604）相对应的说话者的概率（609），其中，确定所述说话者模型（604）的概率（609） ）基于说话者模型的可能性得分（606），并且考虑到关于说话者模型的先验知识（607）。