知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

541 records (took 0.02 seconds)

    Blind diarization of recorded calls with arbitrary number of speakers
    451.
    发明授权
    Blind diarization of recorded calls with arbitrary number of speakers 有权
    用任意数量的扬声器对被录制的通话进行盲目的梳理

    公开(公告)号：US09460722B2

    公开(公告)日：2016-10-04

    申请号：US14319860

    申请日：2014-06-30

    Applicant: Verint Systems Ltd.

    Inventor： Oana Sidi ,  Ron Wein

    IPC: G10L17/16 ,  G10L17/02

    CPC classification number: G10L17/06 ,  G10L15/02 ,  G10L15/26 ,  G10L17/02 ,  G10L17/04 ,  G10L17/16 ,  G10L25/78 ,  G10L2015/025 ,  H04M3/5175 ,  H04M2201/41 ,  H04M2203/303

    Abstract: In a method of diarization of audio data, audio data is segmented into a plurality of utterances. Each utterance is represented as an utterance model representative of a plurality of feature vectors. The utterance models are clustered. A plurality of speaker models are constructed from the clustered utterance models. A hidden Markov model is constructed of the plurality of speaker models. A sequence of identified speaker models is decoded.

    Abstract translation: 在音频数据的分类方法中，将音频数据分割为多个话语。 每个话语被表示为代表多个特征向量的话语模型。 话语模型是聚类的。 从群集话语模型构建多个说话者模型。 由多个扬声器模型构成隐马尔可夫模型。 已识别的扬声器模型的序列被解码。

    Biometrics platform provider
    452.
    发明授权
    Biometrics platform provider 有权
    生物识别平台提供商

    公开(公告)号：US09349373B1

    公开(公告)日：2016-05-24

    申请号：US14213535

    申请日：2014-03-14

    Applicant: Verint Americas Inc.

    Inventor： Jamie Richard Williams ,  Robert John Barnes ,  Ian Linsdell ,  Scott M. Bluman

    IPC: G10L17/00

    CPC classification number: G10L17/06 ,  G10L17/005 ,  G10L17/04 ,  G10L25/51 ,  H04M3/42221 ,  H04M3/4936 ,  H04M3/5166 ,  H04M2201/41 ,  H04M2203/551 ,  H04M2203/6045 ,  H04M2203/6054

    Abstract: Systems and methods for analyzing digital recordings of the human voice in order to find characteristics unique to an individual. A biometrics engine may use an analytics service in a contact center to supply audio streams based on configured rules and providers for biometric detection. The analytics service may provide call audio data and attributes to connected engines based on a provider-set of selection rules. The connected providers send call audio data and attributes through the analytics service. The engines are notified when a new call is available for processing and can then retrieve chunks of audio data and call attributes by polling an analytics service interface. A mathematical model of the human vocal tract in the call audio data is created and/or matched against existing models. The result is analogous to a fingerprint, i.e., a pattern unique to an individual to within some level of probability.

    Abstract translation: 用于分析人声的数字录音的系统和方法，以便发现个人独特的特征。 生物识别引擎可以使用联络中心中的分析服务来基于配置的规则和提供者提供用于生物特征检测的音频流。 分析服务可以基于选择规则的提供者集合来向所连接的引擎提供呼叫音频数据和属性。 连接的提供商通过分析服务发送呼叫音频数据和属性。 当新的呼叫可用于处理时，引擎被通知，然后可以通过轮询分析服务接口来检索音频数据块和呼叫属性块。 在呼叫音频数据中的人声道的数学模型被创建和/或与现有模型匹配。 结果类似于指纹，即，在某种概率水平内对于个人特有的模式。

    Technology for combating mobile phone criminal activity
    458.
    发明授权
    Technology for combating mobile phone criminal activity 有权
    打击手机犯罪活动的技术

    公开(公告)号：US09264532B2

    公开(公告)日：2016-02-16

    申请号：US13756465

    申请日：2013-01-31

    Applicant: International Business Machines Corporation

    Inventor： Fernando A. Camargo

    IPC: G10L15/04 ,  H04M3/22 ,  H04L29/06 ,  H04M3/42

    CPC classification number: H04M3/2281 ,  H04L63/30 ,  H04M3/42221 ,  H04M2201/40 ,  H04M2201/41 ,  H04M2207/18

    Abstract: Technology for crime control includes receiving a voucher identifier for a mobile phone credit voucher purchased under duress by a victim and generating a request for a legal order directing a telecommunication service provider to obtain certain information about use of the voucher. Approval for the legal order is received and the legal order and the voucher identifier are transmitted by a law enforcement agency computer system via a network to a computer system of the telecommunication service provider. A phone number associated with a mobile phone to which a credit associated with the voucher identifier was applied and a recording of a telephone call to or from the phone number are received via the network from the telecommunication service provider computer system and the law enforcement agency computer system performs an automated analysis of the call by a voice recognition process.

    Abstract translation: 犯罪控制技术包括收到受害人胁迫下购买的移动电话信用凭证的凭证标识符，并产生指示电信服务提供商获得关于使用凭证的某些信息的法律命令的请求。 收到法定命令的批准，法律令和凭证标识符由执法机构计算机系统通过网络传送到电信服务提供商的计算机系统。 通过网络从电信服务提供商计算机系统和执法机构计算机接收与应用了与凭证标识符相关联的信用以及从电话号码记录电话呼叫的移动电话相关联的电话号码 系统通过语音识别过程执行呼叫的自动分析。

    Method and system for speaker verification
    459.
    发明授权
    Method and system for speaker verification 有权
    说话人验证方法和系统

    公开(公告)号：US09258425B2

    公开(公告)日：2016-02-09

    申请号：US13900163

    申请日：2013-05-22

    Applicant: Nuance Communications, Inc.

    Inventor： Nir Moshe Krause

    IPC: H04M1/64 ,  H04M3/523 ,  G10L25/51 ,  H04M3/51 ,  G10L17/02 ,  G10L17/00

    CPC classification number: H04M3/523 ,  G10L17/00 ,  G10L17/02 ,  G10L25/51 ,  H04M3/51 ,  H04M2201/41

    Abstract: In many scenarios, speaker verification systems can be given a single-channel audio with recordings of multiple speakers. To perform accurate speaker verification, a system can isolate the speech of a speaker. In one embodiment, a method, and corresponding system, of speaker verification includes extracting a target speaker's speech, using a known speaker voiceprint, from an audio recording that includes the target speaker's speech and the known speaker's speech. The known speaker voiceprint can correspond to the known speaker. Extracting the target speaker's speech can include determining portions of the audio recording where the known speaker voiceprint matches the known speaker's speech above a particular threshold, and extracting the target speaker's speech from other portions of the audio recording. In this manner, speaker verification is performed on the target speaker's speech without interference from the known speaker's speech and allows for a more accurate verification.

    Abstract translation: 在许多情况下，扬声器验证系统可以被给予具有多个扬声器记录的单声道音频。 为了执行准确的扬声器验证，系统可以隔离扬声器的语音。 在一个实施例中，扬声器验证的方法和对应系统包括使用已知的扬声器声纹从包括目标说话者的语音和已知的说话者的语音的音频记录中提取目标说话者的语音。 已知的扬声器声纹可以对应于已知的扬声器。 提取目标说话者的语音可以包括确定音频记录的部分，其中已知的说话者声纹与特定阈值之上的已知说话者的语音匹配，以及从音频记录的其他部分提取目标说话者的语音。 以这种方式，在目标讲话者的语音上执行说话人验证，而不会受到已知讲话人的言语的干扰，并允许进行更准确的验证。

    Speech analytics with adaptive filtering
    460.
    发明授权
    Speech analytics with adaptive filtering 有权
    具有自适应滤波的语音分析

    公开(公告)号：US09257117B2

    公开(公告)日：2016-02-09

    申请号：US14172297

    申请日：2014-02-04

    Applicant: Avaya, Inc.

    Inventor： Valentine C. Matula ,  Shmuel Shaffer

    IPC: G10L15/00 ,  G10L15/08 ,  G10L25/48 ,  H04M3/51

    CPC classification number: G10L15/08 ,  G10L25/48 ,  G10L25/51 ,  G10L2015/088 ,  H04M3/2227 ,  H04M3/51 ,  H04M3/5175 ,  H04M2201/40 ,  H04M2201/41 ,  H04M2203/404

    Abstract: Contact center agents often work in close proximity to other agents. As a primary agent is engaged in a call, a neighboring agent speech may be picked up by the primary agent's microphone. Contact centers using automated speech recognition systems may monitor the agent's speech for key terms and, if detected, respond accordingly. Determining a primary agent spoke a key term, when the true speaker of the key term is a neighboring agent, may cause errors or other problems. Characterizing at least the primary agent's voice and then, once a key term is detected, determining if it was the primary agent that spoke the key term, may help to reduce the errors. Additionally, computational requirements may be reduced as non-key terms may be quickly discarded and optionally, key terms determined to not have been spoke by the primary agent, may also be discarded without further processing.

    Abstract translation: 联络中心代理商经常在其他代理商附近工作。 作为主要代理人进行呼叫，可以由主代理的麦克风拾取相邻代理语音。 使用自动语音识别系统的联络中心可以监控代理人的关键词语，并且如果检测到，则相应地进行响应。 确定主要代理人是一个关键术语，当关键术语的真正发言人是相邻代理人时，可能会导致错误或其他问题。 至少表征主代理的语音，然后，一旦检测到关键术语，确定是否是主要代理人谈论关键术语，可能有助于减少错误。 此外，由于非关键术语可能被快速丢弃，并且可选地，被确定为不被主代理人讲话的关键术语也可能被丢弃，而无需进一步处理，可以减少计算要求。

Patent Agency Ranking