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. 发明申请

WO2019097217A1 AUDIO PROCESSING 审中-公开
公开(公告)号：WO2019097217A1
公开(公告)日：2019-05-23
申请号：PCT/GB2018/053275
申请日：2018-11-13
申请人： CIRRUS LOGIC INTERNATIONAL SEMICONDUCTOR LIMITED
发明人： LESSO, John Paul , MCLEOD, Gordon Richard
IPC分类号： G10L17/02
CPC分类号： G10L17/04 , G10L17/02 , G10L25/45 , G10L25/51
摘要： A method of audio processing comprises receiving an audio signal. A plurality of framed versions of the received audio signal are formed, each of the framed versions having a respective frame start position. One of the plurality of framed versions of the received audio signal is selected. The selected one of the plurality of framed versions of the received audio signal is used in a subsequent process.

2. 发明申请

WO2018232233A1 DATA UPDATE METHOD, CLIENT, AND ELECTRONIC DEVICE 审中-公开
公开(公告)号：WO2018232233A1
公开(公告)日：2018-12-20
申请号：PCT/US2018/037741
申请日：2018-06-15
申请人： ALIBABA GROUP HOLDING LIMITED
发明人： LIU, Gang , ZHAO, Qingen , LIU, Guangxing
IPC分类号： G06F17/20 , G10L15/06 , G10L15/08 , G10L17/22 , G10L25/57
CPC分类号： G10L17/22 , G10L17/02 , G10L17/04 , G10L17/06
摘要： Data update methods, systems, and devices are disclosed. The method includes: acquiring at least a first piece of audio data of a user in a first conversation scenario and at least a second piece of audio data of the user in a second conversation scenario, performing voiceprint recognition on the first and second pieces of audio data based on voiceprint information of the user, acquiring audio feature information of the first and second pieces of audio data, and updating voiceprint information of the user according to the audio feature information.

3. 发明申请

WO2018107810A1 声纹识别方法、装置、电子设备及介质审中-公开
公开(公告)号：WO2018107810A1
公开(公告)日：2018-06-21
申请号：PCT/CN2017/099707
申请日：2017-08-30
申请人：平安科技(深圳)有限公司
发明人：王健宗 , 郭卉 , 肖京
IPC分类号： G10L17/00
CPC分类号： G10L17/00 , G10L17/02 , G10L17/04 , G10L17/18 , G10L17/20 , G10L25/24 , G10L25/30
摘要：提供了适用于身份认证技术领域的声纹识别方法、装置、电子设备及介质。方法包括：对输入的语音进行预处理，获取语音中的有效语音；提取语音的MFCC声学特征，输出包含MFCC维度及语音分帧数的第一和第二特征矩阵；构建长短时递归神经网络模型，并将第一特征矩阵作为输入；利用神经网络模型的训练参数及语音的说话人特征训练特征提取矩阵，每个特征提取矩阵对应一个说话人模型；选取出匹配第二特征矩阵的说话人模型，匹配的说话人模型对应的说话人输出为声纹识别结果。能够从训练语音中挖掘出更合适的声学特征，从而能够更准确地辨别说话人的差异性特征，学习到鲁棒性更强的说话人模型，获取更好的声纹识别效果。

4. 发明申请

WO2018095167A1 声纹识别方法和声纹识别系统审中-公开
公开(公告)号：WO2018095167A1
公开(公告)日：2018-05-31
申请号：PCT/CN2017/106886
申请日：2017-10-19
申请人：北京京东尚科信息技术有限公司 , 北京京东世纪贸易有限公司
发明人：雷利博 , 薛韬 , 罗超
IPC分类号： G10L17/02
CPC分类号： G10L17/02 , G10L17/04
摘要：一种声纹识别方法和系统，方法包括：接收待测试音频并将其分割为第一和第二部分；选择一个样本音频并将其分割为第一和第二部分；通过使用梅尔倒谱系数的提取方法，提取针对待测试音频和样本音频的特征矩阵；通过将待测试音频的第一部分的特征矩阵作为第一类样本，并将所选样本音频的特征矩阵作为第二类样本，执行支持向量机训练，并计算待测试音频的第二部分与第二类样本的匹配程度；以类似方式针对样本音频的第一部分、待测试音频的第一部分和样本音频的第二部分，分别计算其与对应作为第二类样本的待测试音频、所选样本音频和待测试音频的匹配程度；根据各匹配程度，确定待测试音频和样本音频是否来自同一个人的声音。

5. 发明申请

WO2017172113A1 SPEAKER RECOGNITION USING ADAPTIVE THRESHOLDING 审中-公开
标题翻译：使用自适应阈值进行扬声器识别
公开(公告)号：WO2017172113A1
公开(公告)日：2017-10-05
申请号：PCT/US2017/018716
申请日：2017-02-21
申请人： INTEL CORPORATION
发明人： BISWAL, Narayan , CILINGIR, Gokcen
IPC分类号： G10L17/12 , G10L17/20 , G10L21/02
CPC分类号： G10L17/20 , G10L15/20 , G10L15/265 , G10L17/02 , G10L17/04 , G10L17/06 , G10L17/22 , G10L25/78 , G10L25/81
摘要： Techniques related to speaker recognition are discussed. Such techniques may include determining an adaptive speaker recognition threshold based on a speech to noise ratio and noise type label corresponding to received audio and performing speaker recognition based on the adaptive speaker recognition threshold and a speaker recognition score corresponding to received audio.
摘要翻译：讨论了与说话人识别有关的技术。这样的技术可以包括基于与接收到的音频对应的语音噪声比和噪声类型标签来确定自适应说话人识别阈值，并且基于自适应说话人识别阈值和与接收到的音频对应的说话者识别分数来执行说话者识别。

6. 发明申请

WO2017157423A1 SYSTEM, APPARATUS, AND METHOD FOR PERFORMING SPEAKER VERIFICATION USING A UNIVERSAL BACKGROUND MODEL 审中-公开
标题翻译：使用通用背景模型执行扬声器验证的系统，装置和方法
公开(公告)号：WO2017157423A1
公开(公告)日：2017-09-21
申请号：PCT/EP2016/055564
申请日：2016-03-15
申请人： TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
发明人： GRANCHAROV, Volodya , KARLSSON, Erlendur , SVERRISSON, Sigurdur , POBLOTH, Harald
IPC分类号： G10L17/06 , G10L17/04
CPC分类号： G10L17/06 , G10L17/04
摘要： An apparatus (102/108), machine-implemented method (300), and computer program product for determining whether an utterance, U, was spoken by a hypothesized speaker, S, are presented. The apparatus includes a module (110a/110b) that obtains data corresponding to an utterance captured by a microphone. It also obtains a set of two or more Gaussian Mixture Models (GMMs), where each GMM comprises a set of parameter pairs. The module further selects, based on the obtained data corresponding to the captured utterance U, a subset of a first set of parameter pairs of the first GMM and a subset of a second set of parameter pairs of the second GMM. The two subsets are combined to form a Universal Background Model (UBM), λ UBM . The module obtains a GMM model, λ S , associated with the hypothesized speaker, S. The segment λ S , λ UBM , and the obtained data corresponding to U are used to determine whether U was spoken by S.
摘要翻译：呈现了用于确定是否由假设的说话者S说出话语U的设备（102/108），机器实现的方法（300）和计算机程序产品。该设备包括获得与由麦克风捕获的话语相对应的数据的模块（110a / 110b）。它还获得一组两个或更多个高斯混合模型（GMM），其中每个GMM包括一组参数对。该模块还基于所获得的与捕获的话语U相对应的数据选择第一GMM的第一组参数对的子集和第二GMM的第二组参数对的子集。这两个子集合起来形成一个通用背景模型（UBM），即λUBM。该模块获得与假设说话者S相关联的GMM模型λS。该段λS S，λUBM和所获得的对应于U的数据用于确定U是否由S说出。

7. 发明申请

WO2017148523A1 NON-PARAMETRIC AUDIO CLASSIFICATION 审中-公开
标题翻译：非参数音频分类
公开(公告)号：WO2017148523A1
公开(公告)日：2017-09-08
申请号：PCT/EP2016/054586
申请日：2016-03-03
申请人： TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
发明人： GRANCHAROV, Volodya , SVERRISSON, Sigurdur
IPC分类号： G10L15/08 , G10L15/00 , G10L15/10 , G10L15/28 , G10L17/02 , G10L17/04 , G10L25/51
CPC分类号： G10L25/51 , G10L15/005 , G10L15/08 , G10L15/10 , G10L15/285 , G10L17/02 , G10L17/04 , G10L2015/085
摘要： There is provided mechanisms for non-parametric audio classification. A method is performed by a classification device. The method comprises obtaining a short-term frequency representation of an audio waveform, the short-term frequency representation defining an input sequence divided into input vectors. The method comprises determining per-class posterior probabilities for at least two classes. Each per-class posterior probability is based on a weighted sum of pre-stored per-cluster posterior probabilities for the at least two classes. Each class represents a unique audio classification property. The method comprises classifying the input sequence to belong to the class for which the per-class posterior probability is largest.
摘要翻译：提供了非参数音频分类的机制。一种方法由分类装置执行。该方法包括获得音频波形的短期频率表示，该短期频率表示定义了被分成输入向量的输入序列。该方法包括确定至少两类的每类后验概率。每个类别后验概率基于至少两个类别的预先存储的每个群集后验概率的加权和。每个类表示一个独特的音频分类属性。该方法包括将输入序列分类为属于类别后验概率最大的类别。

8. 发明申请

WO2016113521A1 AUTHENTICATION METHOD 审中-公开
标题翻译：认证方法
公开(公告)号：WO2016113521A1
公开(公告)日：2016-07-21
申请号：PCT/GB2015/052771
申请日：2015-09-24
申请人： VALIDSOFT UK LIMITED
发明人： PETERSEN, John , THORNHILL, Daniel
IPC分类号： G10L17/04 , G10L17/22 , G10L17/10 , G06F21/32 , H04L9/32 , H04L29/06
CPC分类号： G06F21/32 , G10L17/04 , G10L17/10 , G10L17/22 , H04L9/3231 , H04L63/0861 , H04L2463/082
摘要： An authentication method. The method comprises comparing user voice data received via an electronic device to a stored voice template to determine a voice authentication parameter. A voice authentication threshold is determined and the voice authentication parameter is compared to the voice authentication threshold to determine whether to authenticate the user. Determining the voice authentication threshold comprises determining a current value of an enrolment counter, then comparing the current value of the enrolment counter to an enrolment counter threshold and determining whether the stored voice template is fully enrolled according to the result. If the stored voice template is fully enrolled, the voice authentication threshold is set to a first voice authentication threshold. If the stored voice template is not fully enrolled then a device attribute received from the electronic device is compared to a stored device attribute. If the received device attribute matches the stored device attribute, the voice authentication threshold is set to a second voice authentication threshold determined by the current value of the enrolment counter. If the received device attribute does not match the stored device attribute, the voice authentication threshold is set to a third voice authentication threshold.
摘要翻译：一种认证方式。该方法包括将通过电子设备接收的用户语音数据与存储的语音模板进行比较，以确定语音认证参数。确定语音认证阈值，并将语音认证参数与语音认证阈值进行比较，以确定是否对用户进行认证。确定语音认证阈值包括确定注册计数器的当前值，然后将注册计数器的当前值与注册计数器阈值进行比较，并根据结果确定所存储的语音模板是否完全注册。如果存储的语音模板完全注册，则将语音认证阈值设置为第一语音认证阈值。如果存储的语音模板未被完全注册，则从电子设备接收的设备属性与存储的设备属性进行比较。如果接收到的设备属性与存储的设备属性相匹配，则将语音认证阈值设置为由注册计数器的当前值确定的第二语音认证阈值。如果接收到的设备属性与存储的设备属性不匹配，则将语音认证阈值设置为第三语音认证阈值。

9. 发明申请

WO2016053523A1 SPEAKER IDENTIFICATION AND UNSUPERVISED SPEAKER ADAPTATION TECHNIQUES 审中-公开
标题翻译：扬声器识别和不可支持的扬声器适配技术
公开(公告)号：WO2016053523A1
公开(公告)日：2016-04-07
申请号：PCT/US2015/047281
申请日：2015-08-27
申请人： APPLE INC.
发明人： KIM, Yoon , KAJAKEKAR, Sachin, S.
IPC分类号： G10L17/04 , G10L17/06 , G10L15/18 , G10L15/26
CPC分类号： G10L17/26 , G10L15/1822 , G10L15/26 , G10L17/04 , G10L17/06
摘要： Systems and processes for generating a speaker profile for use in performing speaker identification for a virtual assistant are provided. One example process can include receiving an audio input including user speech and determining whether a speaker of the user speech is a predetermined user based on a speaker profile for the predetermined user. In response to determining that the speaker of the user speech is the predetermined user, the user speech can be added to the speaker profile and operation of the virtual assistant can be triggered. In response to determining that the speaker of the user speech is not the predetermined user, the user speech can be added to an alternate speaker profile and operation of the virtual assistant may not be triggered. In some examples, contextual information can be used to verify results produced by the speaker identification process.
摘要翻译：提供了用于生成用于为虚拟助理执行说话者识别的扬声器简档的系统和过程。一个示例性过程可以包括基于用于预定用户的扬声器简档来接收包括用户语音的音频输入并且确定用户语音的扬声器是否是预定用户。响应于确定用户语音的扬声器是预定用户，可以将用户语音添加到扬声器简档，并且可以触发虚拟助手的操作。响应于确定用户语音的讲话者不是预定用户，可以将用户语音添加到备用讲话者简档，并且虚拟助理的操作可能不被触发。在一些示例中，可以使用上下文信息来验证由说话者识别过程产生的结果。

10. 发明申请

WO2015161240A2 SPEAKER VERIFICATION 审中-公开
标题翻译：扬声器验证
公开(公告)号：WO2015161240A2
公开(公告)日：2015-10-22
申请号：PCT/US2015/026453
申请日：2015-04-17
申请人： QUALCOMM INCORPORATED
发明人： VISSER, Erik , KIM, Lae-Hoon , JIN, Minho , GUO, Yinyi
IPC分类号： G10L17/04
CPC分类号： G06F3/167 , G10L15/22 , G10L17/04 , G10L17/08 , G10L17/22
摘要： A device includes a memory, a receiver, a processor, and a display. The memory is configured to store a speaker model. The receiver is configured to receive an input audio signal. The processor is configured to determine a first confidence level associated with a first portion of the input audio signal based on the speaker model. The processor is also configured to determine a second confidence level associated with a second portion of the input audio signal based on the speaker model. The display is configured to present a graphical user interface associated with the first confidence level or associated with the second confidence level.
摘要翻译：设备包括存储器，接收器，处理器和显示器。存储器被配置为存储扬声器模型。接收器被配置为接收输入音频信号。处理器被配置为基于说话者模型来确定与输入音频信号的第一部分相关联的第一置信水平。处理器还被配置为基于说话者模型来确定与输入音频信号的第二部分相关联的第二置信度。显示器被配置为呈现与第一置信水平相关联或与第二置信水平相关联的图形用户界面。

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