CN102543076A - Speech training method and corresponding system for phonetic entry method - Google Patents
Speech training method and corresponding system for phonetic entry methodDownload PDFInfo
- Publication number
- CN102543076A CN102543076ACN2011100004314ACN201110000431ACN102543076ACN 102543076 ACN102543076 ACN 102543076ACN 2011100004314 ACN2011100004314 ACN 2011100004314ACN 201110000431 ACN201110000431 ACN 201110000431ACN 102543076 ACN102543076 ACN 102543076A
- Authority
- CN
- China
- Prior art keywords
- information
- voice
- user
- training
- side device
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012549trainingMethods0.000titleclaimsabstractdescription135
- 238000000034methodMethods0.000titleclaimsabstractdescription53
- 239000000284extractSubstances0.000claimsdescription15
- 238000000605extractionMethods0.000claimsdescription13
- 238000004891communicationMethods0.000description2
- 241001672694Citrus reticulataSpecies0.000description1
- 238000012905input functionMethods0.000description1
- 238000003672processing methodMethods0.000description1
Images
Landscapes
- Telephonic Communication Services (AREA)
Abstract
Translated fromChineseDescription
Translated fromChinese技术领域technical field
本发明涉及通信领域，尤其涉及一种用于语音输入法的语音训练方法及相应的系统。The invention relates to the communication field, in particular to a speech training method and a corresponding system for speech input methods.
背景技术Background technique
语音输入法具有使用简便、快速掌握、快速输入的特点，已经得到广泛的应用。现有的语音输入法的处理方式如下：首先终端侧采集语音信息并发送至网络服务器，之后网络服务器对接收到的语音进行特征识别，然后根据识别的语音特征进行声学模型匹配，最终通过语法模型，选择最佳匹配的词句。The voice input method has the characteristics of easy to use, fast mastering, and fast input, and has been widely used. The processing method of the existing voice input method is as follows: first, the terminal side collects voice information and sends it to the network server, and then the network server performs feature recognition on the received voice, and then performs acoustic model matching according to the recognized voice features, and finally passes the grammatical model , select the best matching phrase.
然而，由于用户的语言总是带有各式各样的个人特点，例如发音，语速，重音，连字等等，用户无法完全按照普通话标准进行发音，因此语音识别总是存在一定程度的不准确。However, because the user's language always has a variety of personal characteristics, such as pronunciation, speech rate, accent, ligatures, etc., the user cannot fully pronounce according to the standard of Mandarin, so there is always a certain degree of inadequacy in speech recognition. precise.
为了提高语音引擎对用户语音识别的准确性，主要采用了语音训练方法。用户需要在使用语音输入法前，按照终端提示的文字信息，进行朗读训练，以便语音引擎中的训练模块根据录制的语音信息与标准文本进行自适应，记录用户的个人信息，用于之后的语音识别。In order to improve the accuracy of the speech engine for user speech recognition, speech training methods are mainly used. Before using the voice input method, the user needs to perform reading training according to the text information prompted by the terminal, so that the training module in the voice engine can adapt to the recorded voice information and standard text, and record the user's personal information for future voice identify.
现有的语音训练方法至少存在以下不足：用户在使用前需要按照提示文本进行语音训练，为了有效获取用户的语音信息，语音训练一般需要20分钟甚至更长的时间。对于广大用户来说，这种语音训练是非常繁琐的且占用了用户大量的时间。The existing voice training methods have at least the following disadvantages: the user needs to perform voice training according to the prompt text before use, and in order to effectively obtain the user's voice information, the voice training generally takes 20 minutes or even longer. For the majority of users, this voice training is very cumbersome and takes a lot of time for the users.
发明内容Contents of the invention
根据本申请的一个实施方式，公开一种用于语音输入法的语音训练方法，包括：网络侧设备从终端接收用户实时输入的语音信息；所述网络侧设备对所述语音信息进行语音识别，并将识别出的文字信息发送至所述终端；从所述终端接收反馈信息；以及所述网络侧设备根据所述反馈信息对所述语音信息进行语音训练。According to an embodiment of the present application, a voice training method for a voice input method is disclosed, including: a network side device receives voice information input by a user in real time from a terminal; the network side device performs voice recognition on the voice information, and sending the recognized text information to the terminal; receiving feedback information from the terminal; and performing speech training on the speech information by the network side device according to the feedback information.
根据本申请的另一个实施方式，公开一种用于对语音输入法进行语音训练的系统，包括：语音信息提取模块，在终端提取用户实时输入的语音信息；语音识别模块，从所述语音信息提取模块接收所述语音信息，对所述语音信息进行语音识别；文字信息反馈模块，从所述语音识别模块接收对所述语音信息识别出的文字信息，并回送反馈信息；语音训练模块，从所述文字信息反馈模块接收所述反馈信息，并根据所述反馈信息对所述语音信息进行语音训练。According to another embodiment of the present application, a system for voice training of a voice input method is disclosed, including: a voice information extraction module, which extracts the voice information input by the user in real time at the terminal; a voice recognition module, which extracts the voice information from the voice information The extraction module receives the voice information, and performs voice recognition on the voice information; the text information feedback module receives the text information recognized from the voice information from the voice recognition module, and returns feedback information; the voice training module, from The text information feedback module receives the feedback information, and performs voice training on the voice information according to the feedback information.
附图说明Description of drawings
图1示出了本申请一个实施方式的用于语音输入法的语音训练方法的流程1000；Fig. 1 shows the
图2示出了本申请另一个实施方式的用于语音输入法的语音训练方法的流程2000；Fig. 2 shows the
图3示出了本申请一个实施方式的用于对语音输入法进行语音训练的系统100；FIG. 3 shows a
图4示出了本申请另一个实施方式的用于对语音输入法进行语音训练的系统200；FIG. 4 shows a system 200 for performing speech training on a speech input method according to another embodiment of the present application;
图5示出了本申请另一个实施方式的用于对语音输入法进行语音训练的系统300；FIG. 5 shows a
图6示出了本申请另一个实施方式的用于对语音输入法进行语音训练的系统400；FIG. 6 shows a
图7示出了图6中的网络侧设备的一个实施方式；Fig. 7 shows an embodiment of the network side device in Fig. 6;
图8示出了图6中的网络侧设备的另一个实施方式。FIG. 8 shows another implementation manner of the network side device in FIG. 6 .
具体实施方式Detailed ways
下面将结合本申请实施方式中的附图，对本申请实施方式中的技术方案进行描述。The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.
本申请中出现的终端可以是具有语音输入功能并能够进行通信的任何设备，例如手机、PDA等。以下以手机为例进行说明。The terminal mentioned in this application may be any device that has a voice input function and is capable of communication, such as a mobile phone, a PDA, and the like. The mobile phone is taken as an example for description below.
图1示出了本申请一个实施方式的用于语音输入法的语音训练方法的流程1000。如图1所示，在步骤S110中，网络侧设备从终端接收用户实时输入的语音信息。例如，手机将用户实时输入的语音信息发送给网络侧设备，网络侧设备接收并存储该语音信息。在步骤S120中，对接收到的语音信息进行语音识别，并将识别出的文字信息发送至终端。例如对语音进行特征识别，根据识别出的语音特征比对声学模型，再根据声学模型比对结果在语法模型中查找最佳匹配结果，然后将语音识别得到的文字信息返回该手机，供该用户使用。然后在步骤S130中，从终端接收反馈信息。通常该反馈信息会与语音信息内容一致。例如，终端从网络侧设备接收到的文字信息如果与语音信息一致，即该文字信息是用户所期望的，则终端会将该文字信息作为反馈信息自动回送至网络侧设备，例如在用户输入下一条语音信息时将该反馈信息回送至网络侧设备；如果不一致，则该终端接收用户的指令对该文字信息进行修改，将修改后的文字信息作为反馈信息自动回送至网络侧设备，例如在用户输入下一条语音信息时将该反馈信息回送至网络侧设备。接着在步骤S140中，根据从终端接收的反馈信息对步骤S110中接收到的语音信息进行语音训练。FIG. 1 shows a
在该实施方式中，网络侧设备对用户实时输入的语音信息进行语音识别并将识别出的文字信息返回终端，用户通过终端可以对该文字信息进行编辑、修改，然后将自动发送反馈信息至网络侧设备，网络侧设备根据反馈信息对之前接收到的语音信息进行语音训练。因此，用户不需要在使用语音输入法之前按照提示文本进行专门语音训练，而是在使用语音输入法的过程中进行语音训练，大大提高了用户对语音输入法的用户体验。In this embodiment, the network side device performs voice recognition on the voice information input by the user in real time and returns the recognized text information to the terminal. The user can edit and modify the text information through the terminal, and then automatically sends feedback information to the network. side device, and the network side device performs speech training on the previously received speech information according to the feedback information. Therefore, the user does not need to perform special voice training according to the prompt text before using the voice input method, but conducts voice training during the process of using the voice input method, which greatly improves the user experience of the voice input method.
图2示出了本申请另一个实施方式的用于语音输入法的语音训练方法的流程2000。在步骤S210中，网络侧设备接收来自终端的用户的语音信息。在步骤S211中，网络侧设备获取用户的用户ID。接着在步骤S212中，根据用户ID从语音训练信息数据库中提取与该用户对应的语音训练信息。作为一种选择，语音训练信息数据库可以存储在该网络侧设备中，网络侧设备还可以为每个用户设立对应的数据库用于存储每个用户的语音训练信息。接着在步骤S220中，网络侧设备根据所提取的语音训练信息对步骤S210中接收到的语音信息进行语音识别，并将语音识别得到的文字信息发送给终端。然后在步骤S230中，网络侧设备从终端接收反馈信息。接着在步骤S240中，根据反馈信息对S210中接收到的语音信息进行语音训练。后在步骤S250中，根据步骤S250中语音训练得到结果更新语音训练信息数据库。作为一种选择，还可在步骤S250中删除S210中存储的语音信息。FIG. 2 shows a
在本发明的一个实施方式中，用户ID可以是可以是手机号。当用户ID是手机号时，则在步骤S210中该手机号与用户的语音信息一起发送至网络侧设备，以便网络侧设备在步骤S211中获取该手机号；在步骤S240中，该手机号码与反馈信息一起被发送至网络侧设备，以便网络侧设备根据该手机号将步骤S240中得到的语音训练结果存入该用户的语音训练信息数据库中。作为另一种选择，用户ID还可以是用户预先在网络侧设备注册的ID，在这种情况下，在步骤S210之前，网络侧设备会接收到该用户的登录信息，也就是说用户在使用语音输入法之前需要先通过手机登录到网络侧设备，然后在步骤S211中网络侧设备根据用户的登录信息即可获取该用户的用户ID。In one embodiment of the present invention, the user ID may be a mobile phone number. When the user ID is a mobile phone number, then in step S210, the mobile phone number and the voice information of the user are sent to the network side equipment, so that the network side equipment obtains the mobile phone number in step S211; in step S240, the mobile phone number and The feedback information is sent to the network-side device together, so that the network-side device stores the voice training result obtained in step S240 into the user's voice training information database according to the mobile phone number. As another option, the user ID can also be an ID previously registered by the user on the network side device. In this case, before step S210, the network side device will receive the user's login information, that is to say, the user is using Before the voice input method, it is necessary to log in to the network-side device through the mobile phone, and then in step S211, the network-side device can obtain the user ID of the user according to the user's login information.
此外，根据本申请的另一个实施方式，在步骤S120和/或步骤S220中，网络侧设备还生成用于唯一标识在步骤S110和/或步骤S210中接收到的语音信息的标识符，并将该标识符与步骤S120和/或步骤S220中识别出的文字信息一起发送至终端。然后在步骤S140和/或步骤S240中，终端将该标识符与反馈信息一起发送至网络侧设备，以使步骤S110和/或步骤S210中接收到的语音信息与步骤S140和/或步骤S240中接收到的反馈信息一一对应。并且在步骤S140和/或步骤S240中，网络侧设备还根据接收到的标识符提取在步骤S110和/或步骤S210中接收到的语音信息，以便根据该反馈信息对语音信息进行语音训练。In addition, according to another embodiment of the present application, in step S120 and/or step S220, the network side device also generates an identifier for uniquely identifying the voice information received in step S110 and/or step S210, and The identifier is sent to the terminal together with the text information identified in step S120 and/or step S220. Then in step S140 and/or step S240, the terminal sends the identifier together with the feedback information to the network side device, so that the voice information received in step S110 and/or step S210 is consistent with that in step S140 and/or step S240 The received feedback information corresponds one by one. And in step S140 and/or step S240, the network side device also extracts the speech information received in step S110 and/or step S210 according to the received identifier, so as to perform speech training on the speech information according to the feedback information.
作为一种选择，网络侧设备可以根据预先设定的阈值删除接收到的语音信息。例如，当网络侧设备由于网络原因没有收到终端上传的反馈信息时，会根据预先设定的时间阈值定时删除接收到的语音信息。As an option, the network side device may delete the received voice information according to a preset threshold. For example, when the network side device does not receive the feedback information uploaded by the terminal due to network reasons, it will periodically delete the received voice information according to a preset time threshold.
此外，语音训练信息数据库可以根据预先设定的阈值停止更新。例如，当网络侧设备根据用户的反馈信息，统计得出针对该用户的语音识别的准确性已经达到预设的阈值时，该用户的语音训练信息数据库可停止更新。In addition, the speech training information database can stop updating according to a preset threshold. For example, when the network side device calculates according to the user's feedback information that the accuracy of the user's speech recognition has reached a preset threshold, the user's speech training information database may stop updating.
根据本申请另一个实施方式，当语音训练信息数据库中没有该用户的语音训练信息时，网络侧设备可提取语音训练信息数据库中的通用语音训练信息对接收到的语音信息进行语音识别，以便进行后续的程序。在完成语音训练后，针对该用户创建新的数据库。According to another embodiment of the present application, when there is no voice training information of the user in the voice training information database, the network side device can extract the general voice training information in the voice training information database to perform voice recognition on the received voice information, so as to perform follow-up procedure. After completing the voice training, create a new database for this user.
图3示出了本申请一个实施方式的用于对语音输入法进行语音训练的系统100。如图3所示，系统100包括语音信息提取模块11、语音识别模块12、文字信息反馈模块13和语音训练模块14。在一个实施方式中，语音提取模块11和文字信息反馈模块13可以设置在用户侧设备中，语音识别模块12和语音训练模块14可以设置在网络侧设备中。其中，语音信息提取模块11在终端提取用户实时输入的语音信息，并将所提取的语音信息发送给语音识别模块12。例如语音信息提取模块11提取用户向手机实时输入的语音信息，并将该语音信息发送给语音识别模块12。语音识别模块12接收语音信息，对该语音信息进行语音识别，并将识别出的文字信息发送给文字信息反馈模块13。例如语音识别模块12将对接收到的语音信息识别出的文字信息返回手机，供该用户使用。例如，如果文字信息反馈模块13接收到的文字信息与语音信息提取模块11提取的语音信息一致，则将该文字信息作为反馈信息发送至语音训练模块14；如果不一致，则根据用户指令对接收到的文字信息进行修改，并将修改后文字信息作为反馈信息发送至语音训练模块14。语音训练模块14接收反馈信息，并根据反馈信息对语音识别模块12接收的语音信息进行语音训练。Fig. 3 shows a
图4示出了本申请另一个实施方式的用于对语音输入法进行语音训练的系统200。如图所示，系统200除了包括语音信息提取模块21、语音识别模块22、文字信息反馈模块23和语音训练模块24之外，还包括ID获取模块25、语音训练信息数据库26以及标识符生成模块27。其中，ID获取模块25获取用户的用户ID。语音训练信息数据库26，用于存储语音训练信息。语音训练模块24还用于根据ID获取模块25获取的用户ID从语音训练信息数据库26中提取与该用户对应的语音训练信息。作为一种选择，语音训练信息数据库26可以为每个用户设立对应的数据库，用于存储每个用户的语音训练信息。语音识别模块22根据所提取的语音训练信息对接收到的语音信息进行语音识别。标识符生成模块27生成语音用于唯一标识该语音信息的标识符，并将该标识符发送给文字信息反馈模块23。作为一种选择，语音识别模块22识别出的文字信息与标识符生成模块27生成的标识符同时发送给文字信息反馈模块23。文字信息反馈模块23将反馈信息与接收到的标识符一起发送至语音训练模块24，以使语音信息与反馈信息一一对应。语音训练模块24根据接收到的标识符提取语音识别模块22接收到的语音信息，并根据该反馈信息对语音信息进行语音训练。语音训练信息数据库26根据语音训练模块24得到的结果进行更新。FIG. 4 shows a system 200 for performing speech training on a speech input method according to another embodiment of the present application. As shown in the figure, the system 200 includes an ID acquisition module 25, a speech training information database 26 and an identifier generation module in addition to a speech information extraction module 21, a speech recognition module 22, a text information feedback module 23 and a speech training module 24. 27. Wherein, the ID obtaining module 25 obtains the user ID of the user. The speech training information database 26 is used for storing speech training information. The speech training module 24 is also used for extracting the speech training information corresponding to the user from the speech training information database 26 according to the user ID obtained by the ID obtaining module 25 . As an option, the speech training information database 26 may set up a corresponding database for each user, for storing the speech training information of each user. The voice recognition module 22 performs voice recognition on the received voice information according to the extracted voice training information. The identifier generation module 27 generates an identifier used by voice to uniquely identify the voice information, and sends the identifier to the text information feedback module 23 . As an option, the text information recognized by the speech recognition module 22 and the identifier generated by the identifier generation module 27 are sent to the text information feedback module 23 at the same time. The text information feedback module 23 sends the feedback information together with the received identifier to the speech training module 24, so that the speech information corresponds to the feedback information one by one. The speech training module 24 extracts the speech information received by the speech recognition module 22 according to the received identifier, and performs speech training on the speech information according to the feedback information. The speech training information database 26 is updated according to the results obtained by the speech training module 24 .
作为一种选择，用户ID可以是手机号。当用户ID是手机号时，语音信息提取模块21将用户实时输入的语音信息与手机号码一起发送至语音识别模块22，以便ID获取模块25获取该手机号码。并且反馈信息模块23还可将该手机号码与反馈信息一起发送至语音训练模块24，以便语音训练模块24根据该手机号将得到的语音训练结果存入与该用户对应的语音训练信息数据库26中。作为另一种选择，用户ID还可以是用户预先在网络侧设备注册的ID，在这种情况下，语音识别模块22在接收语音信息的同时或之前，会接收到用户的登录信息，也就是说用户在使用该手机发送语音信息之前需要先通过该手机登录到网络侧设备。ID获取模块25根据用户的登录信息即可获取该用户的用户ID。As an option, the user ID can be a mobile phone number. When the user ID is a mobile phone number, the voice information extraction module 21 sends the voice information input by the user in real time together with the mobile phone number to the voice recognition module 22, so that the ID acquisition module 25 obtains the mobile phone number. And the feedback information module 23 can also send the mobile phone number and the feedback information to the voice training module 24, so that the voice training module 24 stores the voice training result obtained according to the mobile phone number in the voice training information database 26 corresponding to the user . As another option, the user ID can also be the ID registered by the user on the network side device in advance. In this case, the voice recognition module 22 will receive the user's login information at the same time or before receiving the voice information, that is, It is said that the user needs to log in to the network side device through the mobile phone before using the mobile phone to send voice messages. The ID acquiring module 25 can acquire the user ID of the user according to the user's login information.
在本申请的一个实施方式中，上述的语音提取模块21和文字信息反馈模块可以设置在用户侧设备中，语音识别模块22、语音训练模块24、ID获取模块25、语音训练信息数据库26以及标识符生成模块27可以设置在网络侧设备。In one embodiment of the present application, the above-mentioned voice extraction module 21 and the text information feedback module can be arranged in the user side equipment, the voice recognition module 22, the voice training module 24, the ID acquisition module 25, the voice training information database 26 and the identification The symbol generation module 27 can be set on the network side device.
图5示出了本申请一个实施方式的用于对语音输入法进行语音训练的系统300。如图5所示，系统300包括用户侧设备31和网络侧设备32。其中，用户侧设备31提取用户实时输入的语音信息，并将所提取的语音信息发送给网络侧设备32。例如手机提取用户实时输入的语音信息并将该语音信息发送给网络侧设备32。网络侧设备32接收语音信息，对该语音信息进行语音识别，并将识别出的文字信息发送给用户侧设备31。例如，如果用户侧设备31接收的文字信息与语音信息一致，则将该文字信息作为反馈信息发送至网络侧设备32；如果不一致，则根据用户指令对接收到的文字信息进行修改，并将修改后的文字信息作为反馈信息发送至网络侧设备32。网络侧设备32接收反馈信息，并根据反馈信息对语音识别模块12接收的语音信息进行语音训练。FIG. 5 shows a
图6示出了本申请另一个实施方式的用于对语音输入法进行语音训练的系统400。如图所示，系统200除了包括用户侧设备41和网络侧设备42之外，还包括语音训练信息数据库43，用于存储语音训练信息。其中，网络侧设备42还获取用户的用户ID，根据用户ID从语音训练信息数据库43中提取与该用户对应的语音训练信息。网络侧设备42根据所提取的语音训练信息对接收到的语音信息进行语音识别。网络侧设备42还生成唯一标识语音信息的标识符，并将该标识符与识别出的文字信息一起发送给用户侧设备41。用户侧设备41将反馈信息与接收到的标识符一起回送至网络侧设备42，以使语音信息与反馈信息一一对应。网络侧设备42根据接收到的标识符提取接收到的语音信息，并根据该反馈信息对所提取的语音信息进行语音训练。语音训练信息数据库43还可以根据网络侧设备42得到的结果进行更新。作为一种选择，语音训练信息数据库43可以存储在网络侧设备42中。FIG. 6 shows a
作为一种选择，用户ID可以是手机号。当用户ID是手机号时，用户侧设备41将用户的语音信息与手机号码一起发送至网络侧设备42。而且，用户侧设备41还可以将该手机号码与反馈信息一起回送至网络侧设备42，以便网络侧设备42根据该手机号将得到的语音训练结果存入与该用户对应的语音训练信息数据库43中。作为另一种选择，用户ID还可以是用户预先在网络侧设备注册的ID，在这种情况下，用户侧设备41需要预先登录网络侧设备42，而网络侧设备42则根据用户的登录信息即可获取该用户的用户ID。As an option, the user ID can be a mobile phone number. When the user ID is a mobile phone number, the user-
在本申请的一个实施方式中，上述网络侧设备42可包括语音识别模块421以及语音训练模块422。其中，语音识别模块421从用户侧设备41接收用户实时输入的语音信息，对接收到的语音信息进行语音识别，并将识别出的文字信息发送给用户侧设备41。语音训练模块422从用户侧设备41接收反馈信息，并根据反馈信息对语音信息进行语音训练。其中，通常反馈信息与语音信息的内容是一致的。In an embodiment of the present application, the
在本申请的另一个实施方式中，上述网络侧设备42还可包括ID获取模块423以及标识符生成模块424。其中，ID获取模块423获取用户的用户ID。语音训练模块422还用于根据ID获取模块423获取的用户ID从语音训练信息数据库43中提取与该用户对应的语音训练信息。作为一种选择，语音训练信息数据库43可以设置在网络侧设备42中，并且可以为每个用户设立对应的数据库用于存储每个用户的语音训练信息。语音识别模块422根据所提取的语音训练信息对接收到的语音信息进行语音识别，并将得到的文字信息发送给用户侧设备41。In another embodiment of the present application, the
作为一种选择，上述网络侧设备200还可包括标识符生成模块424，用于生成标识符，该标识符唯一地标识语音识别模块421接收到的语音信息。语音识别模块421将该标识符与识别出的文字信息一起发送至用户侧设备41。语音训练模块422从用户侧设备41接收与反馈信息一起回送的该标识符，以使语音识别模块421接收到的语音信息与反馈信息一一对应。语音训练模块422根据接收到的标识符提取语音信息，以便根据该反馈信息对所提取的语音信息进行语音训练。As an option, the network-side device 200 may further include an
作为一种选择，网络侧设备42还可以根据预先设定的阈值删除接收到的语音信息。例如，当网络侧设备42由于网络故障等原因没有收到终端上传的反馈信息时，网络侧设备42会根据预先设定的时间阈值定时删除接收到的语音信息。As an option, the
此外，语音训练信息数据库43可以根据预先设定的阈值停止更新语音训练信息数据库。例如，当语音训练信息数据库43根据用户的反馈信息，统计得出针对该用户的语音识别的准确性已经达到预设的阈值时，则停止更新该用户的语音训练信息数据库。In addition, the speech
根据本申请另一个实施方式，语音训练信息数据库43包括第一数据库表，用来保存语音信息和文字信息的匹配关系。第一数据库表可以包括三个字段：TransID，用来保存标识符生成模块424生成的标识符；VoiceInfo，用来保存语音识别模块421接收到的语音数据；Time，表示该语音信息采集的时间。According to another embodiment of the present application, the voice
网络侧设备42接收到用户侧设备41上传的语音信息并进行语音识别后，将该语音数据存入上述数据表中。After receiving the voice information uploaded by the user-
当网络侧设备42接收到终端上传的反馈信息时，语音训练模块422根据标识符从语音训练信息数据库43中提取语音信息，然后对该语音信息进行语音训练，对该语音信息的语音训练完成后可以删除该语音信息。When the
如果由于各种原因，语音识别模块421没有收到用户侧设备41的反馈信息，网络侧设备42会定时检查该数据表中的语音数据是否超时，例如通过Time字段判定该数据表中的语音数据是否超时。如果超时，则删除该条记录。If due to various reasons, the
语音训练信息数据库43还可以包括第二数据库表，第二数据库表可以包括两个字段，例如UserID和UserInfo。UserID用来存储用户唯一ID；UserInfo用来保存用户信息，在语音特征提取时作为辅助信息。The speech
根据本申请另一个实施方式，当语音训练信息数据库43中没有该用户的语音训练信息时，语音识别模块421可以提取语音训练信息数据库43中的通用语音训练信息对接收到的语音信息进行语音识别，以便进行后续的程序。在完成语音训练后，可以针对该用户创建新的数据库。According to another embodiment of the present application, when there is no speech training information of the user in the speech
以上仅为本申请的示例性实施方式，本领域技术人员根据上述实施方式，在本申请权利要求限定的范围内，可以对上述各个实施方式进行修改。The above are only exemplary implementations of the present application, and those skilled in the art may modify the above implementations within the scope defined by the claims of the application based on the above implementations.
Claims (14)
Translated fromChinesePriority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100004314ACN102543076A (en) | 2011-01-04 | 2011-01-04 | Speech training method and corresponding system for phonetic entry method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100004314ACN102543076A (en) | 2011-01-04 | 2011-01-04 | Speech training method and corresponding system for phonetic entry method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102543076Atrue CN102543076A (en) | 2012-07-04 |
Family
ID=46349816
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100004314APendingCN102543076A (en) | 2011-01-04 | 2011-01-04 | Speech training method and corresponding system for phonetic entry method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102543076A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103093316A (en)* | 2013-01-24 | 2013-05-08 | 广东欧珀移动通信有限公司 | Method and device of bill generation |
| CN104123930A (en)* | 2013-04-27 | 2014-10-29 | 华为技术有限公司 | Guttural identification method and device |
| CN104538034A (en)* | 2014-12-31 | 2015-04-22 | 深圳雷柏科技股份有限公司 | Voice recognition method and system |
| CN104732975A (en)* | 2013-12-20 | 2015-06-24 | 华为技术有限公司 | Method and device for voice instant messaging |
| CN104751844A (en)* | 2015-03-12 | 2015-07-01 | 深圳市富途网络科技有限公司 | Voice identification method and system used for security information interaction |
| CN107895572A (en)* | 2016-09-30 | 2018-04-10 | 法乐第(北京)网络科技有限公司 | A kind of speech recognition training method and system |
| CN108682416A (en)* | 2018-04-11 | 2018-10-19 | 深圳市卓翼科技股份有限公司 | local adaptive voice training method and system |
| CN109215638A (en)* | 2018-10-19 | 2019-01-15 | 珠海格力电器股份有限公司 | Voice learning method and device, voice equipment and storage medium |
| US10984795B2 (en) | 2018-04-12 | 2021-04-20 | Samsung Electronics Co., Ltd. | Electronic apparatus and operation method thereof |
| WO2021134551A1 (en)* | 2019-12-31 | 2021-07-08 | 李庆远 | Human merging and training of multiple machine translation outputs |
| CN113129867A (en)* | 2019-12-28 | 2021-07-16 | 中移(上海)信息通信科技有限公司 | Training method of voice recognition model, voice recognition method, device and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1703923A (en)* | 2002-10-18 | 2005-11-30 | 中国科学院声学研究所 | Portable digital mobile communication apparatus and voice control method and system thereof |
| CN101170757A (en)* | 2006-10-26 | 2008-04-30 | 英华达(上海)电子有限公司 | A method and device for controlling text input in mobile device |
| US7440894B2 (en)* | 2005-08-09 | 2008-10-21 | International Business Machines Corporation | Method and system for creation of voice training profiles with multiple methods with uniform server mechanism using heterogeneous devices |
| CN101923856A (en)* | 2009-06-12 | 2010-12-22 | 华为技术有限公司 | Speech recognition training processing, control method and device |
- 2011
- 2011-01-04CNCN2011100004314Apatent/CN102543076A/enactivePending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1703923A (en)* | 2002-10-18 | 2005-11-30 | 中国科学院声学研究所 | Portable digital mobile communication apparatus and voice control method and system thereof |
| US7440894B2 (en)* | 2005-08-09 | 2008-10-21 | International Business Machines Corporation | Method and system for creation of voice training profiles with multiple methods with uniform server mechanism using heterogeneous devices |
| CN101170757A (en)* | 2006-10-26 | 2008-04-30 | 英华达(上海)电子有限公司 | A method and device for controlling text input in mobile device |
| CN101923856A (en)* | 2009-06-12 | 2010-12-22 | 华为技术有限公司 | Speech recognition training processing, control method and device |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103093316B (en)* | 2013-01-24 | 2016-03-23 | 广东欧珀移动通信有限公司 | A kind of bill generation method and device |
| CN103093316A (en)* | 2013-01-24 | 2013-05-08 | 广东欧珀移动通信有限公司 | Method and device of bill generation |
| CN104123930A (en)* | 2013-04-27 | 2014-10-29 | 华为技术有限公司 | Guttural identification method and device |
| CN104732975A (en)* | 2013-12-20 | 2015-06-24 | 华为技术有限公司 | Method and device for voice instant messaging |
| CN104538034B (en)* | 2014-12-31 | 2018-08-28 | 深圳雷柏科技股份有限公司 | A kind of audio recognition method and system |
| CN104538034A (en)* | 2014-12-31 | 2015-04-22 | 深圳雷柏科技股份有限公司 | Voice recognition method and system |
| CN104751844A (en)* | 2015-03-12 | 2015-07-01 | 深圳市富途网络科技有限公司 | Voice identification method and system used for security information interaction |
| CN107895572A (en)* | 2016-09-30 | 2018-04-10 | 法乐第(北京)网络科技有限公司 | A kind of speech recognition training method and system |
| CN108682416A (en)* | 2018-04-11 | 2018-10-19 | 深圳市卓翼科技股份有限公司 | local adaptive voice training method and system |
| CN108682416B (en)* | 2018-04-11 | 2021-01-01 | 深圳市卓翼科技股份有限公司 | Local adaptive speech training method and system |
| US10984795B2 (en) | 2018-04-12 | 2021-04-20 | Samsung Electronics Co., Ltd. | Electronic apparatus and operation method thereof |
| CN109215638A (en)* | 2018-10-19 | 2019-01-15 | 珠海格力电器股份有限公司 | Voice learning method and device, voice equipment and storage medium |
| CN113129867A (en)* | 2019-12-28 | 2021-07-16 | 中移(上海)信息通信科技有限公司 | Training method of voice recognition model, voice recognition method, device and equipment |
| CN113129867B (en)* | 2019-12-28 | 2024-05-24 | 中移(上海)信息通信科技有限公司 | Training method of voice recognition model, voice recognition method, device and equipment |
| WO2021134551A1 (en)* | 2019-12-31 | 2021-07-08 | 李庆远 | Human merging and training of multiple machine translation outputs |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102543076A (en) | Speech training method and corresponding system for phonetic entry method | |
| US10270736B2 (en) | Account adding method, terminal, server, and computer storage medium | |
| US10593334B2 (en) | Method and apparatus for generating voiceprint information comprised of reference pieces each used for authentication | |
| CN105895103B (en) | Voice recognition method and device | |
| KR101768509B1 (en) | On-line voice translation method and device | |
| CN103187053B (en) | Input method and electronic equipment | |
| WO2016101577A1 (en) | Voice recognition method, client and terminal device | |
| CN104185868A (en) | Voice authentication and speech recognition system and method | |
| CN103413549A (en) | Voice interaction method and system and interaction terminal | |
| CN102074231A (en) | Speech recognition method and speech recognition system | |
| CN111883140A (en) | Authentication method, device, equipment and medium based on knowledge graph and voiceprint recognition | |
| CN104219195A (en) | Identity verifying method, device and system | |
| CN106713111B (en) | Processing method for adding friends, terminal and server | |
| CN103841272B (en) | A kind of method and device sending speech message | |
| CN104538034A (en) | Voice recognition method and system | |
| JP2017003611A (en) | Voice recognition device, voice recognition system, terminal used in voice recognition system and method for generating speaker identification model | |
| CN108447489B (en) | A continuous voiceprint authentication method and system with feedback | |
| CN118284930A (en) | Method and system for device feature analysis to improve user experience | |
| TW200304638A (en) | Network-accessible speaker-dependent voice models of multiple persons | |
| CN109003600B (en) | Message processing method and device | |
| CN113782026A (en) | An information processing method, apparatus, medium and equipment | |
| CN106782567A (en) | Method and device for establishing voiceprint model | |
| CN116825097A (en) | Vehicle-mounted voice recognition method and device | |
| EP2913822A1 (en) | Speaker recognition method | |
| CN111933157A (en) | Voiceprint recognition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20120704 |