WO2019210567A1 - 推送消息播报方法、可读存储介质、终端设备及装置 - Google Patents

推送消息播报方法、可读存储介质、终端设备及装置 Download PDF

Info

Publication number
WO2019210567A1
WO2019210567A1 PCT/CN2018/093708 CN2018093708W WO2019210567A1 WO 2019210567 A1 WO2019210567 A1 WO 2019210567A1 CN 2018093708 W CN2018093708 W CN 2018093708W WO 2019210567 A1 WO2019210567 A1 WO 2019210567A1
Authority
WO
WIPO (PCT)
Prior art keywords
message
feature vector
voiceprint feature
voice information
voice
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.)
Ceased
Application number
PCT/CN2018/093708
Other languages
English (en)
French (fr)
Inventor
姜翔
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Publication of WO2019210567A1 publication Critical patent/WO2019210567A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/55Push-based network services
    • GPHYSICS
    • G08SIGNALLING
    • G08BSIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B3/00Audible signalling systems, e.g. audible personal calling systems
    • G08B3/10Audible signalling systems, e.g. audible personal calling systems using electric transmission; using electromagnetic transmission
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L17/00Speaker identification or verification techniques
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination

Definitions

  • the present application belongs to the field of computer technology, and in particular, to a push message broadcast method, a computer readable storage medium, a terminal device and a device.
  • the commonly used push messages are usually text messages, and some important real-time messages, such as stock account trading information, when the user cannot view the mobile phone, the real-time performance of the message is not guaranteed, which may cause economic loss to the user.
  • some manufacturers have adopted the practice of converting the push message for voice conversion and broadcasting the push message in the form of voice, but this approach may cause other people around to hear the push message, and the privacy is poor. Causes the disclosure of user privacy.
  • the embodiment of the present application provides a push message broadcast method, a computer readable storage medium, a terminal device and a device, so as to solve the problem that the existing push message broadcast method has poor privacy and is easy to cause leakage of user privacy.
  • a first aspect of the embodiment of the present application provides a push message broadcast method, which may include:
  • the reference voiceprint feature vector is a voiceprint feature vector extracted from a voice of a specified user
  • the push message is broadcasted in a voice form.
  • a second aspect of embodiments of the present application provides a computer readable storage medium storing computer readable instructions that, when executed by a processor, implement the push message broadcast method step.
  • a third aspect of an embodiment of the present application provides a push message broadcast terminal device including a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, the processor executing The computer readable instructions implement the steps of the push message broadcast method described above.
  • a fourth aspect of the embodiments of the present application provides a push message broadcast apparatus, which may include a module for implementing the steps of the push message broadcast method.
  • the embodiment of the present application has the beneficial effects that: when receiving the push message, the embodiment of the present application uses a message prompt tone to remind the user that if the user believes that there is no danger of privacy leakage in the current environment, And issuing a preset instruction for instructing to broadcast the push message, if the instruction is verified to be correct, further determining whether the issued instruction is a designated user by voiceprint matching, thereby greatly improving the privacy of the push message broadcast and reducing user privacy. The chance of leaking.
  • FIG. 1 is a flowchart of an embodiment of a method for pushing a message broadcast according to an embodiment of the present application
  • 2 is a schematic diagram of a message queue
  • 3 is a schematic diagram of a read message queue and an unread message queue
  • FIG. 5 is a structural diagram of an embodiment of a push message broadcast apparatus according to an embodiment of the present application.
  • FIG. 6 is a schematic block diagram of a push message broadcast terminal device according to an embodiment of the present application.
  • an embodiment of a push message broadcast method in the embodiment of the present application may include:
  • Step S101 Receive a push message sent by the server push server.
  • the push message may be stored in the terminal device in the form of a queue.
  • a message queue R can be configured to store the push message received by the terminal device, and multiple messages in the message queue can be sequentially stored in the order of receiving time, as shown in FIG. 2 .
  • the push message recently received by the terminal device that is, the message P, can be stored to the end of the queue of the message queue R.
  • the message queue R can be further divided into: a read message queue and an unread message queue, wherein the messages stored in the unread message queue are unread messages, that is, are not broadcasted.
  • the message stored in the read message queue is a message that has already been reported.
  • the broadcast message may be assigned a broadcast label, and the initial value of the broadcast label is: unread, and the push message is stored in the tail of the unread message queue;
  • a message in the unread message queue can update the broadcast label of the broadcasted message to read: and store the broadcasted message to the end of the queue of the read message queue. For example, when the terminal device broadcasts the unread message 0 in the unread message queue, the unread message 0 can be marked as read and stored to the end of the queue of the read message queue.
  • each message queue may be further divided into a read message queue and an unread message queue.
  • the terminal device After receiving the push message sent by the message push server, the terminal device sequentially matches the push message with the keyword in the preset keyword set, and determines the message type corresponding to the successfully matched keyword as the Pushing the message type of the message, and then storing the push message into an unread message queue corresponding to the message type of the push message.
  • the keyword set includes respective subsets of keywords corresponding to preset various message types, and each keyword subset includes more than one keyword.
  • the push message may be classified into a message type such as an advertisement class, a news information class, and a system notification class, and each type of push information is set with a corresponding keyword subset.
  • the keyword subset of the advertisement category may include keywords such as "advertisement”, “promotion”, “offer”, and "price reduction”.
  • priority can also be set for each message queue, for example, the priority of the system notification can be set to the highest, the priority of the news information is second, and the priority of the advertisement is the lowest.
  • the message messages of different message types are classified and stored through multiple message queues, which facilitates the classification management and broadcast of push messages, and further improves the flexibility of message broadcast control.
  • the preset message prompt tone may also be broadcasted.
  • the audio content of the prompt tone may be a default fixed configuration, or may be set by the user according to the preference of the user.
  • the prompt tone is used to indicate that the terminal device has received the push message. If the user hears the prompt tone and wants to listen to the push message in the form of voice, the user may trigger an instruction to broadcast the message by voice, and the terminal device may broadcast the report according to the instruction.
  • the push message is described, thereby ensuring that the user does not miss the push message when conveniently listening to the message.
  • Step S102 Collect voice information in a preset time period.
  • the time period may be set according to actual conditions. Generally, it may be set to be within half a minute or within 1 minute after the message prompt tone is broadcast.
  • the voice information in the external environment of the terminal device in the time period is collected by the microphone of the terminal device.
  • Step S103 Determine whether the voice information includes a preset instruction for indicating that the push message is broadcasted.
  • the preset instruction for indicating the broadcast of the push message may be set according to actual conditions, for example, it may be set as a voice "broadcast message”, “broadcast”, “voice message”, and the like.
  • step S104 If the voice information does not include a preset instruction for indicating that the push message is broadcasted, step S104 is performed, and if the voice information includes a preset instruction for indicating that the push message is broadcasted, step S105 is performed. And its subsequent steps.
  • Step S104 Determine not to broadcast the push message.
  • Step S105 Extract a voiceprint feature vector in the voice information.
  • step S105 may include a process as shown in FIG. 4:
  • Step S1051 The voice information is divided into M voice sub-segments.
  • M is an integer greater than one.
  • Step S1052 Calculate a Meyer spectrum cepstral coefficient vector of each speech sub-segment.
  • the Meer spectral scrambling coefficient vector of each speech sub-segment can be calculated according to the following formula:
  • n is The sequence number of the frequency spectrum scrambling coefficient of the Mel, 1 ⁇ n ⁇ N, N is the preset dimension of the frequency spectrum coefficient of the Mel spectrum, and MelCoe m,n is the nth Meier spectrum frequency-factoring coefficient of the mth voice sub-segment .
  • Step S1053 Calculate a weight coefficient of each voice sub-segment.
  • the weight coefficients of the respective voice sub-segments can be calculated according to the following formula:
  • Weight m is the weight coefficient of the mth voice subsection.
  • Step S1054 Constructing a voiceprint feature vector in the voice information.
  • the voiceprint feature vector in the voice information may be constructed according to the following formula:
  • VoPrintVec (VpElem 1 , VpElem 2 , ..., VpElem n , ..., VpElem N )
  • VpElem n is the nth element of the voiceprint feature vector in the voice information
  • VoPrintVec is the voiceprint feature vector in the voice information.
  • Step S106 Calculate a matching degree between the voiceprint feature vector in the voice information and a preset reference voiceprint feature vector.
  • the reference voiceprint feature vector is a voiceprint feature vector extracted from the voice of the specified user (for example, the owner of the terminal device), and the specific calculation process is similar to step S105, and details are not described herein again.
  • the degree of matching between the voiceprint feature vector in the voice information and the reference voiceprint feature vector may be calculated according to the following formula:
  • StVpElem n is the nth element of the reference voiceprint feature vector
  • MatDeg is a degree of matching between the voiceprint feature vector in the voice information and the reference voiceprint feature vector.
  • Step S107 Determine whether a matching degree between the voiceprint feature vector and the reference voiceprint feature vector in the voice information is greater than a preset matching degree threshold.
  • the matching degree threshold may be set according to actual conditions, for example, it may be set to 70%, 80% or 90%, and the like.
  • Step S104 If the matching degree between the voiceprint feature vector and the reference voiceprint feature vector in the voice information is less than or equal to the matching degree threshold, it is indicated that the designated user is not the specified user, and should be executed at this time. Step S104.
  • step S108 If the matching degree between the voiceprint feature vector and the reference voiceprint feature vector in the voice information is greater than the matching degree threshold, it is indicated that the designated user is the announcement command, and step S108 is performed.
  • Step S108 Broadcast the push message in a voice form.
  • the unread message queue with the highest priority and the queue is not empty is selected as the current message queue, the queue header message is broadcasted in the voice form, and the queue header message is deleted from the current message queue.
  • the queue header message is a push message arranged at a head position of the current message queue, and then determining whether the current message queue is empty. If the current message queue is not empty, returning to perform the voice form The step of broadcasting the head start message until the current message queue is empty, and if the current message queue is empty, returning to the unread message queue that performs the selection preset with the highest priority and the queue is not empty as the current The steps of the message queue until each unread message queue is empty.
  • other instructions may be set, for example, an instruction for indicating the previous message, an instruction for indicating the next message, An instruction for indicating a replay message, an instruction for instructing to stop the broadcast message, and the like.
  • the terminal device When the terminal device receives an instruction for indicating that the previous message is broadcast, if the terminal device is currently broadcasting an unread message in a certain message queue, the terminal device may then broadcast the last read read message in the message queue; If the read message in a certain message queue is currently being broadcast, the terminal device can then broadcast the read message in the message queue that is before the read message currently being broadcast.
  • the terminal device can acquire and broadcast the read message P in the read message queue when receiving the instruction for indicating the previous message. -1.
  • the terminal device may The broadcast control command broadcasts the read message P-2 in the read message queue.
  • the terminal device may broadcast a message to be broadcasted after the message currently being broadcasted in the message queue currently being broadcast.
  • the terminal device when the terminal device receives the broadcast instruction for indicating the next message, it can obtain the unread in the unread message queue. Message P+1, and broadcast. If the terminal device is currently broadcasting the read message 0 of the read message queue in the message queue R, when the terminal device receives the broadcast instruction for indicating the next message, the read message in the read message queue can be obtained. 1, and broadcast.
  • the terminal device may repeatedly broadcast the message that was last broadcasted in the message queue currently being broadcast.
  • the terminal device can repeatedly broadcast the read message. Read message P-1 of the queue.
  • the terminal device may stop the broadcast and clear the push message buffered in the terminal device.
  • the terminal device receives an instruction for indicating the broadcast message, the terminal does not store the message in the unread message queue, or the terminal device receives the instruction for indicating the previous message, the terminal The device has already broadcasted to the head of the message queue; or, when the terminal device receives the broadcast control command for indicating the next message, the terminal device has already broadcasted to the tail of the message queue, and the terminal device can return the prompt information,
  • the prompt message may be a voice message "no more messages", which is used to prompt the user that there is no message available for broadcast.
  • the message prompt tone is used to remind the user. If the user believes that there is no danger of privacy leakage in the current environment, the user may be instructed to indicate the pre-posting of the push message. If the instruction is verified, the voiceprint matching is further used to determine whether the instruction is a designated user, thereby greatly improving the privacy of the push message broadcast and reducing the probability of user privacy leakage.
  • FIG. 5 is a structural diagram of an embodiment of a push message broadcast apparatus provided by an embodiment of the present application.
  • a push message broadcast apparatus may include:
  • the push message receiving module 501 is configured to receive a push message sent by the message push server;
  • the voice information collecting module 502 is configured to collect voice information in a preset time period
  • the command determining module 503 is configured to determine whether the voice information includes a preset instruction for indicating that the push message is broadcasted;
  • the voiceprint feature vector extraction module 504 is configured to extract a voiceprint feature vector in the voice information if the voice information includes a preset instruction for indicating that the push message is broadcasted;
  • the matching degree calculation module 505 is configured to calculate a matching degree between the voiceprint feature vector in the voice information and a preset reference voiceprint feature vector, where the reference voiceprint feature vector is extracted from the voice of the specified user.
  • Voiceprint feature vector ;
  • the push message broadcast module 506 is configured to broadcast the push message in a voice form if the degree of matching between the voiceprint feature vector and the reference voiceprint feature vector in the voice information is greater than a preset match degree threshold.
  • the voiceprint feature vector extraction module may include:
  • a voice sub-segment dividing unit configured to divide the voice information into M voice sub-segments, where M is an integer greater than one;
  • a Meyer spectrum cepstral coefficient vector calculation unit is configured to calculate a Meyer spectral cepstral coefficient vector of each speech sub-segment according to the following formula:
  • n is The sequence number of the frequency spectrum scrambling coefficient of the Mel, 1 ⁇ n ⁇ N, N is the preset dimension of the frequency spectrum coefficient of the Mel spectrum, and MelCoe m,n is the nth Meier spectrum frequency-factoring coefficient of the mth voice sub-segment ;
  • a weight coefficient calculation unit is configured to calculate a weight coefficient of each voice sub-segment according to the following formula:
  • Weight m is the weight coefficient of the mth voice subsection
  • a voiceprint feature vector construction unit configured to construct a voiceprint feature vector in the voice information according to the following formula:
  • VoPrintVec (VpElem 1 , VpElem 2 , ..., VpElem n , ..., VpElem N )
  • VpElem n is the nth element of the voiceprint feature vector in the voice information
  • VoPrintVec is the voiceprint feature vector in the voice information.
  • the matching degree calculation module may include:
  • a matching degree calculation unit configured to calculate a matching degree between the voiceprint feature vector in the voice information and the reference voiceprint feature vector according to the following formula:
  • StVpElem n is the nth element of the reference voiceprint feature vector
  • MatDeg is a degree of matching between the voiceprint feature vector in the voice information and the reference voiceprint feature vector.
  • the push message broadcast apparatus may further include:
  • a keyword matching module configured to sequentially match the push message with a keyword in a preset keyword set, and determine a message type corresponding to the successfully matched keyword as a message type of the push message, where
  • the keyword set includes a subset of each keyword corresponding to each of the preset various message types, and each keyword subset includes more than one keyword;
  • a push message storage module configured to store the push message into an unread message queue corresponding to the message type of the push message.
  • the push message broadcast module may include:
  • the current message queue selection unit is configured to select an unread message queue with the highest priority and the queue is not empty as the current message queue;
  • a head-end message broadcast unit configured to broadcast a team head message in a voice form, the head-to-head message being a push message arranged at a head position of the current message queue;
  • a headend message deleting unit configured to delete the team header message from the current message queue
  • the message queue determining unit is configured to determine whether the current message queue is empty.
  • FIG. 6 is a schematic block diagram of a push message broadcast terminal device provided by an embodiment of the present application. For convenience of description, only parts related to the embodiment of the present application are shown.
  • the push message broadcast terminal device 6 may be a computing device such as a mobile phone, a tablet computer, a smart watch, a smart bracelet, and smart glasses.
  • the push message broadcast terminal device 6 may include a processor 60, a memory 61, and computer readable instructions 62 stored in the memory 61 and operable on the processor 60, such as performing the push message broadcast method described above.
  • the processor 60 performs the steps in the embodiments of the respective push message broadcast methods when the computer readable instructions 62 are executed.
  • the functional units in the various embodiments of the present application may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of computer readable instructions are included to cause a computer device to perform all or part of the steps of the methods described in various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Signal Processing (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Computational Linguistics (AREA)
  • Electromagnetism (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Telephonic Communication Services (AREA)

Abstract

本申请属于计算机技术领域,尤其涉及一种推送消息播报方法、计算机可读存储介质、终端设备及装置。所述方法接收消息推送服务器下发的推送消息;采集预设的时间段内的语音信息;判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度;若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。通过本申请实施例,大大提高了推送消息播报的私密性,减少了用户隐私泄露的几率。

Description

推送消息播报方法、可读存储介质、终端设备及装置
本申请要求于2018年5月3日提交中国专利局、申请号为201810416166.X、发明名称为“一种推送消息播报方法、计算机可读存储介质及终端设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请属于计算机技术领域,尤其涉及一种推送消息播报方法、计算机可读存储介质、终端设备及装置。
背景技术
目前常用的推送消息通常为文字消息,一些重要的实时消息,如股票账户的买卖信息,在用户无法查看手机时,消息的实时性得不到保证,可能造成用户的经济损失。为了解决这一问题,部分厂商采用了将推送消息进行语音转化,并以语音形式播报推送消息的做法,但这种做法可能会使周边的其他人也听到该推送消息,私密性差,极易造成用户隐私的泄露。
技术问题
有鉴于此,本申请实施例提供了一种推送消息播报方法、计算机可读存储介质、终端设备及装置,以解决现有的推送消息播报方法私密性差,极易造成用户隐私的泄露的问题。
技术解决方案
本申请实施例的第一方面提供了一种推送消息播报方法,可以包括:
接收消息推送服务器下发的推送消息;
采集预设的时间段内的语音信息;
判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;
计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
本申请实施例的第二方面提供了一种计算机可读存储介质,所述计算机可读存储介质存储有计算机可读指令,所述计算机可读指令被处理器执行时实现上述推送消息播报方法的步骤。
本申请实施例的第三方面提供了一种推送消息播报终端设备,包括存储器、处理 器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令,所述处理器执行所述计算机可读指令时实现上述推送消息播报方法的步骤。
本申请实施例的第四方面提供了一种推送消息播报装置,可以包括用于实现上述推送消息播报方法的步骤的模块。
有益效果
本申请实施例与现有技术相比存在的有益效果是:本申请实施例在收到推送消息时,会用消息提示音来提醒用户,若用户认为当前环境不存在隐私泄露的危险,则可以发出用于指示播报所述推送消息的预设指令,若该指令核实无误,则进一步通过声纹匹配判断发出指令的是否为指定用户,从而大大提高了推送消息播报的私密性,减少了用户隐私泄露的几率。
附图说明
图1为本申请实施例中一种推送消息播报方法的一个实施例流程图;
图2为消息队列的示意图;
图3为已读消息队列和未读消息队列的示意图;
图4为提取声纹特征向量的示意流程图;
图5为本申请实施例中一种推送消息播报装置的一个实施例结构图;
图6为本申请实施例中一种推送消息播报终端设备的示意框图。
本发明的实施方式
请参阅图1,本申请实施例中一种推送消息播报方法的一个实施例可以包括:
步骤S101、接收消息推送服务器下发的推送消息。
优选地,所述推送消息可以以队列的形式存储在终端设备中。在其中的一种具体实现中,可以配置一个消息队列R,该消息队列R用于存储终端设备接收到的推送消息,该消息队列中的多个消息可以按照接收时间顺序依次存储,如图2所示,可以将终端设备最近接收到的推送消息,也即消息P存储至消息队列R的队尾。
进一步地,如图3所示,该消息队列R还可以进一步划分为:已读消息队列和未读消息队列,其中,未读消息队列中存储的消息均为未读消息,即未被播报过的消息,已读消息队列中存储的消息为已播报过的消息。终端设备每接收到一个推送消息,可以为该推送消息分配一个播报标签,该播报标签的初始值为:未读,并将该推送消息存储至未读消息队列的队尾;终端设备每播报完一个未读消息队列中的消息,即可将该播报过的消息的播报标签更新为:已读,并将该播报过的消息存储至已读消息队列的队尾。例如,当终端设备播报完未读消息队列中的未读消息0后,可以将该未读消息0标记为已读,并存储至已读消息队列的队尾。
进一步地,还可以配置对应于消息类型的至少两个消息队列,同样的,每个消息队列也可以进一步划分为已读消息队列和未读消息队列。终端设备在接收消息推送服务器下发的推送消息之后,将所述推送消息与预设的关键词集合中的关键词依次进行匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,然后,将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词。例如,可以将推送消息分为广告类、新闻资讯类、系统通知类等消息类型,每种类型的推送信息均设置与之对应的关键词子集。例如,广告类的关键词子集中可以包括:“广告”、“促销”、“优惠”、“降价”等关键词。
优选地,还可以为各个消息队列设置优先级,例如,可以将系统通知的优先级设为最高,新闻资讯的优先级次之,广告的优先级最低。
通过多个消息队列对不同消息类型的推送消息进行分类存储,便于实现对推送消息的分类管理和播报,进一步提高了消息播报控制时的灵活性。
优选地,在接收消息推送服务器下发的推送消息之后,还可以播报预设的消息提示音。该提示音的音频内容可以为默认的固定配置,也可以由用户根据自己的喜好自主设置,本申请实施例对此不做限定。该提示音用于指示终端设备接收到了推送消息,若用户听到了该提示音,并希望以语音的形式收听该推送消息,则可以通过语音触发播报消息的指令,终端设备可以根据该指令播报所述推送消息,由此即可保证用户在方便收听消息的时候不会错过所述推送消息。
步骤S102、采集预设的时间段内的语音信息。
所述时间段可以根据实际情况进行设置,一般地,可以将其设置为播报消息提示音后的半分钟或者1分钟之内。通过终端设备的麦克风采集该时间段内终端设备的外部环境中的语音信息。
步骤S103、判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令。
用于指示播报所述推送消息的预设指令可以根据实际情况进行设置,例如,可以将其设置为语音“播报消息”、“播报”、“语音消息”等等。
若所述语音信息中不包含用于指示播报所述推送消息的预设指令,则执行步骤S104,若所述语音信息中包含用于指示播报所述推送消息的预设指令,则执行步骤S105及其后续步骤。
步骤S104、确定不对所述推送消息进行播报。
步骤S105、提取所述语音信息中的声纹特征向量。
具体地,步骤S105可以包括如图4所示的过程:
步骤S1051、将所述语音信息划分为M个语音子段。
其中,M为大于1的整数。
步骤S1052、计算各个语音子段的梅尔频谱倒频系数向量。
具体地,可以根据下式计算各个语音子段的梅尔频谱倒频系数向量:
MelVec m=MFCCFuc(SubVoice m)
其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数。
步骤S1053、计算各个语音子段的权重系数。
具体地,可以根据下式计算各个语音子段的权重系数:
Figure PCTCN2018093708-appb-000001
其中,Weight m为第m个语音子段的权重系数。
步骤S1054、构造所述语音信息中的声纹特征向量。
具体地,可以根据下式构造所述语音信息中的声纹特征向量:
VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
Figure PCTCN2018093708-appb-000002
VoPrintVec为所述语音信息中的声纹特征向量。
步骤S106、计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度。
所述基准声纹特征向量为从指定用户(例如,终端设备的主人)的语音中提取的声纹特征向量,其具体的计算过程与步骤S105类似,在此不再赘述。
优选地,可以根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
Figure PCTCN2018093708-appb-000003
其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
步骤S107、判断所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度是否大于预设的匹配度阈值。
所述匹配度阈值可以根据实际情况进行设置,例如,可以将其设置为70%、80%或者90%等等。
若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度小于或等于所述匹配度阈值,则说明发出播报指令的并不是所述指定用户,此时应执行步骤S104。
若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于所述匹配度阈值,则说明发出播报指令的就是所述指定用户,此时执行步骤S108。
步骤S108、以语音形式播报所述推送消息。
具体地,首先选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列,以语音形式播报队头消息,并将所述队头消息从所述当前消息队列中删除,所述队头消息为排列在所述当前消息队列的队头位置的推送消息,然后判断所述当前消息队列是否为空,若所述当前消息队列不为空,则返回执行所述以语音形式播报队头消息的步骤,直至所述当前消息队列为空为止,若所述当前消息队列为空,则返回执行所述选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列的步骤,直至各个未读消息队列均为空为止。
进一步地,除了上述的用于指示播报所述推送消息的预设指令之外,还可以设置其它的指令,例如,用于指示播报上一条消息的指令、用于指示播报下一条消息的指令、用于指示重播消息的指令以及用于指示停止播报消息的指令等等。
当终端设备接收到用于指示播报上一条消息的指令时,若终端设备当前正在播报某一消息队列中的未读消息,则终端设备可以接着播报该消息队列中最后存入的已读消息;若当前正在播报某一消息队列中的已读消息,则终端设备可以接着播报该消息队列中位于当前正在播报的已读消息之前的已读消息。
例如,假设终端设备当前正在播报未读消息队列中的未读消息P,则当终端设备接收到用于指示播报上一条消息的指令时,可以获取并播报已读消息队列中的已读消息P-1。或者,若终端设备接收到该用于指示播报上一条消息的指令时,没有正在播报的消息,且最近一次播报的消息为已读消息队列中的已读消息P-1,则终端设备可以根据该播报控制指令播报已读消息队列中的已读消息P-2。
当终端设备接收到用于指示播报下一条消息的指令时,终端设备可以播报当前正 在播报的消息队列中位于当前正在播报的消息之后待播报的消息。
例如,假设终端设备当前正在播报消息队列R中未读消息队列的未读消息P,则当终端设备接收到用于指示播报下一条消息的播报指令时,可以获取未读消息队列中的未读消息P+1,并进行播报。若终端设备当前正在播报消息队列R中已读消息队列的已读消息0,则当终端设备接收到用于指示播报下一条消息的播报指令时,可以获取该已读消息队列中的已读消息1,并进行播报。
当终端设备接收到用于指示重播消息的指令时,终端设备可以重复播报当前正在播报的消息队列中最近一次播报过的消息。
例如,假设终端设备在接收到用于指示重播消息的指令时,最近播报过的消息为消息队列R中已读消息队列的已读消息P-1,则该终端设备可以重复播报该已读消息队列的已读消息P-1。
当终端设备接收到用于指示停止播报消息的指令时,终端设备可以停止播报,并且清除终端设备中缓存的推送消息。
需要说明的是,若终端设备接收到用于指示播报消息的指令时,消息队列中未读消息队列中没有存储消息,或者,终端设备在接收到用于指示播报上一条消息的指令时,终端设备已经播报至消息队列的队首;或者,终端设备在接收到用于指示播报下一条消息的播报控制指令时,终端设备已经播报至消息队列的队尾,则终端设备可以返回提示信息,该提示信息可以为语音消息“无更多消息”,用于提示用户没有可供播报的消息。
综上所述,本申请实施例在收到推送消息时,会用消息提示音来提醒用户,若用户认为当前环境不存在隐私泄露的危险,则可以发出用于指示播报所述推送消息的预设指令,若该指令核实无误,则进一步通过声纹匹配判断发出指令的是否为指定用户,从而大大提高了推送消息播报的私密性,减少了用户隐私泄露的几率。
对应于上文实施例所述的一种推送消息播报方法,图5示出了本申请实施例提供的一种推送消息播报装置的一个实施例结构图。
本实施例中,一种推送消息播报装置可以包括:
推送消息接收模块501,用于接收消息推送服务器下发的推送消息;
语音信息采集模块502,用于采集预设的时间段内的语音信息;
指令判断模块503,用于判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
声纹特征向量提取模块504,用于若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;
匹配度计算模块505,用于计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
推送消息播报模块506,用于若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
进一步地,所述声纹特征向量提取模块可以包括:
语音子段划分单元,用于将所述语音信息划分为M个语音子段,其中,M为大于1的整数;
梅尔频谱倒频系数向量计算单元,用于根据下式计算各个语音子段的梅尔频谱倒频系数向量:
MelVec m=MFCCFuc(SubVoice m)
其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数;
权重系数计算单元,用于根据下式计算各个语音子段的权重系数:
Figure PCTCN2018093708-appb-000004
其中,Weight m为第m个语音子段的权重系数;
声纹特征向量构造单元,用于根据下式构造所述语音信息中的声纹特征向量:
VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
Figure PCTCN2018093708-appb-000005
VoPrintVec为所述语音信息中的声纹特征向量。
进一步地,所述匹配度计算模块可以包括:
匹配度计算单元,用于根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
Figure PCTCN2018093708-appb-000006
其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
进一步地,所述推送消息播报装置还可以包括:
关键词匹配模块,用于将所述推送消息与预设的关键词集合中的关键词依次进行匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词;
推送消息存储模块,用于将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。
进一步地,所述推送消息播报模块可以包括:
当前消息队列选取单元,用于选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列;
队头消息播报单元,用于以语音形式播报队头消息,所述队头消息为排列在所述当前消息队列的队头位置的推送消息;
队头消息删除单元,用于并将所述队头消息从所述当前消息队列中删除;
消息队列判断单元,用于判断所述当前消息队列是否为空。
图6示出了本申请实施例提供的一种推送消息播报终端设备的示意框图,为了便于说明,仅示出了与本申请实施例相关的部分。
在本实施例中,所述推送消息播报终端设备6可以是手机、平板电脑、智能手表、智能手环以及智能眼镜等计算设备。该推送消息播报终端设备6可包括:处理器60、存储器61以及存储在所述存储器61中并可在所述处理器60上运行的计算机可读指令62,例如执行上述的推送消息播报方法的计算机可读指令。所述处理器60执行所述计算机可读指令62时实现上述各个推送消息播报方法实施例中的步骤。
在本申请各个实施例中的各功能单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读存储介质中。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干计算机可读指令用以使得一台计算机设备执行本申请各个实施例所述方法的全部或部分步骤。

Claims (20)

  1. 一种推送消息播报方法,其特征在于,包括:
    接收消息推送服务器下发的推送消息;
    采集预设的时间段内的语音信息;
    判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
    若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;
    计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
    若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
  2. 根据权利要求1所述的推送消息播报方法,其特征在于,所述提取所述语音信息中的声纹特征向量包括:
    将所述语音信息划分为M个语音子段,其中,M为大于1的整数;
    根据下式计算各个语音子段的梅尔频谱倒频系数向量:
    MelVec m=MFCCFuc(SubVoice m)
    其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数;
    根据下式计算各个语音子段的权重系数:
    Figure PCTCN2018093708-appb-100001
    其中,Weight m为第m个语音子段的权重系数;
    根据下式构造所述语音信息中的声纹特征向量:
    VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
    其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
    Figure PCTCN2018093708-appb-100002
    VoPrintVec为所述语音信息中的声纹特征向量。
  3. 根据权利要求2所述的推送消息播报方法,其特征在于,所述计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度包括:
    根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
    Figure PCTCN2018093708-appb-100003
    其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
  4. 根据权利要求1至3中任一项所述的推送消息播报方法,其特征在于,在接收消息推送服务器下发的推送消息之后,还包括:
    将所述推送消息与预设的关键词集合中的关键词依次进行匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词;
    将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。
  5. 根据权利要求4所述的推送消息播报方法,其特征在于,所述以语音形式播报所述推送消息包括:
    选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列;
    以语音形式播报队头消息,并将所述队头消息从所述当前消息队列中删除,所述队头消息为排列在所述当前消息队列的队头位置的推送消息;
    判断所述当前消息队列是否为空;
    若所述当前消息队列不为空,则返回执行所述以语音形式播报队头消息的步骤,直至所述当前消息队列为空为止;
    若所述当前消息队列为空,则返回执行所述选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列的步骤,直至各个未读消息队列均为空为止。
  6. 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机可读指令,其特征在于,所述计算机可读指令被处理器执行时实现如下步骤:
    接收消息推送服务器下发的推送消息;
    采集预设的时间段内的语音信息;
    判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
    若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音 信息中的声纹特征向量;
    计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
    若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
  7. 根据权利要求6所述的计算机可读存储介质,其特征在于,所述提取所述语音信息中的声纹特征向量包括:
    将所述语音信息划分为M个语音子段,其中,M为大于1的整数;
    根据下式计算各个语音子段的梅尔频谱倒频系数向量:
    MelVec m=MFCCFuc(SubVoice m)
    其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数;
    根据下式计算各个语音子段的权重系数:
    Figure PCTCN2018093708-appb-100004
    其中,Weight m为第m个语音子段的权重系数;
    根据下式构造所述语音信息中的声纹特征向量:
    VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
    其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
    Figure PCTCN2018093708-appb-100005
    VoPrintVec为所述语音信息中的声纹特征向量。
  8. 根据权利要求7所述的计算机可读存储介质,其特征在于,所述计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度包括:
    根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
    Figure PCTCN2018093708-appb-100006
    其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
  9. 根据权利要求6至8中任一项所述的计算机可读存储介质,其特征在于,在接收消息推送服务器下发的推送消息之后,还包括:
    将所述推送消息与预设的关键词集合中的关键词依次进行匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词;
    将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。
  10. 根据权利要求9所述的计算机可读存储介质,其特征在于,所述以语音形式播报所述推送消息包括:
    选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列;
    以语音形式播报队头消息,并将所述队头消息从所述当前消息队列中删除,所述队头消息为排列在所述当前消息队列的队头位置的推送消息;
    判断所述当前消息队列是否为空;
    若所述当前消息队列不为空,则返回执行所述以语音形式播报队头消息的步骤,直至所述当前消息队列为空为止;
    若所述当前消息队列为空,则返回执行所述选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列的步骤,直至各个未读消息队列均为空为止。
  11. 一种推送消息播报终端设备,包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令,其特征在于,所述处理器执行所述计算机可读指令时实现如下步骤:
    接收消息推送服务器下发的推送消息;
    采集预设的时间段内的语音信息;
    判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
    若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;
    计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
    若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
  12. 根据权利要求11所述的推送消息播报终端设备,其特征在于,所述提取所述语音信息中的声纹特征向量包括:
    将所述语音信息划分为M个语音子段,其中,M为大于1的整数;
    根据下式计算各个语音子段的梅尔频谱倒频系数向量:
    MelVec m=MFCCFuc(SubVoice m)
    其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数;
    根据下式计算各个语音子段的权重系数:
    Figure PCTCN2018093708-appb-100007
    其中,Weight m为第m个语音子段的权重系数;
    根据下式构造所述语音信息中的声纹特征向量:
    VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
    其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
    Figure PCTCN2018093708-appb-100008
    VoPrintVec为所述语音信息中的声纹特征向量。
  13. 根据权利要求12所述的推送消息播报终端设备,其特征在于,所述计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度包括:
    根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
    Figure PCTCN2018093708-appb-100009
    其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
  14. 根据权利要求11至13中任一项所述的推送消息播报终端设备,其特征在于,在接收消息推送服务器下发的推送消息之后,还包括:
    将所述推送消息与预设的关键词集合中的关键词依次进行匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词;
    将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。
  15. 根据权利要求14所述的推送消息播报终端设备,其特征在于,所述以语音形式播报所述推送消息包括:
    选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列;
    以语音形式播报队头消息,并将所述队头消息从所述当前消息队列中删除,所述队头消息为排列在所述当前消息队列的队头位置的推送消息;
    判断所述当前消息队列是否为空;
    若所述当前消息队列不为空,则返回执行所述以语音形式播报队头消息的步骤,直至所述当前消息队列为空为止;
    若所述当前消息队列为空,则返回执行所述选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列的步骤,直至各个未读消息队列均为空为止。
  16. 一种推送消息播报装置,其特征在于,包括:
    推送消息接收模块,用于接收消息推送服务器下发的推送消息;
    语音信息采集模块,用于采集预设的时间段内的语音信息;
    指令判断模块,用于判断所述语音信息中是否包含用于指示播报所述推送消息的预设指令;
    声纹特征向量提取模块,用于若所述语音信息中包含用于指示播报所述推送消息的预设指令,则提取所述语音信息中的声纹特征向量;
    匹配度计算模块,用于计算所述语音信息中的声纹特征向量与预设的基准声纹特征向量之间的匹配度,所述基准声纹特征向量为从指定用户的语音中提取的声纹特征向量;
    推送消息播报模块,用于若所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度大于预设的匹配度阈值,则以语音形式播报所述推送消息。
  17. 根据权利要求16所述的推送消息播报装置,其特征在于,所述声纹特征向量提取模块包括:
    语音子段划分单元,用于将所述语音信息划分为M个语音子段,其中,M为大于 1的整数;
    梅尔频谱倒频系数向量计算单元,用于根据下式计算各个语音子段的梅尔频谱倒频系数向量:
    MelVec m=MFCCFuc(SubVoice m)
    其中,m为语音子段的序号,1≤m≤M,SubVoice m为第m个语音子段,MFCCFuc为预设的梅尔频谱倒频系数计算函数,MelVec m为第m个语音子段的梅尔频谱倒频系数向量,且MelVec m=(MelCoe m,1,MelCoe m,2,......,MelCoe m,n,......,MelCoe m,N),n为梅尔频谱倒频系数的序号,1≤n≤N,N为梅尔频谱倒频系数向量的预设维度,MelCoe m,n为第m个语音子段的第n个梅尔频谱倒频系数;
    权重系数计算单元,用于根据下式计算各个语音子段的权重系数:
    Figure PCTCN2018093708-appb-100010
    其中,Weight m为第m个语音子段的权重系数;
    声纹特征向量构造单元,用于根据下式构造所述语音信息中的声纹特征向量:
    VoPrintVec=(VpElem 1,VpElem 2,......,VpElem n,......,VpElem N)
    其中,VpElem n为所述语音信息中的声纹特征向量的第n个元素,且
    Figure PCTCN2018093708-appb-100011
    VoPrintVec为所述语音信息中的声纹特征向量。
  18. 根据权利要求17所述的推送消息播报装置,其特征在于,所述匹配度计算模块包括:
    匹配度计算单元,用于根据下式计算所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度:
    Figure PCTCN2018093708-appb-100012
    其中,StVpElem n为所述基准声纹特征向量的第n个元素,MatDeg为所述语音信息中的声纹特征向量与所述基准声纹特征向量之间的匹配度。
  19. 根据权利要求16至18中任一项所述的推送消息播报装置,其特征在于,还包括:
    关键词匹配模块,用于将所述推送消息与预设的关键词集合中的关键词依次进行 匹配,并将与匹配成功的关键词对应的消息类型确定为所述推送消息的消息类型,所述关键词集合中包括与预设的各种消息类型分别对应的各个关键词子集,每个关键词子集中包括一个以上的关键词;
    推送消息存储模块,用于将所述推送消息存储至与所述推送消息的消息类型对应的未读消息队列中。
  20. 根据权利要求19所述的推送消息播报装置,其特征在于,所述推送消息播报模块包括:
    当前消息队列选取单元,用于选取预设的优先级最高且队列不为空的未读消息队列作为当前消息队列;
    队头消息播报单元,用于以语音形式播报队头消息,所述队头消息为排列在所述当前消息队列的队头位置的推送消息;
    队头消息删除单元,用于并将所述队头消息从所述当前消息队列中删除;
    消息队列判断单元,用于判断所述当前消息队列是否为空。
PCT/CN2018/093708 2018-05-03 2018-06-29 推送消息播报方法、可读存储介质、终端设备及装置 Ceased WO2019210567A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810416166.X 2018-05-03
CN201810416166.XA CN108769131B (zh) 2018-05-03 2018-05-03 一种推送消息播报方法、计算机可读存储介质及终端设备

Publications (1)

Publication Number Publication Date
WO2019210567A1 true WO2019210567A1 (zh) 2019-11-07

Family

ID=64009350

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/093708 Ceased WO2019210567A1 (zh) 2018-05-03 2018-06-29 推送消息播报方法、可读存储介质、终端设备及装置

Country Status (2)

Country Link
CN (1) CN108769131B (zh)
WO (1) WO2019210567A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111276123B (zh) * 2018-11-16 2023-01-24 阿拉的(深圳)人工智能有限公司 一种语音播报留言的方法、装置、计算机设备及存储介质
CN110224848B (zh) * 2019-03-29 2022-12-06 北京龙鼎源科技股份有限公司 告警的播报方法及装置
CN111241512B (zh) * 2020-01-09 2022-06-17 珠海格力电器股份有限公司 留言信息播报方法、装置、电子设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8626246B2 (en) * 2001-09-05 2014-01-07 Vocera Communications, Inc. Voice-controlled communications system and method using a badge application
CN103956168A (zh) * 2014-03-29 2014-07-30 深圳创维数字技术股份有限公司 一种语音识别方法、装置及终端
CN106961376A (zh) * 2016-01-08 2017-07-18 腾讯科技(深圳)有限公司 即时通讯信息的处理方法和装置
CN107506166A (zh) * 2017-08-04 2017-12-22 珠海市魅族科技有限公司 信息提示方法及装置、计算机装置及可读存储介质
CN107895578A (zh) * 2017-11-15 2018-04-10 百度在线网络技术(北京)有限公司 语音交互方法和装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012075039A (ja) * 2010-09-29 2012-04-12 Sony Corp 制御装置、および制御方法
CN102131156A (zh) * 2011-01-27 2011-07-20 上海华美络信息技术有限公司 一种医院即时通讯系统及即时通讯方法
CN102404330B (zh) * 2011-11-30 2018-07-20 上海博泰悦臻电子设备制造有限公司 多用户同时在线管理方法和系统
CN205427990U (zh) * 2016-02-14 2016-08-03 北京得意音通技术有限责任公司 一种基于动态数字验证码的语音门禁系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8626246B2 (en) * 2001-09-05 2014-01-07 Vocera Communications, Inc. Voice-controlled communications system and method using a badge application
CN103956168A (zh) * 2014-03-29 2014-07-30 深圳创维数字技术股份有限公司 一种语音识别方法、装置及终端
CN106961376A (zh) * 2016-01-08 2017-07-18 腾讯科技(深圳)有限公司 即时通讯信息的处理方法和装置
CN107506166A (zh) * 2017-08-04 2017-12-22 珠海市魅族科技有限公司 信息提示方法及装置、计算机装置及可读存储介质
CN107895578A (zh) * 2017-11-15 2018-04-10 百度在线网络技术(北京)有限公司 语音交互方法和装置

Also Published As

Publication number Publication date
CN108769131A (zh) 2018-11-06
CN108769131B (zh) 2021-05-28

Similar Documents

Publication Publication Date Title
US12099674B2 (en) Systems and methods for identifying a set of characters in a media file
US11037560B2 (en) Method, apparatus and storage medium for wake up processing of application
CN107211061B (zh) 用于空间会议回放的优化虚拟场景布局
US10270736B2 (en) Account adding method, terminal, server, and computer storage medium
US9386256B1 (en) Systems and methods for identifying a set of characters in a media file
CN108766418A (zh) 语音端点识别方法、装置及设备
CN106250400B (zh) 一种音频数据处理方法、装置以及系统
CN110517082A (zh) 广告推送方法、装置、设备及计算机可读存储介质
US8412527B2 (en) Automatic disclosure detection
CN111666059B (zh) 提醒信息播报方法、装置,以及电子设备
CN107210036A (zh) 会议词语云
JP5171501B2 (ja) 重要語を抽出するサーバ、システム、方法およびプログラム
CN112399269B (zh) 视频分割方法、装置、设备及存储介质
US20120053937A1 (en) Generalizing text content summary from speech content
WO2022142031A1 (zh) 无效通话判断方法、装置、计算机设备及存储介质
WO2019210567A1 (zh) 推送消息播报方法、可读存储介质、终端设备及装置
CN115150660B (zh) 一种基于字幕的视频编辑方法和相关设备
JP2009238199A5 (zh)
CN109582823A (zh) 视频信息链式存储方法、装置、计算机设备及存储介质
JP2016062333A (ja) 検索サーバ、及び検索方法
US20140129221A1 (en) Sound recognition device, non-transitory computer readable storage medium stored threreof sound recognition program, and sound recognition method
CN112634906A (zh) 一种主播话术语音识别统计方法、装置、设备和存储介质
CN103247316A (zh) 一种音频检索中构建索引的方法及系统
TWI815400B (zh) 情緒分析系統
CN112509567B (zh) 语音数据处理的方法、装置、设备、存储介质及程序产品

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18917069

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 19.02.2021.)

122 Ep: pct application non-entry in european phase

Ref document number: 18917069

Country of ref document: EP

Kind code of ref document: A1