WO2017050120A1 - 儿童锁启动方法及装置 - Google Patents

儿童锁启动方法及装置 Download PDF

Info

Publication number
WO2017050120A1
WO2017050120A1 PCT/CN2016/098070 CN2016098070W WO2017050120A1 WO 2017050120 A1 WO2017050120 A1 WO 2017050120A1 CN 2016098070 W CN2016098070 W CN 2016098070W WO 2017050120 A1 WO2017050120 A1 WO 2017050120A1
Authority
WO
WIPO (PCT)
Prior art keywords
sound
child lock
sound feature
extracting
limited user
Prior art date
Application number
PCT/CN2016/098070
Other languages
English (en)
French (fr)
Inventor
龚松
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2017050120A1 publication Critical patent/WO2017050120A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L17/00—Speaker identification or verification techniques
    • G10L17/04—Training, enrolment or model building

Definitions

  • the present invention relates to the field of communications, and in particular to a child lock activation method and apparatus.
  • the embodiment of the invention provides a method and a device for starting a child lock, so as to at least solve the problem that the startup mode of the child lock is not intelligent enough in the related art.
  • a child lock activation method including: extracting a sound feature of a limited time length of a limited user; determining whether the sound feature matches a preset sound feature; In the case of the case, the child lock mode is activated.
  • extracting the sound feature of the limited user for a predetermined length of time comprises: acquiring, by the recording device, a sound file of the limited user for a predetermined length of time in the child lock interface of the set top box; extracting the sound feature according to the sound file .
  • extracting the sound feature according to the sound file includes: preprocessing a voice signal of the sound file, including: removing a non-speech signal and a silent voice signal, and framing the voice signal; extracting each a Mel Frequency Cepstrum Coefficient (MFCC) parameter of a frame of speech signal is saved and saved; the Gaussian mixture model of the limited user is trained using the extracted MFCC parameter to obtain the sound of the limited user feature.
  • MFCC Mel Frequency Cepstrum Coefficient
  • determining whether the sound feature matches the preset sound feature comprises: calculating a probability of a pre-saved Gaussian mixture model in the currently collected Gaussian mixture model matching database, and controlling a probability threshold to obtain the currently extracted Whether the sound characteristics match the pre-stored sound features.
  • the method further comprises: closing the child lock mode by restarting the set top box.
  • a child lock activation device comprising: an extraction module configured to extract a sound feature of a limited time length of a limited user; and a determination module configured to determine the sound feature and Whether the preset sound characteristics match; the startup module is set to start the child lock mode if the judgment result is YES.
  • the extracting module includes: an acquiring unit, configured to acquire, by using a recording device, a sound file of the limited user for a predetermined length of time in a child lock interface of the set top box; and an extracting unit configured to extract the sound file according to the sound file The characteristics of the sound.
  • the extracting unit includes: a preprocessing subunit, configured to preprocess the voice signal of the sound file, including: removing the non-speech signal and the silent voice signal, and framing the voice signal; and extracting a subunit, configured to extract a Meer frequency cepstral MFCC parameter of each frame of the speech signal and save; a training subunit configured to train the Gaussian mixture model of the restricted user using the extracted MFCC parameter to obtain the subject Limit the user's voice characteristics.
  • a preprocessing subunit configured to preprocess the voice signal of the sound file, including: removing the non-speech signal and the silent voice signal, and framing the voice signal
  • extracting a subunit configured to extract a Meer frequency cepstral MFCC parameter of each frame of the speech signal and save
  • a training subunit configured to train the Gaussian mixture model of the restricted user using the extracted MFCC parameter to obtain the subject Limit the user's voice characteristics.
  • the determining module includes: a calculating unit configured to calculate a probability of a Gaussian mixture model pre-stored in the currently collected Gaussian mixture model matching database, and control a probability threshold to obtain a currently extracted sound feature and a predetermined Whether the stored sound characteristics match.
  • the apparatus further comprises: a shutdown module configured to close the child lock mode by restarting the set top box.
  • a storage medium is also provided.
  • the storage medium is arranged to store program code for performing the following steps:
  • the storage medium is further arranged to store program code for performing the following steps:
  • Extracting the sound feature of the limited user for a predetermined length of time comprises: acquiring, by the recording device, a sound file of the limited user for a predetermined length of time at the child lock interface of the set top box; and extracting the sound feature according to the sound file.
  • Extracting the sound feature according to the sound file includes: preprocessing a voice signal of the sound file, including: removing a non-speech signal and a silent voice signal, and framing the voice signal; extracting each frame of the voice signal
  • the Mel Frequency Cepstrum Coefficient (MFCC) parameter is saved and saved; the Gaussian mixture model of the limited user is trained using the extracted MFCC parameters to obtain the sound characteristics of the limited user.
  • MFCC Mel Frequency Cepstrum Coefficient
  • Determining whether the sound feature matches the preset sound feature comprises: calculating a probability of a pre-saved Gaussian mixture model in the currently collected Gaussian mixture model matching database, and controlling a probability threshold to obtain the currently extracted sound feature and the advance Whether the stored sound characteristics match.
  • the method further includes closing the child lock mode by restarting the set top box.
  • the sound feature of the predetermined time length of the limited user is extracted; whether the sound feature matches the preset sound feature is determined; if the determination result is yes, the child lock mode is activated, and the relevant
  • the method of starting the child lock is not smart enough, and the child lock can be activated according to the sound, thereby improving the user experience.
  • FIG. 1 is a first flowchart of a child lock activation method according to an embodiment of the present invention
  • FIG. 2 is a block diagram of a child lock activation device in accordance with an embodiment of the present invention.
  • Figure 3 is a block diagram 1 of a child lock activation device in accordance with a preferred embodiment of the present invention.
  • Figure 4 is a block diagram 2 of a child lock activation device in accordance with a preferred embodiment of the present invention.
  • Figure 5 is a block diagram 3 of a child lock activation device in accordance with a preferred embodiment of the present invention.
  • Figure 6 is a block diagram 4 of a child lock activation device in accordance with a preferred embodiment of the present invention.
  • FIG. 7 is a flow chart showing a booting phase of a set top box according to an embodiment of the present invention.
  • FIG. 8 is a flow chart of daily background monitoring of a set top box according to an embodiment of the present invention.
  • FIG. 1 is a flowchart 1 of a child lock activation method according to an embodiment of the present invention.
  • Step S102 extracting a sound feature of a limited time length of the limited user
  • Step S104 determining whether the sound feature matches a preset sound feature
  • step S106 if the result of the determination is YES, the child lock mode is activated.
  • the sound file of the limited user for a predetermined length of time is acquired by the recording device at the child lock interface of the set top box; the sound feature is extracted according to the sound file.
  • extracting the sound feature according to the sound file includes: preprocessing the voice signal of the sound file, including: removing the non-speech signal and the silent voice signal, and performing frame division on the voice signal; extracting each frame of the voice signal
  • the Mel frequency cepstrum MFCC parameters are saved; the Gaussian mixture model of the restricted user is trained using the extracted MFCC parameters to obtain the sound characteristics of the limited user.
  • determining whether the sound feature matches the preset sound feature may include: calculating a probability of a pre-saved Gaussian mixture model in the currently collected Gaussian mixture model matching database, and controlling a probability threshold to obtain a Whether the previously extracted sound features match the pre-stored sound features.
  • the child lock mode is turned off by restarting the set top box.
  • FIG. 2 is a block diagram of a child lock activation device according to an embodiment of the present invention. As shown in FIG. 2, the method includes:
  • the extracting module 22 is configured to extract a sound feature of the limited user for a predetermined length of time
  • the determining module 24 is configured to determine whether the sound feature matches a preset sound feature
  • the startup module 26 is set to activate the child lock mode if the determination result is YES.
  • the extraction module 22 includes:
  • the obtaining unit 32 is configured to acquire, by the recording device, a sound file of the limited user for a predetermined length of time in the child lock interface of the set top box;
  • the extracting unit 34 is arranged to extract the sound feature from the sound file.
  • the extraction unit 34 includes:
  • the pre-processing sub-unit 42 is configured to pre-process the voice signal of the sound file, including: removing the non-speech signal and the silent voice signal, and framing the voice signal;
  • Extracting sub-unit 44 configured to extract the Mel frequency cepstral MFCC parameters of each frame of the speech signal and save;
  • the training sub-unit 46 is arranged to train the Gaussian mixture model of the limited user using the extracted MFCC parameters to obtain the sound characteristics of the limited user.
  • FIG. 5 is a block diagram 3 of a child lock activation device according to a preferred embodiment of the present invention. As shown in FIG. 5, the determination module 24 includes:
  • the calculating unit 52 is configured to calculate a probability of the Gaussian mixture model stored in the currently collected Gaussian mixture model matching database, and control a probability threshold to determine whether the currently extracted sound feature matches the pre-stored sound feature.
  • Figure 6 is a block diagram 4 of a child lock activation device in accordance with a preferred embodiment of the present invention. As shown in Figure 6, the device further includes:
  • the shutdown module 62 is configured to close the child lock mode by restarting the set top box.
  • the embodiment of the invention proposes an automatic activation mechanism of the child lock, that is, the voiceprint recognition method is used to complete the function.
  • the parent enters the child lock interface of the set top box, and the child's 10 second sound is recorded into the set top box through the microphone device, and is saved.
  • the function will take effect immediately. Unless the child is added to the home, there is no need to modify it. This function is normally open in the set-top box.
  • FIG. 7 is a flowchart of a boot-up phase of a set-top box according to an embodiment of the present invention.
  • the voiceprint monitoring service is started by the system, and the current check is started after the startup.
  • the voiceprint feature file already exists, and if it exists, it is read out, and the sound data in the environment is collected by the microphone itself in the background. If it is checked that there is no voiceprint feature file at present, it will exit immediately.
  • the parents enter the child's voice through the microphone in the setting interface of the set-top box.
  • the setting module of the box will extract the 10 seconds sound collected at this time, through the voiceprint feature extraction, and the voiceprint modeling finally generates the voiceprint feature file. In the box.
  • FIG. 8 is a flowchart of daily background monitoring of a set top box according to an embodiment of the present invention.
  • the voiceprint monitoring service collects 10 seconds of audio data in the environment every half minute during daily operation of the box.
  • the voiceprint data is extracted from the sound data.
  • the child's voiceprint data recorded before the box is compared. If the judgment is consistent, the child lock is immediately opened, and the child lock mode is activated.
  • the above voiceprint feature extraction includes the following steps:
  • the code stream collected by the recording device is pre-processed, that is, the audio data is first framed by sampling the number of frames in a fixed time, and then the squared summation operation is performed on the voice signal of each frame. Way to remove silence (time period without sound) sound;
  • the signal is finally transformed into a set of MFCC Mel frequency cepstrum parameters describing the speech features
  • the step of matching the voiceprint feature described above is to calculate the probability that the currently collected speaker's speech sequence matches the speech Gaussian mixture model in the data by using the posterior probability, and control a probability threshold to obtain the currently collected speech signal and Whether there is matching data in the database, if it matches, the child lock is automatically activated to achieve the goal.
  • Parents and children are watching TV. After the parents leave, the children are left alone. At this time, because the voiceprint monitoring service is resident in the background, once the parents leave, the box will enter the children mode in a short time. .
  • modules or steps of the present invention described above can be implemented by a general-purpose computing device that can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein.
  • the steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps thereof are fabricated as a single integrated circuit module.
  • the invention is not limited to any specific combination of hardware and software.
  • the sound feature of the predetermined time length of the limited user is extracted; whether the sound feature matches the preset sound feature is determined; if the determination result is yes, the child lock mode is activated, and the relevant
  • the method of starting the child lock is not smart enough, and the child lock can be activated according to the sound, thereby improving the user experience.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Telephone Function (AREA)
  • Electrically Operated Instructional Devices (AREA)

Abstract

一种儿童锁启动方法及装置，其中，该方法包括：提取受限用户的预定时间长度的声音特征（S102）；判断该声音特征与预先设置的声音特征是否匹配（S104）；在判断结果为是的情况下，启动儿童锁模式（S106），解决了相关技术中对儿童锁启动方式不够智能的问题，能够根据声音启动儿童锁，提高了用户体验。

Description

儿童锁启动方法及装置 技术领域
本发明涉及通信领域，具体而言，涉及一种儿童锁启动方法及装置。
背景技术
目前市面上的机顶盒已广泛使用于用户家庭的客厅布局中，其中针对儿童锁的操作方式基本都是通过家长每次去主动设置儿童锁的开关或者密码而进行的，这种操作方式存在的缺陷在于家长在观看结束时，要靠自身意志来控制儿童锁的开关，这样的做法是不够完善的，完全取决于家长是否记得这件事情，不够智能。
针对相关技术中对儿童锁的启动方式不够智能的问题，还未提出有效的解决方案。
发明内容
本发明实施例提供了一种儿童锁启动方法及装置，以至少解决相关技术中对儿童锁的启动方式不够智能的问题。
根据本发明实施例的一个方面，提供了一种儿童锁启动方法，包括：提取受限用户的预定时间长度的声音特征；判断所述声音特征与预先设置的声音特征是否匹配；在判断结果为是的情况下，启动儿童锁模式。
可选地，提取受限用户的预定时间长度的声音特征包括：在机顶盒的儿童锁界面通过录音设备获取所述受限用户的预定时间长度的声音文件；根据所述声音文件提取所述声音特征。
可选地，根据所述声音文件提取所述声音特征包括：对所述声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对所述语音信号进行分帧；提取每一帧语音信号的梅尔频率倒谱(Mel Frequency Cepstrum Coefficient，简称为MFCC)参数并保存；使用提取的所述MFCC参数训练所述受限用户的高斯混合模型，得到所述受限用户的声音特征。
可选地，判断所述声音特征与预先设置的声音特征是否匹配包括：计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
可选地，在启动所述儿童锁模式之后，所述方法还包括：通过重启所述机顶盒关闭所述儿童锁模式。
根据本发明实施例的另一方面，还提供了一种儿童锁启动装置，包括：提取模块，设置为提取受限用户的预定时间长度的声音特征；判断模块，设置为判断所述声音特征与预先设置的声音特征是否匹配；启动模块，设置为在判断结果为是的情况下，启动儿童锁模式。
可选地，所述提取模块包括：获取单元，设置为在机顶盒的儿童锁界面通过录音设备获取所述受限用户的预定时间长度的声音文件；提取单元，设置为根据所述声音文件提取所述声音特征。
可选地，所述提取单元包括：预处理子单元，设置为对所述声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对所述语音信号进行分帧；提取子单元，设置为提取每一帧语音信号的梅尔频率倒谱MFCC参数并保存；训练子单元，设置为使用提取的所述MFCC参数训练所述受限用户的高斯混合模型，得到所述受限用户的声音特征。
可选地，所述判断模块包括：计算单元，设置为计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
可选地，所述装置还包括：关闭模块，设置为通过重启所述机顶盒关闭所述儿童锁模式。
根据本发明的又一个实施例，还提供了一种存储介质。该存储介质设置为存储用于执行以下步骤的程序代码：
提取受限用户的预定时间长度的声音特征；判断所述声音特征与预先设置的声音特征是否匹配；在判断结果为是的情况下，启动儿童锁模式。
可选地，存储介质还设置为存储用于执行以下步骤的程序代码：
提取受限用户的预定时间长度的声音特征包括：在机顶盒的儿童锁界面通过录音设备获取所述受限用户的预定时间长度的声音文件；根据所述声音文件提取所述声音特征。
根据所述声音文件提取所述声音特征包括：对所述声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对所述语音信号进行分帧；提取每一帧语音信号的梅尔频率倒谱(Mel Frequency Cepstrum Coefficient，简称为MFCC)参数并保存；使用提取的所述MFCC参数训练所述受限用户的高斯混合模型，得到所述受限用户的声音特征。
判断所述声音特征与预先设置的声音特征是否匹配包括：计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
在启动所述儿童锁模式之后，所述方法还包括：通过重启所述机顶盒关闭所述儿童锁模式。
通过本发明实施例，采用提取受限用户的预定时间长度的声音特征；判断所述声音特征与预先设置的声音特征是否匹配；在判断结果为是的情况下，启动儿童锁模式，解决了相关技术中对儿童锁的启动方式不够智能的问题，能够根据声音启动儿童锁，提高了用户体验。
附图说明
此处所说明的附图用来提供对本发明的进一步理解，构成本申请的一部分，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：
图1是根据本发明实施例的儿童锁启动方法的流程图一；
图2是根据本发明实施例的儿童锁启动装置的框图；
图3是根据本发明优选实施例的儿童锁启动装置的框图一；
图4是根据本发明优选实施例的儿童锁启动装置的框图二；
图5是根据本发明优选实施例的儿童锁启动装置的框图三；
图6是根据本发明优选实施例的儿童锁启动装置的框图四；
图7是根据本发明实施例的机顶盒开机阶段的流程图；
图8是根据本发明实施例的机顶盒日常后台监听的流程图。
具体实施方式
下文中将参考附图并结合实施例来详细说明本发明。需要说明的是，在不冲突的情况下，本申请中的实施例及实施例中的特征可以相互组合。
本发明实施例提供了一种儿童锁启动方法，图1是根据本发明实施例的儿童锁启动方法的流程图一，如图1所示，包括：
步骤S102，提取受限用户的预定时间长度的声音特征；
步骤S104，判断该声音特征与预先设置的声音特征是否匹配；
步骤S106，在判断结果为是的情况下，启动儿童锁模式。
通过上述步骤，提取受限用户的预定时间长度的声音特征；判断该声音特征与预先设置的声音特征是否匹配；在判断结果为是的情况下，启动儿童锁模式，解决了相关技术中对儿童锁的启动方式不够智能的问题，能够根据声音启动儿童锁，提高了用户体验。
可选地，在机顶盒的儿童锁界面通过录音设备获取该受限用户的预定时间长度的声音文件；根据该声音文件提取该声音特征。
可选地，根据该声音文件提取该声音特征包括：对该声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对该语音信号进行分帧；提取每一帧语音信号的梅尔频率倒谱MFCC参数并保存；使用提取的该MFCC参数训练该受限用户的高斯混合模型，得到该受限用户的声音特征。
可选地，判断该声音特征与预先设置的声音特征是否匹配可以包括：计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当 前提取的声音特征与预先储存的声音特征是否匹配。
可选地，在启动该儿童锁模式之后，通过重启该机顶盒关闭该儿童锁模式。
本发明实施例还提供了一种儿童锁启动装置，图2是根据本发明实施例的儿童锁启动装置的框图，如图2所示，包括：
提取模块22，设置为提取受限用户的预定时间长度的声音特征；
判断模块24，设置为判断该声音特征与预先设置的声音特征是否匹配；
启动模块26，设置为在判断结果为是的情况下，启动儿童锁模式。
图3是根据本发明优选实施例的儿童锁启动装置的框图一，如图3所示，提取模块22包括：
获取单元32，设置为在机顶盒的儿童锁界面通过录音设备获取该受限用户的预定时间长度的声音文件；
提取单元34，设置为根据该声音文件提取该声音特征。
图4是根据本发明优选实施例的儿童锁启动装置的框图二，如图4所示，提取单元34包括：
预处理子单元42，设置为对该声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对该语音信号进行分帧；
提取子单元44，设置为提取每一帧语音信号的梅尔频率倒谱MFCC参数并保存；
训练子单元46，设置为使用提取的该MFCC参数训练该受限用户的高斯混合模型，得到该受限用户的声音特征。
图5是根据本发明优选实施例的儿童锁启动装置的框图三，如图5所示，判断模块24包括：
计算单元52，设置为计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
图6是根据本发明优选实施例的儿童锁启动装置的框图四，如图6所示，该装置还包括：
关闭模块62，设置为通过重启该机顶盒关闭该儿童锁模式。
下面结合优选实施例对本发明实施例进行进一步说明。
本发明实施例提出了一种儿童锁的自动激活机制，即通过声纹识别的方法来完成这个功能，首先家长进入机顶盒的儿童锁界面，通过麦克风设备将儿童10秒的声音录入机顶盒中，保存以后该功能就立即生效，除非家里新增了儿童否则无需再修改，该功能在机顶盒正常开 机后会检查当前盒子是否已录入了童音声纹文件，如有的话便立即启动后台声纹监听模式，通过即时声纹特征提取来和之前录入的声音特征进行对比，如判断为一致则自动打开童锁，整个用户操作除了第一次的儿童声音数据提取和童锁相关功能设置外，后续都无需其他的人为控制，达到家庭客厅的儿童模式自动值守功能。需要说明的是，该功能可以单独使用或配合传统手动设置童锁功能一起使用均可。
图7是根据本发明实施例的机顶盒开机阶段的流程图，如图7所示，开机阶段，机顶盒开机后，监听到开机成功的广播后，声纹监听服务被***启动，启动后检查当前是否已存在声纹特征文件，如已存在则将其读出，自身常驻于后台即时通过麦克采集环境中的声音数据，如检查到当前并不存在任何声纹特征文件，则立即退出。
声纹首次输入阶段，家长在机顶盒的设置界面通过麦克录入儿童的声音，盒子的设置模块将此时采集到的10秒声音，通过声纹特征提取，声纹建模最后生成声纹特征文件保存于盒子中。
图8是根据本发明实施例的机顶盒日常后台监听的流程图，如图8所示，日常监听阶段，声纹监听服务在盒子日常运行中，每隔半分钟采集一次环境中的10秒音频数据，对其中的声音数据进行声纹特征提取，提取完成后，和盒子之前录入的儿童声纹数据进行对比，如判断一致的话则立即打开童锁，启动童锁模式。
上述的声纹特征提取包括以下步骤：
1、首先对录音设备采集到的码流进行预处理，即先通过固定时间内采样个数为一帧的方式对音频数据进行分帧，然后对其每帧的语音信号做平方求和运算的方式，去除静默(没声音的时间段)声音；
2、然后对每帧信号加窗减少吉布斯效应信号重复，屏蔽掉；
3、通过FFT(快速傅里叶变换)将难以看出特性的时域信号变换为信号的功率谱，通过功率谱上的不同能量分布，就能代表不同语音的特性；
4、使用三角带通滤波器，模拟人耳的掩蔽效应(屏蔽掉人耳听不到的声音)，将频谱平滑化，消除谐波；
5、通过离散余弦变换，将信号最终变换成一组描述语音特征的MFCC梅尔频率倒谱参数；
6、最后对上述的MFCC参数建立高斯混合模型，保存于数据库中。
上述的声纹特征匹配的步骤是，通过后验概率计算当前采集到的说话者的语音序列匹配上数据中语音高斯混合模型的概率，并控制一个概率阀值，来得出当前收集的语音信号和数据库中是否有匹配数据，若匹配则自动启动童锁，达到目的。
需要说明的是，上述的声纹特征提取以及声纹特征匹配的过程与相关技术中相同，具体细节不再赘述。
下面对儿童锁启动的几种场景进行进一步说明。
1、只有家长观看电视：启动机顶盒后，只有家长在观看电视时，儿童锁由于声纹监听服务采集到的声音数据并不能和已保存的儿童声音数据匹配，故此处家长无需操作，儿童锁也不会被打开。
2、只有儿童在看电视：启动机顶盒后，只有儿童在观看电视时，儿童锁由于声纹监听服务采集到的声音数据和已保存的儿童声音数据相匹配，故儿童锁会被打开，儿童使用机顶盒受到了保护。
3、家长和儿童一块在看电视，之后家长走了，剩下儿童一个人，这时由于声纹监听服务是常驻后台的，所以家长一旦走了后，短时间内盒子就会进入儿童模式。
4、家长和儿童一块在看电视，之后儿童走了之后，电视可能仍处于儿童模式中，这时家长只需重启机顶盒即可关闭儿童模式。
5、家长和儿童一块在看电视：启动机顶盒后，由于有家长在，此时声纹监听服务采集到的声音数据并不一定能匹配上，所以儿童锁并不一定会打开，不过由于此时家长在场，此时是否打开了儿童锁并没有关系。
显然，本领域的技术人员应该明白，上述的本发明的各模块或各步骤可以用通用的计算装置来实现，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上，可选地，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，并且在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本发明不限制于任何特定的硬件和软件结合。
以上所述仅为本发明的优选实施例而已，并不用于限制本发明，对于本领域的技术人员来说，本发明可以有各种更改和变化。凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。
工业实用性
通过本发明实施例，采用提取受限用户的预定时间长度的声音特征；判断所述声音特征与预先设置的声音特征是否匹配；在判断结果为是的情况下，启动儿童锁模式，解决了相关技术中对儿童锁的启动方式不够智能的问题，能够根据声音启动儿童锁，提高了用户体验。

Claims (10)

  1. 一种儿童锁启动方法，包括：
    提取受限用户的预定时间长度的声音特征；
    判断所述声音特征与预先设置的声音特征是否匹配；
    在判断结果为是的情况下，启动儿童锁模式。
  2. 根据权利要求1所述的方法，其中，提取受限用户的预定时间长度的声音特征包括：
    在机顶盒的儿童锁界面通过录音设备获取所述受限用户的预定时间长度的声音文件；
    根据所述声音文件提取所述声音特征。
  3. 根据权利要求2所述的方法，其中，根据所述声音文件提取所述声音特征包括：
    对所述声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对所述语音信号进行分帧；
    提取每一帧语音信号的梅尔频率倒谱MFCC参数并保存；
    使用提取的所述MFCC参数训练所述受限用户的高斯混合模型，得到所述受限用户的声音特征。
  4. 根据权利要求3所述的方法，其中，判断所述声音特征与预先设置的声音特征是否匹配包括：
    计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
  5. 根据权利要求1至4中任一项所述的方法，其中，在启动所述儿童锁模式之后，所述方法还包括：
    通过重启所述机顶盒关闭所述儿童锁模式。
  6. 一种儿童锁启动装置，包括：
    提取模块，设置为提取受限用户的预定时间长度的声音特征；
    判断模块，设置为判断所述声音特征与预先设置的声音特征是否匹配；
    启动模块，设置为在判断结果为是的情况下，启动儿童锁模式。
  7. 根据权利要求6所述的装置，其中，所述提取模块包括：
    获取单元，设置为在机顶盒的儿童锁界面通过录音设备获取所述受限用户的预定时间长度的声音文件；
    提取单元，设置为根据所述声音文件提取所述声音特征。
  8. 根据权利要求7所述的装置，其中，所述提取单元包括：
    预处理子单元，设置为对所述声音文件的语音信号进行预处理，包括：去除非语音信号和静默语音信号，对所述语音信号进行分帧；
    提取子单元，设置为提取每一帧语音信号的梅尔频率倒谱MFCC参数并保存；
    训练子单元，设置为使用提取的所述MFCC参数训练所述受限用户的高斯混合模型，得到所述受限用户的声音特征。
  9. 根据权利要求8所述的装置，其中，所述判断模块包括：
    计算单元，设置为计算当前采集到的高斯混合模型匹配数据库中预先保存的高斯混合模型的概率，并控制一个概率阈值，得出当前提取的声音特征与预先储存的声音特征是否匹配。
  10. 根据权利要求6至9中任一项所述的装置，其中，所述装置还包括：
    关闭模块，设置为通过重启所述机顶盒关闭所述儿童锁模式。
PCT/CN2016/098070 2015-09-21 2016-09-05 儿童锁启动方法及装置 WO2017050120A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510605365.1 2015-09-21
CN201510605365.1A CN106548779A (zh) 2015-09-21 2015-09-21 儿童锁启动方法及装置

Publications (1)

Publication Number Publication Date
WO2017050120A1 true WO2017050120A1 (zh) 2017-03-30

Family

ID=58364506

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/098070 WO2017050120A1 (zh) 2015-09-21 2016-09-05 儿童锁启动方法及装置

Country Status (2)

Country Link
CN (1) CN106548779A (zh)
WO (1) WO2017050120A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109559740A (zh) * 2017-09-27 2019-04-02 浙江绍兴苏泊尔生活电器有限公司 设备控制方法和装置
CN109769148A (zh) * 2019-02-13 2019-05-17 深圳创维数字技术有限公司 智能电视儿童锁控制方法、装置、智能电视及存储介质

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107452236A (zh) * 2017-07-26 2017-12-08 常州大学 一种儿童趣味益智遥控小车***
CN107415882A (zh) * 2017-08-10 2017-12-01 上海博泰悦臻网络技术服务有限公司 一种儿童锁智能控制***和方法
CN108154588B (zh) * 2017-12-29 2020-11-27 深圳市艾特智能科技有限公司 解锁方法、***、可读存储介质及智能设备
CN108307218A (zh) * 2018-01-25 2018-07-20 广州视源电子科技股份有限公司 电视锁定方法及***
CN108874469B (zh) * 2018-07-16 2021-10-01 广东小天才科技有限公司 一种家教设备的应用管控方法及家教设备
CN110320815B (zh) * 2019-07-31 2023-03-03 广东美的制冷设备有限公司 家电设备的控制方法、装置和电子设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1760566A1 (en) * 2005-08-29 2007-03-07 Top Digital Co., Ltd. Voiceprint-lock system for electronic data
US8036891B2 (en) * 2008-06-26 2011-10-11 California State University, Fresno Methods of identification using voice sound analysis
CN102324232A (zh) * 2011-09-12 2012-01-18 辽宁工业大学 基于高斯混合模型的声纹识别方法及***
CN103035242A (zh) * 2012-12-06 2013-04-10 大连奥林匹克电子城腾飞办公设备商行 一种基于语音控制的计算机***
CN103812656A (zh) * 2013-12-06 2014-05-21 南通芯迎设计服务有限公司 一种具有身份认证功能的数字家庭***
CN104575492A (zh) * 2014-12-31 2015-04-29 深圳市航盛电子股份有限公司 一种声纹识别方法及装置和无钥匙车锁***及实现方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1760566A1 (en) * 2005-08-29 2007-03-07 Top Digital Co., Ltd. Voiceprint-lock system for electronic data
US8036891B2 (en) * 2008-06-26 2011-10-11 California State University, Fresno Methods of identification using voice sound analysis
CN102324232A (zh) * 2011-09-12 2012-01-18 辽宁工业大学 基于高斯混合模型的声纹识别方法及***
CN103035242A (zh) * 2012-12-06 2013-04-10 大连奥林匹克电子城腾飞办公设备商行 一种基于语音控制的计算机***
CN103812656A (zh) * 2013-12-06 2014-05-21 南通芯迎设计服务有限公司 一种具有身份认证功能的数字家庭***
CN104575492A (zh) * 2014-12-31 2015-04-29 深圳市航盛电子股份有限公司 一种声纹识别方法及装置和无钥匙车锁***及实现方法

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109559740A (zh) * 2017-09-27 2019-04-02 浙江绍兴苏泊尔生活电器有限公司 设备控制方法和装置
CN109769148A (zh) * 2019-02-13 2019-05-17 深圳创维数字技术有限公司 智能电视儿童锁控制方法、装置、智能电视及存储介质
CN109769148B (zh) * 2019-02-13 2021-03-12 深圳创维数字技术有限公司 智能电视儿童锁控制方法、装置、智能电视及存储介质

Also Published As

Publication number Publication date
CN106548779A (zh) 2017-03-29

Similar Documents

Publication Publication Date Title
WO2017050120A1 (zh) 儿童锁启动方法及装置
US11042616B2 (en) Detection of replay attack
US9177131B2 (en) User authentication method and apparatus based on audio and video data
CN104575504A (zh) 采用声纹和语音识别进行个性化电视语音唤醒的方法
US20170110125A1 (en) Method and apparatus for initiating an operation using voice data
WO2020181824A1 (zh) 声纹识别方法、装置、设备以及计算机可读存储介质
US20130006633A1 (en) Learning speech models for mobile device users
CN112820291B (zh) 智能家居控制方法、***和存储介质
WO2015103836A1 (zh) 一种语音控制方法及装置
US10984795B2 (en) Electronic apparatus and operation method thereof
CN111343028A (zh) 配网控制方法及装置
US20210118464A1 (en) Method and apparatus for emotion recognition from speech
CN116490920A (zh) 用于针对由自动语音识别***处理的语音输入检测音频对抗性攻击的方法、对应的设备、计算机程序产品和计算机可读载体介质
TWI839834B (zh) 語音喚醒方法和相關裝置
CN113129893B (zh) 一种语音识别方法、装置、设备及存储介质
CN108847246A (zh) 一种动画制作方法、装置、终端及可读介质
CN114999472A (zh) 一种空调控制方法、装置及一种空调
EP3499502A1 (en) Voice information processing method and apparatus
WO2017177629A1 (zh) 远讲语音识别方法及装置
US9626967B2 (en) Information processing method and electronic device
WO2023185004A1 (zh) 一种音色切换方法及装置
CN110661923A (zh) 一种在会议中记录发言信息的方法和装置
CN108648758B (zh) 医疗场景中分离无效语音的方法及***
CN113160821A (zh) 一种基于语音识别的控制方法及装置
CN113380244A (zh) 一种设备播放音量的智能调节方法和***

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: 16848002

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16848002

Country of ref document: EP

Kind code of ref document: A1