WO2018068422A1 - 一种提高油烟机语音识别率的方法及*** - Google Patents

一种提高油烟机语音识别率的方法及*** Download PDF

Info

Publication number
WO2018068422A1
WO2018068422A1 PCT/CN2016/113278 CN2016113278W WO2018068422A1 WO 2018068422 A1 WO2018068422 A1 WO 2018068422A1 CN 2016113278 W CN2016113278 W CN 2016113278W WO 2018068422 A1 WO2018068422 A1 WO 2018068422A1
Authority
WO
WIPO (PCT)
Prior art keywords
voice command
voice
range hood
fan speed
module
Prior art date
Application number
PCT/CN2016/113278
Other languages
English (en)
French (fr)
Inventor
宋夏
Original Assignee
广州视源电子科技股份有限公司
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 广州视源电子科技股份有限公司 filed Critical 广州视源电子科技股份有限公司
Publication of WO2018068422A1 publication Critical patent/WO2018068422A1/zh

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24C—DOMESTIC STOVES OR RANGES ; DETAILS OF DOMESTIC STOVES OR RANGES, OF GENERAL APPLICATION
    • F24C15/00—Details
    • F24C15/20—Removing cooking fumes
    • F24C15/2021—Arrangement or mounting of control or safety systems
    • 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
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • 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
    • G10L15/00—Speech recognition
    • G10L15/20—Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech
    • 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
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/223—Execution procedure of a spoken command
    • 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
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/226—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics
    • G10L2015/228—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics of application context

Definitions

  • Embodiments of the present invention relate to the field of voice recognition technology, and in particular, to a method and system for improving a voice recognition rate of a range hood.
  • voice control range hoods have entered people's family life, and different languages can be used.
  • voice commands to control the various functions of the range hood for example, the function of controlling the switch of the range hood, the position of the fan and the timing through voice commands, which greatly facilitates people's lives, and also saves the previous function buttons on the range hood. It is especially convenient for the elderly.
  • the speech recognition model is usually trained by increasing the complexity of the speech recognition algorithm or according to the running noise, but if the speech recognition algorithm is complicated, it is bound to increase the CPU (Central Processing Unit).
  • the workload of the device affects the response speed of the device and increases the power consumption of the device.
  • the range hood is usually provided with multiple operating speeds in different kitchen environments. The noise generated during operation is also different, so the speech recognition model obtained by training may not be suitable for the actual use environment of the range hood.
  • the embodiment of the invention provides a method and a system for improving the speech recognition rate of the range hood, so as to improve the speech recognition rate by appropriately reducing the running noise of the range hood when the speech recognition is difficult.
  • an embodiment of the present invention provides a method for improving a voice recognition rate of a range hood, the method comprising:
  • the method further includes:
  • the method further includes: if the recognition of the second voice command fails within a set time, recovering the fan speed of the range hood, and deleting the saved first voice command.
  • the method further includes:
  • an embodiment of the present invention provides a system for improving a voice recognition rate of a range hood, the system comprising:
  • a reducing module configured to reduce a current fan speed of the range hood when the number of consecutive failures of identifying the first voice command exceeds a set value
  • a receiving module configured to receive a second voice command, and identify the second voice command
  • an instruction execution module configured to perform an indication matching the second voice command if the second voice command is successfully identified within a set time.
  • system further includes:
  • a first voice command saving module configured to save the first voice command before reducing a current fan speed of the range hood after the number of consecutive failures of identifying the first voice command exceeds a set value
  • a query module configured to query, in the database based on the current fan speed of the range hood, whether there is a voice sample record matching the first voice command
  • the indication execution module is further configured to: if there is a voice sample record matching the first voice command in the database, perform an indication matching the voice sample record;
  • the reducing module is further configured to perform the operation of reducing a current fan speed of the range hood if there is no voice sample record matching the first voice command in the database.
  • system further includes:
  • a recovery module configured to recover if the second voice command fails to be recognized within a set time The fan speed of the hood is removed and the saved first voice command is deleted.
  • system further includes:
  • a saving module configured to perform, after the indicating that the second voice command is successful within a set time, performing an instruction matching the second voice command, if the first voice command and the first voice command If the similarity of the two voice commands is greater than the similarity threshold, the first voice command, the current fan speed of the range hood, and the second voice command are associated and saved in a database.
  • system further includes:
  • a deleting module configured to perform, after the indicating that the second voice command is successful within a set time, performing an instruction matching the second voice command, if the first voice command and the first voice command If the similarity of the two voice commands is not greater than the similarity threshold, the saved first voice command is deleted.
  • the method for improving the voice recognition rate of the range hood reduces the current fan speed of the range hood when the number of consecutive failures of the first voice command is determined to exceed the set value;
  • the second voice command is used to identify; if the second voice command is successfully identified within the set time, the technical means for performing the indication matching the second voice command is implemented, so that when the voice recognition is difficult to monitor,
  • the speech recognition rate is improved by appropriately reducing the running noise of the hood itself.
  • FIG. 1 is a schematic flow chart of a method for improving a voice recognition rate of a range hood according to Embodiment 1 of the present invention
  • FIG. 2 is a schematic flow chart of a method for improving a voice recognition rate of a range hood provided by Embodiment 2 of the present invention
  • FIG. 3 is a flow chart of a method for improving the voice recognition rate of a range hood provided by Embodiment 3 of the present invention.
  • Embodiment 4 is a schematic flow chart of a method for improving a voice recognition rate of a range hood provided by Embodiment 4 of the present invention
  • FIG. 5 is a schematic structural diagram of a system for improving the voice recognition rate of a range hood provided by Embodiment 5 of the present invention.
  • FIG. 1 is a flow chart of a method for improving a voice recognition rate of a range hood according to a first embodiment of the present invention.
  • the method is applicable to a situation in which a voice command issued by a user is recognized and a corresponding operation is performed in the process of working in the range hood. It can be performed by a system that improves the voice recognition rate of the range hood.
  • the system can pass Implemented in hardware and/or software.
  • the method specifically includes the following steps:
  • Step 110 When the number of consecutive failures of identifying the first voice command exceeds a set value, reducing the current fan speed of the range hood.
  • the first voice command and the second voice command referred to in the present invention all refer to a voice command that is input by the user to control the function of the range hood.
  • the first voice command and the second voice command may be the same or different, but only for convenience.
  • the description is defined by “first” and “second”; specifically, the voice of the user controlling the functions of the range hood by speaking, such as “turning on the light”, “starting the large exhaust fan”, “closing” or “timing” 10 minutes” and other speech utterances.
  • the voice recognition system on the range hood recognizes the voice command and responds to the voice command to achieve the purpose of controlling various functions of the range hood by voice.
  • the hood does not perform the corresponding operation according to the voice command, that is, the hood fails to recognize the voice command once; usually if the hood does not respond to the voice command, the user will issue the same voice again.
  • the command of course, the soot opportunity performs another recognition operation of the voice command.
  • the number of consecutive failures of identifying a certain voice command exceeds the set value, it is considered that there is difficulty in voice recognition at this time, and it may be a range hood.
  • the running noise of itself is too large.
  • the “number of consecutive failures of identifying the first voice command” in the present solution does not refer to the number of consecutive failures of multiple recognition of a single received voice command, but refers to the number of consecutive failures, but refers to The number of consecutive failures to identify the same voice command received multiple times.
  • the set value can be set according to design experience, for example, it can be 3 times, etc. In order to enhance the user experience, the set value is generally not set too large, and the situation that multiple voice commands are issued but no response is avoided.
  • the acoustic feature value of the voice command input by the user is extracted every time.
  • the similarity of the adjacent two acoustic feature values is higher than the preset value, the adjacent two voice commands are considered to be the same. Language Sound command.
  • the operation of the range hood itself is reduced by reducing the current fan speed of the range hood. Noise to improve speech recognition rate.
  • the speed of the fan does not give the user too complicated control mode.
  • the fan speed is selected through multiple gear positions.
  • the current fan speed of the hood is reduced, not random. Reduce the current fan speed, but reduce the fan speed to the adjacent gear.
  • Step 120 Receive a second voice command, and identify the second voice command.
  • the voice command input by the user is received again, and the received voice command is voice-recognized to achieve the purpose of responding to the voice command of the user.
  • the present invention does not limit the speech recognition algorithm used, and the speech recognition algorithms existing in the prior art can be used in the present invention.
  • Step 130 If the second voice command is successfully identified within the set time, an indication matching the second voice command is performed.
  • the voice command input by the user is successfully recognized within a reasonable set time, the user's voice command is immediately responded to, and an indication matching the voice command is performed; for example, "turning on the light” or "timing".
  • the method for improving the voice recognition rate of the range hood provides a method for reducing the current fan speed of the range hood when the number of consecutive failures of identifying the first voice command exceeds a set value;
  • the second voice command is used for identification; if the second voice command is successfully identified within the set time, the technical means for performing the indication matching the second voice command is implemented, and when the voice recognition is difficult to monitor, Reduce the fan speed of the range hood to properly reduce the running noise of the range hood itself to improve the speech recognition rate.
  • FIG. 2 is a schematic flowchart of a method for improving a voice recognition rate of a range hood according to Embodiment 2 of the present invention.
  • the embodiment is further optimized, and the benefit of optimization is that the stored The voice sample records matched by the voice command respond to the user's control commands, thereby improving the voice recognition rate.
  • the method specifically includes the following:
  • Step 210 Save the first voice command when the number of consecutive failures of identifying the first voice command exceeds a set value.
  • Step 220 Query whether there is a voice sample record matching the first voice command in the database based on the current fan speed of the hood. If yes, execute step 230. Otherwise, perform step 240.
  • the corresponding voice recognition data is “111111”, and when the current fan speed of the range hood is 100 rpm, the range hood It will generate a certain noise.
  • the corresponding voice data recognized by the first voice command issued by the user to "turn on the light” is "111100”, then if there is such a voice sample record in the database - "100- 111100-111111”, it can be judged that when the current fan speed of the range hood is 100 rpm, the first voice command identified is "111100”, and the corresponding voice command is substantially "turn on the light”, then the soot
  • the central control unit of the machine immediately performs the operation of turning on the light.
  • Step 230 Perform an indication matching the voice sample record, and end the process.
  • Step 240 Reduce the current fan speed of the range hood.
  • the purpose of reducing the current fan speed of the range hood is to reduce the noise generated by the range hood itself to a certain extent and improve the recognition rate of voice commands.
  • Step 250 Receive a second voice command, and identify the second voice command.
  • Step 260 If the second voice command is successfully identified within the set time, execute and An indication of the second voice command match.
  • the method for improving the voice recognition rate of the range hood provided by the embodiment, on the basis of the technical solution of the foregoing embodiment, when the number of consecutive failures of identifying the first voice command exceeds a set value, saving the first Voice command, and based on the current fan speed query of the range hood, whether there is a voice sample record matching the first voice command in the database, and if so, a technical means for performing an indication matching the voice sample record is implemented
  • the voice sample records stored in the database that match the voice commands respond to the user's control commands, thereby improving the voice recognition rate.
  • FIG. 3 is a schematic flowchart of a method for improving a voice recognition rate of a range hood according to Embodiment 3 of the present invention.
  • the embodiment is further optimized, and the benefit of optimization is to accurately eliminate invalid voice. command.
  • the method is as follows:
  • Step 310 Save the first voice command when the number of consecutive failures of identifying the first voice command exceeds a set value.
  • Step 320 Query whether there is a voice sample record matching the first voice command in the database based on the current fan speed of the hood. If yes, go to step 330. Otherwise, go to step 340.
  • Step 330 Perform an indication matching the voice sample record.
  • Step 340 reducing the current fan speed of the range hood.
  • Step 350 Receive a second voice command, and identify the second voice command.
  • Step 360 If the second voice command is successfully identified within the set time, step 370 is performed; otherwise, step 380 is performed.
  • Step 370 Perform an indication that matches the second voice command.
  • Step 380 Restore the fan speed of the range hood, and delete the saved first voice command.
  • the noise generated by the operation of the range hood itself has been reduced.
  • the second voice command fails to be recognized within the set time, the second voice command is considered to be An invalid voice command may be caused by other external sound interference, or it may be a dialogue between the user's family members, and is not a control command for the function of the hood.
  • the method for improving the voice recognition rate of the range hood provided by the embodiment, on the basis of the technical solution of the foregoing embodiment, recovering the hood by failing to identify the second voice command within a set time
  • the fan speed and the technical means of deleting the saved first voice command achieve the purpose of accurately eliminating invalid voice commands.
  • Embodiment 4 is a schematic flowchart of a method for improving a voice recognition rate of a range hood according to Embodiment 4 of the present invention.
  • the embodiment is further optimized, and the benefit of optimization is realizing updating the database in real time. Improve the next speech recognition rate.
  • the method is as follows:
  • Step 410 Save the first voice command when the number of consecutive failures of identifying the first voice command exceeds a set value.
  • Step 420 Query whether there is a voice sample record matching the first voice command in the database based on the current fan speed of the hood. If yes, execute step 430. Otherwise, perform step 440.
  • Step 430 Perform an indication matching the voice sample record.
  • Step 440 Reduce the current fan speed of the range hood.
  • Step 450 Receive a second voice command, and identify the second voice command.
  • Step 460 If the second voice command is successfully identified within the set time, step 470 is performed; otherwise, step 480 is performed.
  • Step 470 Perform an indication matching the second voice command, and continue to perform steps 490a and 490b.
  • Step 480 Restore the fan speed of the range hood, and delete the saved first voice command.
  • Step 490a if the similarity between the first voice command and the second voice command is greater than a similarity threshold, correlating the first voice command, the current fan speed of the range hood, and the second voice command, And save to the database.
  • the similarity between the first voice command and the second voice command is greater than a similarity threshold, if the first voice command and the second voice command are considered to be the same voice command, then the first voice is The command, the current fan speed of the range hood, and the second voice command are associated and saved to the database for use in the next speech recognition.
  • the corresponding voice recognition data is “111111”, and when the current fan speed of the range hood is 100 rpm, the first voice command corresponding to the first voice command is recognized.
  • the identification data is “111100”, and the similarity between the first voice command and the second voice command is greater than a similarity threshold, and the first voice command, the current fan speed of the range hood, and the second voice are The command is associated to obtain such a voice sample record - "100-111100-111111", and the voice sample record - "100-111100-111111” is saved in the database, so as to facilitate the next speech recognition query call to improve Speech recognition rate. If the fan speed is limited to operate in a plurality of gear positions, the first voice command and the second voice command are directly saved in association with the gear positions corresponding to the first voice command.
  • Step 490b If the similarity between the first voice command and the second voice command is not greater than a similarity threshold, deleting the saved first voice command.
  • the similarity between the first voice command and the second voice command is not greater than the similarity threshold, the first voice command and the second voice command are considered to be different voice commands, and the The first voice command is an invalid voice command and is discarded.
  • the method for improving the voice recognition rate of the range hood provided by the embodiment, on the basis of the technical solution of the foregoing embodiment, by adding “if the similarity between the first voice command and the second voice command is greater than the similarity threshold And the first voice command, the current fan speed of the range hood and the second voice command are associated, and saved in a database, otherwise the technical means of deleting the saved first voice command is achieved, real-time Update the database to improve the next speech recognition rate and accurately exclude the purpose of invalid voice commands.
  • FIG. 5 is a structural block diagram of a system for improving a voice recognition rate of a range hood according to Embodiment 5 of the present invention, and the system specifically includes the following:
  • the reducing module 510 is configured to: when the number of consecutive failures of identifying the first voice command exceeds the set value, reduce the current fan speed of the range hood; and the receiving module 520 is configured to receive the second voice command, and The second voice command is used to identify; the instruction execution module 530 is configured to: if the second voice command is successfully identified within the set time, perform an indication that matches the second voice command;
  • system may further include:
  • a first voice command saving module configured to save the first voice command before reducing a current fan speed of the range hood after the number of consecutive failures of identifying the first voice command exceeds a set value
  • a query module configured to query, in the database based on the current fan speed of the range hood, whether there is a voice sample record matching the first voice command
  • the indication execution module 530 is further configured to: if there is a voice sample record matching the first voice command in the database, perform an indication matching the voice sample record;
  • the reducing module 510 is further configured to perform the operation of reducing the current fan speed of the range hood if there is no voice sample record matching the first voice command in the database.
  • system may further include:
  • a recovery module configured to restore a fan speed of the range hood if the recognition of the second voice command fails within a set time, and delete the saved first voice command.
  • system may further include:
  • a saving module configured to perform, after the indicating that the second voice command is successful within a set time, performing an instruction matching the second voice command, if the first voice command and the first voice command If the similarity of the two voice commands is greater than the similarity threshold, the first voice command, the current fan speed of the range hood, and the second voice command are associated and saved in a database.
  • system may further include:
  • a deleting module configured to perform, after the indicating that the second voice command is successful within a set time, performing an instruction matching the second voice command, if the first voice command and the first voice command If the similarity of the two voice commands is not greater than the similarity threshold, the saved first voice command is deleted.
  • the system for improving the voice recognition rate of the range hood reduces the current fan speed of the range hood when the number of consecutive failures of the first voice command is determined to exceed the set value;
  • the second voice command performs identification; if the second voice command is successfully identified within the set time, the technical means for performing the indication matching the second voice command is implemented When it is found that the speech recognition is difficult, the speech recognition rate is improved by appropriately reducing the running noise of the hood itself.
  • the above product can perform the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
  • a program instructing related hardware may be completed by a program instructing related hardware, and the program is stored in a storage medium, and includes a plurality of instructions for making a device (which may be a single chip microcomputer). , a chip, etc. or a processor performs all or part of the steps of the methods described in various embodiments of the present application.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program codes.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Combustion & Propulsion (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Air Conditioning Control Device (AREA)
  • Ventilation (AREA)

Abstract

一种提高油烟机语音识别率的方法及***，该方法具体包括如下步骤：当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；然后接收第二语音命令，并对第二语音命令进行识别；若在设定的时间内对第二语音命令识别成功，则执行与第二语音命令匹配的指示。实现该方法的***包括：减小模块（510）、接收模块（520）和指示执行模块（530）。其中，减小模块（510）用于当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；接收模块（520）用于接收第二语音命令，并对第二语音命令进行识别；指示执行模块（530）用于若在设定的时间内对第二语音命令识别成功，则执行与第二语音命令匹配的指示。该方法和***实现了当监测到语音识别困难时，通过适当减小油烟机自身的运行噪音来提高语音识别率。

Description

一种提高油烟机语音识别率的方法及*** 技术领域
本发明实施例涉及语音识别技术领域，尤其涉及一种提高油烟机语音识别率的方法及***。
背景技术
随着电子技术以及语音识别技术的飞速发展，语音识别技术已经广泛应用于各行各业，近些年来越来越多的语音控制油烟机走进了人们的家庭生活，实现了可以使用不同的语种以及语音命令控制油烟机的各项功能，例如，通过语音命令控制油烟机的开关、风机档位以及定时等功能，极大地方便了人们的生活，同时也省去了油烟机上以往的功能按钮，特别方便了老年人使用。
但是，由于油烟机本身运行时噪音比较大，较大的运行噪音势必影响对用户发出的语音命令的识别，导致语音识别率下降。为了改善上述问题，现有技术中通常通过提高语音识别算法的复杂度或者根据运行的噪音来训练语音识别模型，但是如果语音识别算法及其复杂，势必会增加CPU(Central Processing Unit，中央处理器)的工作负担，影响设备的响应速度，并且增加了设备的功耗；而对于通过运行的噪音来训练语音识别模型的方法，通常油烟机都设置有多个运行转速，在不同的厨房环境中运行时产生的噪音也不同，因此训练得到的语音识别模型未必适合油烟机实际的使用环境。
发明内容
本发明实施例提供了一种提高油烟机语音识别率的方法及***，以实现当监测到语音识别困难时，通过适当减小油烟机自身的运行噪音来提高语音识别率。
第一方面，本发明实施例提供了一种提高油烟机语音识别率的方法，该方法包括：
当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；
接收第二语音命令，并对所述第二语音命令进行识别；
若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示。
进一步地，所述当对第一语音命令进行识别连续失败的次数超过设定值时之后，减小油烟机当前的风机转速之前，还包括：
保存所述第一语音命令；
基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行与所述语音样本记录匹配的指示，否则，执行所述减小油烟机当前的风机转速的操作。
进一步地，所述方法还包括：若在设定的时间内对所述第二语音命令识别失败，则恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
进一步地，所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，还包括：
若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联， 并保存到数据库中；
若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
第二方面，本发明实施例还提供了一种提高油烟机语音识别率的***，该***包括：
减小模块，用于当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；
接收模块，用于接收第二语音命令，并对所述第二语音命令进行识别；
指示执行模块，用于若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示。
进一步地，所述***还包括：
第一语音命令保存模块，用于在所述当对第一语音命令进行识别连续失败的次数超过设定值时之后，减小油烟机当前的风机转速之前，保存所述第一语音命令；
查询模块，用于基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录；
对应地，所述指示执行模块，还用于若所述数据库中存在与所述第一语音命令匹配的语音样本记录，则执行与所述语音样本记录匹配的指示；
所述减小模块，还用于若所述数据库中不存在与所述第一语音命令匹配的语音样本记录，执行所述减小油烟机当前的风机转速的操作。
进一步地，所述***还包括：
恢复模块，用于若在设定的时间内对所述第二语音命令识别失败，则恢复 所述油烟机的风机转速，并删除保存的所述第一语音命令。
进一步地，所述***还包括：
保存模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中。
进一步地，所述***还包括：
删除模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
本发明实施例提供的一种提高油烟机语音识别率的方法，通过当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；之后对接收到的第二语音命令进行识别；若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示的技术手段，实现了当监测到语音识别困难时，通过适当减小油烟机自身的运行噪音来提高语音识别率。
附图说明
图1是本发明实施例一提供的一种提高油烟机语音识别率的方法流程示意图；
图2是本发明实施例二提供的一种提高油烟机语音识别率的方法流程示意图；
图3是本发明实施例三提供的一种提高油烟机语音识别率的方法流程示意 图；
图4是本发明实施例四提供的一种提高油烟机语音识别率的方法流程示意图；
图5是本发明实施例五提供的一种提高油烟机语音识别率的***结构示意图。
具体实施方式
下面结合附图和实施例对本发明作进一步的详细说明。可以理解的是，此处所描述的具体实施例仅仅用于解释本发明，而非对本发明的限定。另外还需要说明的是，为了便于描述，附图中仅示出了与本发明相关的部分而非全部结构。
在更加详细地讨论示例性实施例之前应当提到的是，一些示例性实施例被描述成作为流程图描绘的处理或方法。虽然流程图将各项步骤描述成顺序的处理，但是其中的许多步骤可以被并行地、并发地或者同时实施。此外，各项步骤的顺序可以被重新安排。当其步骤完成时所述处理可以被终止，但是还可以具有未包括在附图中的附加步骤。所述处理可以对应于方法、函数、规程、子例程、子程序等等。
实施例一
图1为本发明实施例一提供的一种提高油烟机语音识别率的方法流程图，该方法可适用于在油烟机工作的过程中，对用户发出的语音命令进行识别并执行相应操作的情况，可以由提高油烟机语音识别率的***来执行。该***可通 过硬件和/或软件的方式实现。该方法具体包括如下步骤：
步骤110、当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速。
其中，本发明所指的第一语音命令以及第二语音命令均指用户输入的控制油烟机功能的语音命令，所述第一语音命令以及第二语音命令可能相同，也可能不同，只是为了方便描述才通过“第一”、“第二”进行限定；具体可以是用户通过说话发出的控制油烟机各项功能的语音，例如“打开灯光”、“启动大排风扇”、“关闭”或者“定时10分钟”等语音话语。当用户发出语音命令，油烟机上的语音识别***会对所述语音命令进行识别，并响应所述语音命令，以实现通过语音控制油烟机各项功能的目的。如果用户发出某一语音命令后，油烟机没有根据所述语音命令执行相应的操作即认为油烟机对所述语音命令识别失败一次；通常如果油烟机不响应语音命令，用户会再次发出相同的语音命令，当然油烟机会再执行一次对所述语音命令的识别操作，当对某一相同语音命令进行识别连续失败的次数超过设定值时，即认为此时存在语音识别困难，有可能是油烟机本身的运行噪声太大导致。由此可以理解的是，本方案中所说的“对第一语音命令进行识别连续失败的次数”并不是指对单次接收到的语音命令进行多次识别而连续失败的次数，而是指对多次接收到的同一语音命令进行识别而连续失败的次数。所述设定值可以根据设计经验去设定，例如可以是3次等，为增强用户体验，设定值一般不会设置太大，避免出现多次发出语音命令却无响应的情况。
示例性地，提取每次用户输入的语音命令的声学特征值，当相邻两次的声学特征值的相似度高于预设值时，即认为所述相邻两次的语音命令为相同的语 音命令。如果用户多次(如两次)输入相同的语音命令，且油烟机均不对其进行响应，则认为此时存在语音识别困难，通过减小油烟机当前的风机转速来减小油烟机本身的运行噪音，以提高语音识别率。在油烟机的运行过程中，风机的转速并不会给用户太复杂的控制方式，一般通过多个档位来选择风机转速，与此对应的，在减小油烟机当前的风机转速，不是随机减小当前的风机转速，而是减小到相邻档位对应的风机转速。
步骤120、接收第二语音命令，并对所述第二语音命令进行识别。
当减小油烟机当前的风机转速后，再次接收用户输入的语音命令，并对接收到的语音命令进行语音识别，以实现能够响应用户的语音命令的目的。需要说明的是，本发明并不对采用的语音识别算法进行限定，现有技术中存在的语音识别算法均可以用到本发明中。
步骤130、若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示。
如果在合理的设定时间内对用户输入的语音命令识别成功，则立即响应用户的语音命令，执行与所述语音命令匹配的指示；例如“打开灯光”或者“定时”等操作。
本实施例提供的一种提高油烟机语音识别率的方法，通过当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；之后对接收到的第二语音命令进行识别；若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示的技术手段，实现了当监测到语音识别困难时，通过减小油烟机的风机转速来适当减小油烟机自身的运行噪音来提高语音识别率。
实施例二
图2为本发明实施例二提供的一种提高油烟机语音识别率的方法流程示意图，在实施例一的基础上，本实施例进行了进一步优化，优化的好处是通过数据库中已存储的与语音命令匹配的语音样本记录来响应用户的控制命令，进而提高语音识别率。具体参见图2所示，该方法具体包括如下：
步骤210、当对第一语音命令进行识别连续失败的次数超过设定值时，保存所述第一语音命令。
步骤220、基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行步骤230，否则，执行步骤240。
例如，假设用户在安静的环境中发出的所述第一语音命令为“打开灯光”时，对应的语音识别数据为“111111”，当油烟机当前的风机转速为100转/分时，油烟机本身会产生一定的噪音，在这种环境下用户发出“打开灯光”的第一语音命令被识别出的对应语音数据为“111100”，那么如果数据库中存在这样的语音样本记录——“100-111100-111111”，则可以判断在油烟机当前的风机转速为100转/分时，识别到的所述第一语音命令为“111100”时，对应的语音命令实质为“打开灯光”，则油烟机的中央控制机构立即执行打开灯光的操作。
步骤230、执行与所述语音样本记录匹配的指示，结束流程。
步骤240、减小油烟机当前的风机转速。
减小油烟机当前的风机转速的目的是在一定程度减小油烟机本身产生的噪声，提高对语音命令的识别率。
步骤250、接收第二语音命令，并对所述第二语音命令进行识别。
步骤260、若在设定的时间内对所述第二语音命令识别成功，则执行与所 述第二语音命令匹配的指示。
本实施例提供的一种提高油烟机语音识别率的方法，在上述实施例技术方案的基础上，通过当对第一语音命令进行识别连续失败的次数超过设定值时，保存所述第一语音命令，并基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行与所述语音样本记录匹配的指示的技术手段，实现了通过数据库中已存储的与语音命令匹配的语音样本记录来响应用户的控制命令，进而提高了语音识别率。
实施例三
图3为本发明实施例三提供的一种提高油烟机语音识别率的方法流程示意图，在上述实施例的基础上，本实施例进行了进一步优化，优化的好处是实现了准确排除无效的语音命令。具体参见图3所示，该方法具体如下：
步骤310、当对第一语音命令进行识别连续失败的次数超过设定值时，保存所述第一语音命令。
步骤320、基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行步骤330，否则，执行步骤340。
步骤330、执行与所述语音样本记录匹配的指示。
步骤340、减小油烟机当前的风机转速。
步骤350、接收第二语音命令，并对所述第二语音命令进行识别。
步骤360、若在设定的时间内对所述第二语音命令识别成功，则执行步骤370，否则，执行步骤380。
步骤370、执行与所述第二语音命令匹配的指示。
步骤380、恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
在减小油烟机的风机转速之后，已经降低了油烟机本身运行产生的噪声，此时，如果在设定的时间内对所述第二语音命令识别失败，则认为所述第二语音命令为一个无效的语音命令，可能是外界的其他声音干扰，也可能是用户的家庭成员之间的对话，并不是对油烟机功能的控制命令。
本实施例提供的一种提高油烟机语音识别率的方法，在上述实施例技术方案的基础上，通过若在设定的时间内对所述第二语音命令识别失败，则恢复所述油烟机的风机转速，并删除保存的所述第一语音命令的技术手段，达到了准确排除无效的语音命令的目的。
实施例四
图4为本发明实施例四提供的一种提高油烟机语音识别率的方法流程示意图，在上述实施例的基础上，本实施例进行了进一步优化，优化的好处是实现了实时更新数据库，以提高下次的语音识别率。具体参见图4所示，该方法具体如下：
步骤410、当对第一语音命令进行识别连续失败的次数超过设定值时，保存所述第一语音命令。
步骤420、基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行步骤430，否则，执行步骤440。
步骤430、执行与所述语音样本记录匹配的指示。
步骤440、减小油烟机当前的风机转速。
步骤450、接收第二语音命令，并对所述第二语音命令进行识别。
步骤460、若在设定的时间内对所述第二语音命令识别成功，则执行步骤470，否则，执行步骤480。
步骤470、执行与所述第二语音命令匹配的指示，并继续执行步骤490a和步骤490b。
步骤480、恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
步骤490a、若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中。
若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则认为所述第一语音命令与所述第二语音命令为相同的语音命令，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中，以备下次进行语音识别时使用。
例如，所述第二语音命令为“打开灯光”时，对应的语音识别数据为“111111”，当油烟机当前的风机转速为100转/分时，识别到的所述第一语音命令对应的识别数据为“111100”，且所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联得到这样的语音样本记录——“100-111100-111111”，将所述语音样本记录——“100-111100-111111”保存到数据库中，方便下次语音识别时查询调用，以提高语音识别率。如果风机转速限定在多个档位运转，那么第一语音命令和第二语音命令直接与第一语音命令对应的档位关联保存即可。
步骤490b、若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则认为所述第一语音命令与所述第二语音命令是不相同的语音命令，此时认为所述第一语音命令为无效的语音命令，进行舍弃。
本实施例提供的一种提高油烟机语音识别率的方法，在上述实施例技术方案的基础上，通过增加“若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中，否则删除保存的所述第一语音命令”的技术手段，达到了实时更新数据库，以提高下次的语音识别率，且准确排除无效的语音命令的目的。
实施例五
图5为本发明实施例五提供的一种提高油烟机语音识别率的***的结构框图，该***具体包括如下：
减小模块510、接收模块520和指示执行模块530；
其中，减小模块510，用于当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；接收模块520，用于接收第二语音命令，并对所述第二语音命令进行识别；指示执行模块530，用于若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示；
进一步地，所述***还可以包括：
第一语音命令保存模块，用于在所述当对第一语音命令进行识别连续失败的次数超过设定值时之后，减小油烟机当前的风机转速之前，保存所述第一语音命令；
查询模块，用于基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录；
对应地，指示执行模块530，还用于若所述数据库中存在与所述第一语音命令匹配的语音样本记录，则执行与所述语音样本记录匹配的指示；
减小模块510，还用于若所述数据库中不存在与所述第一语音命令匹配的语音样本记录，执行所述减小油烟机当前的风机转速的操作。
进一步地，所述***还可以包括：
恢复模块，用于若在设定的时间内对所述第二语音命令识别失败，则恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
进一步地，所述***还可以包括：
保存模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中。
进一步地，所述***还可以包括：
删除模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
本发明实施例提供的一种提高油烟机语音识别率的***，通过当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；之后对接收到的第二语音命令进行识别；若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示的技术手段，实现了当监 测到语音识别困难时，通过适当减小油烟机自身的运行噪音来提高语音识别率。
上述产品可执行本发明任意实施例所提供的方法，具备执行方法相应的功能模块和有益效果。未在本实施例中详尽描述的技术细节，可参见本发明任意实施例所提供的方法。
本领域技术人员可以理解实现上述实施例方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序存储在一个存储介质中，包括若干指令用以使得一个设备(可以是单片机，芯片等)或处理器(processor)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，RandomAccess Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
注意，上述仅为本发明的较佳实施例及所运用技术原理。本领域技术人员会理解，本发明不限于这里所述的特定实施例，对本领域技术人员来说能够进行各种明显的变化、重新调整和替代而不会脱离本发明的保护范围。因此，虽然通过以上实施例对本发明进行了较为详细的说明，但是本发明不仅仅限于以上实施例，在不脱离本发明构思的情况下，还可以包括更多其他等效实施例，而本发明的范围由所附的权利要求范围决定。

Claims (10)

  1. 一种提高油烟机语音识别率的方法，其特征在于，包括：
    当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；
    接收第二语音命令，并对所述第二语音命令进行识别；
    若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示。
  2. 根据权利要求1所述的方法，其特征在于，所述当对第一语音命令进行识别连续失败的次数超过设定值时之后，减小油烟机当前的风机转速之前，还包括：
    保存所述第一语音命令；
    基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录，若存在，则执行与所述语音样本记录匹配的指示，否则，执行所述减小油烟机当前的风机转速的操作。
  3. 根据权利要求2所述的方法，其特征在于，还包括：
    若在设定的时间内对所述第二语音命令识别失败，则恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
  4. 根据权利要求3所述的方法，其特征在于，其特征在于，所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，还包括：
    若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中。
  5. 根据权利要求3所述的方法，其特征在于，其特征在于，所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，还包括：
    若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
  6. 一种提高油烟机语音识别率的***，其特征在于，包括：
    减小模块，用于当对第一语音命令进行识别连续失败的次数超过设定值时，减小油烟机当前的风机转速；
    接收模块，用于接收第二语音命令，并对所述第二语音命令进行识别；
    指示执行模块，用于若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示。
  7. 根据权利要求6所述的***，其特征在于，还包括：
    第一语音命令保存模块，用于在所述当对第一语音命令进行识别连续失败的次数超过设定值时之后，减小油烟机当前的风机转速之前，保存所述第一语音命令；
    查询模块，用于基于油烟机当前的风机转速查询数据库中是否存在与所述第一语音命令匹配的语音样本记录；
    对应地，所述指示执行模块，还用于若所述数据库中存在与所述第一语音命令匹配的语音样本记录，则执行与所述语音样本记录匹配的指示；
    所述减小模块，还用于若所述数据库中不存在与所述第一语音命令匹配的语音样本记录，执行所述减小油烟机当前的风机转速的操作。
  8. 根据权利要求7所述的***，其特征在于，还包括：
    恢复模块，用于若在设定的时间内对所述第二语音命令识别失败，则恢复所述油烟机的风机转速，并删除保存的所述第一语音命令。
  9. 根据权利要求8所述的***，其特征在于，其特征在于，还包括：
    保存模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度大于相似度阈值，则将所述第一语音命令、油烟机当前的风机转速以及所述第二语音命令进行关联，并保存到数据库中。
  10. 根据权利要求8所述的***，其特征在于，其特征在于，还包括：
    删除模块，用于在所述若在设定的时间内对所述第二语音命令识别成功，则执行与所述第二语音命令匹配的指示之后，若所述第一语音命令与所述第二语音命令的相似度不大于相似度阈值，则删除保存的所述第一语音命令。
PCT/CN2016/113278 2016-10-14 2016-12-29 一种提高油烟机语音识别率的方法及*** WO2018068422A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610899191.9 2016-10-14
CN201610899191.9A CN106439967B (zh) 2016-10-14 2016-10-14 一种提高油烟机语音识别率的方法及***

Publications (1)

Publication Number Publication Date
WO2018068422A1 true WO2018068422A1 (zh) 2018-04-19

Family

ID=58175056

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/113278 WO2018068422A1 (zh) 2016-10-14 2016-12-29 一种提高油烟机语音识别率的方法及***

Country Status (2)

Country Link
CN (1) CN106439967B (zh)
WO (1) WO2018068422A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3879184A1 (en) * 2020-03-10 2021-09-15 BSH Hausgeräte GmbH Noise reduction method and apparatus for household appliance, household appliance, and cooker hood
US20220180865A1 (en) * 2020-12-03 2022-06-09 International Business Machines Corporation Runtime topic change analyses in spoken dialog contexts

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106875945B (zh) * 2017-03-09 2020-06-26 广东美的制冷设备有限公司 语音控制方法、装置和空调器
CN111076245B (zh) * 2018-10-19 2021-07-23 宁波方太厨具有限公司 一种油烟机智能降噪方法
CN112086096B (zh) * 2019-06-14 2024-04-05 北京京东尚科信息技术有限公司 数据处理方法、装置、***、介质

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030163325A1 (en) * 2002-02-27 2003-08-28 Jens Maase Electrical household appliance and methods for testing and for initializing a voice operating unit therein
EP1752712A1 (en) * 2005-08-10 2007-02-14 Best S.p.a. Hood with telephonic communication functions
CN202561865U (zh) * 2012-05-08 2012-11-28 成都众询科技有限公司 一种语音控制抽油烟机
CN103673008A (zh) * 2013-09-16 2014-03-26 宁波方太厨具有限公司 一种智能吸油烟机以及该吸油烟机的控制方法
CN104464729A (zh) * 2014-12-16 2015-03-25 佛山市顺德区美的电热电器制造有限公司 语音控制电器及其语音控制方法
CN104505092A (zh) * 2014-12-10 2015-04-08 广东美的制冷设备有限公司 空调器语音控制方法和***
CN104517607A (zh) * 2014-12-16 2015-04-15 佛山市顺德区美的电热电器制造有限公司 滤除语音控制电器中的噪声的方法及语音控制电器
CN104992709A (zh) * 2015-06-25 2015-10-21 广州视源电子科技股份有限公司 一种语音指令的执行方法及语音识别设备
CN105653233A (zh) * 2015-12-30 2016-06-08 芜湖美智空调设备有限公司 关联语音信号与控制指令的方法及控制终端

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4622124B2 (ja) * 2001-03-13 2011-02-02 株式会社デンソー ナビゲーション装置
JP2009229899A (ja) * 2008-03-24 2009-10-08 Toshiba Corp 音声認識装置および音声認識方法
JP2009230620A (ja) * 2008-03-25 2009-10-08 Panasonic Corp 情報処理装置
CN102761721A (zh) * 2011-04-27 2012-10-31 海尔集团公司 电视机音量的控制方法及控制***
US20130332159A1 (en) * 2012-06-08 2013-12-12 Apple Inc. Using fan throttling to enhance dictation accuracy
CN104715752B (zh) * 2015-04-09 2019-01-08 刘文军 语音识别方法、装置及***
CN205208686U (zh) * 2015-12-22 2016-05-04 嵊州市乐普斯电器有限公司 一种语音输入控制吸油烟机

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030163325A1 (en) * 2002-02-27 2003-08-28 Jens Maase Electrical household appliance and methods for testing and for initializing a voice operating unit therein
EP1752712A1 (en) * 2005-08-10 2007-02-14 Best S.p.a. Hood with telephonic communication functions
CN202561865U (zh) * 2012-05-08 2012-11-28 成都众询科技有限公司 一种语音控制抽油烟机
CN103673008A (zh) * 2013-09-16 2014-03-26 宁波方太厨具有限公司 一种智能吸油烟机以及该吸油烟机的控制方法
CN104505092A (zh) * 2014-12-10 2015-04-08 广东美的制冷设备有限公司 空调器语音控制方法和***
CN104464729A (zh) * 2014-12-16 2015-03-25 佛山市顺德区美的电热电器制造有限公司 语音控制电器及其语音控制方法
CN104517607A (zh) * 2014-12-16 2015-04-15 佛山市顺德区美的电热电器制造有限公司 滤除语音控制电器中的噪声的方法及语音控制电器
CN104992709A (zh) * 2015-06-25 2015-10-21 广州视源电子科技股份有限公司 一种语音指令的执行方法及语音识别设备
CN105653233A (zh) * 2015-12-30 2016-06-08 芜湖美智空调设备有限公司 关联语音信号与控制指令的方法及控制终端

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3879184A1 (en) * 2020-03-10 2021-09-15 BSH Hausgeräte GmbH Noise reduction method and apparatus for household appliance, household appliance, and cooker hood
US20220180865A1 (en) * 2020-12-03 2022-06-09 International Business Machines Corporation Runtime topic change analyses in spoken dialog contexts

Also Published As

Publication number Publication date
CN106439967A (zh) 2017-02-22
CN106439967B (zh) 2019-01-04

Similar Documents

Publication Publication Date Title
WO2018068422A1 (zh) 一种提高油烟机语音识别率的方法及***
US9818407B1 (en) Distributed endpointing for speech recognition
EP3522153B1 (en) Voice control system, wakeup method and wakeup apparatus therefor, electrical appliance and co-processor
CN109410952B (zh) 一种语音唤醒方法、装置及***
TWI474317B (zh) 訊號處理裝置以及訊號處理方法
JP6200516B2 (ja) 発話認識電力管理
WO2017206725A1 (zh) 一种智能冰箱、服务器及语言控制***和方法
GB2552082A (en) Voice user interface
WO2019119431A1 (zh) 人机交互方法、装置、终端设备及计算机可读存储介质
CN112382285B (zh) 语音控制方法、装置、电子设备和存储介质
CN110838296B (zh) 录音过程的控制方法、***、电子设备和存储介质
US20190228773A1 (en) Speech interaction method, apparatus and computer readable storage medium
CN111667818A (zh) 一种训练唤醒模型的方法及装置
WO2016124048A1 (zh) 一种应用程序启动方法和电子设备
US20190302869A1 (en) Information processing method and electronic device
TWI557722B (zh) 語音干擾的濾除方法、系統，與電腦可讀記錄媒體
KR20230002690A (ko) 발화의 음성 인식 오류 교정
US20220068267A1 (en) Method and apparatus for recognizing speech, electronic device and storage medium
JP7208951B2 (ja) 音声対話の方法、装置、デバイス及びコンピュータ読み取り可能な記憶媒体
CN103543814A (zh) 信号处理装置以及信号处理方法
CN105609103A (zh) 一种语音即时识别***
JP2024062993A (ja) 自動音声認識処理結果の減衰
CN112207811B (zh) 一种机器人控制方法、装置、机器人及存储介质
CN109065026B (zh) 一种录音控制方法及装置
WO2021169711A1 (zh) 指令执行方法、装置、存储介质及电子设备

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: 16918571

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 04/10/2019)

122 Ep: pct application non-entry in european phase

Ref document number: 16918571

Country of ref document: EP

Kind code of ref document: A1