WO2015096504A1 - Voice command correcting method, correcting system and electronic device - Google Patents

Abstract

A voice command correcting method, correcting system and electronic device, the voice command correcting method comprising: obtaining a voice command to be corrected in a voice characteristic library (110); prompting a user to input voice information corresponding to the voice command (120); acquiring voice information inputted by the user (130); obtaining voice characteristic information according to the acquired voice information (140); according to the obtained voice characteristic information, updating the voice characteristic information of the voice command to be corrected so as to correct the voice command (150).

Description

 Method for correcting voice command, correction system and electronic device

This application claims priority to Chinese Patent Application No. 201310719786.8, entitled "A Voice Command Correction Method, Calibration System, and Electronic Equipment", filed on December 23, 2013, the entire contents of which are hereby incorporated by reference. The citations are incorporated herein by reference.

Technical field

 The present invention relates to the field of electronic technologies, and in particular, to a method for correcting voice commands, a correction system, and an electronic device. Background technique

 A human-computer interaction method in which voice command operation is very common is also a very natural way of human-computer interaction. With the increasing processing power of user terminal devices (such as smart phones, tablets, iPads, etc.) in recent years, the application of voice command operations on user terminal devices has become a trend.

 When receiving the input voice information, the electronic device extracts the voice feature information according to the input voice information, and then searches for a voice command corresponding to the extracted voice feature information in the voice feature database, and executes the voice command to complete the voice. Command operation.

Because the voice feature library is the key data that distinguishes different users and distinguishes different commands, it is the personalized data of the user's "exclusive", and is also affected by many factors such as hardware conditions and usage environment of the user terminal equipment. Therefore, after the user terminal device is replaced or the voice feature library is upgraded, the voice feature information of some or all voice commands of the voice feature library may be deviated from the corresponding voice feature information of the device set, so that when the voice command is operated, there will be The case of identifying inaccurate or incapable of recognizing the input voice information. Summary of the invention

 The present invention provides a voice command correction method, a correction system, and an electronic device, which solves the voice feature information of some or all voice commands of the voice feature library after the user terminal device is replaced or the voice feature library is upgraded in the prior art. The corresponding voice feature information of the device set may be deviated, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 A first aspect of the embodiments of the present invention provides a method for correcting a voice command, where the method includes: obtaining a voice command that needs to be corrected in a voice feature database; prompting a user to input voice information corresponding to the voice command; The voice information input by the user; obtaining the voice feature information according to the collected voice information; updating the voice feature information of the voice command to be corrected according to the obtained voice feature information, to correct the voice command.

 In a first possible implementation manner of the first aspect, the obtaining the voice command that needs to be corrected in the voice feature database includes:

 Receiving the voice feature library;

 Determining a need according to a specific voice command in the voice feature library preset by the system, or according to a frequency of use of the voice command in the voice feature library, or a voice command selected in the voice feature library according to input information of the user The corrected voice command is a voice command that needs to be corrected each time the voice feature library is received.

 With reference to the first aspect, in a second possible implementation manner of the first aspect, the obtaining a voice command that needs to be corrected in the voice feature library includes:

 The updated voice command and its voice feature information are received, and the updated voice command is determined to be a voice command that needs to be corrected.

 With reference to the first aspect, the first or the second possible implementation manner of the first aspect, in a third possible implementation manner of the first aspect, the updating the voice command to be corrected according to the obtained voice feature information The voice feature information is specifically: replacing the obtained voice feature information with the voice feature information of the voice command that needs to be corrected.

In combination with any of the first to third possible implementations of the first aspect, the fourth aspect is In an implementation manner, the prompting user inputs voice information corresponding to the voice command, specifically:

 Generating voice information according to voice feature information of the voice command and playing; or

 Playing the machine sound corresponding to the voice command; or

 Display text or image information corresponding to the voice command.

 A second aspect of the embodiments of the present invention provides a method for correcting a voice command, where the method includes: obtaining a voice command set that needs to be corrected in a voice feature database, where the voice command set includes M voice commands, and the M is greater than An integer equal to 2;

 Performing on each of the first to Nth voice commands in the voice command set: prompting the user to input voice information corresponding to the first voice command, collecting voice information input by the user, according to the set Obtaining voice feature information, updating and updating voice feature information of the first voice command in the voice feature database according to the obtained voice feature information to correct the first voice command, where N is smaller than the M An integer greater than 0, the first voice command being any one of the first to Nth voice commands;

 Performing on each of the N+1th to Mth voice commands in the voice command set: N of the pre-corrected and corrected voice feature information according to the corrected N first voice commands a deviation value, determining a manner of correcting the second voice command, correcting the second voice command according to the determined correction mode, where the second voice command is the (N+1)th to the Mth Any of the voice commands.

 In a first possible implementation manner of the second aspect, the determining, according to the N deviation values of the pre-corrected and corrected speech feature information, the N first voice commands that have been corrected are determined to be the second voice command. The way to correct, including:

 If the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, the change rule is searched according to the N deviation values;

If the change rule is found, determining to update the voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining to perform the second voice command: prompting user input The voice information corresponding to the second voice command is used for the voice information The voice information input by the user obtains the voice feature information according to the collected voice information, and updates the voice feature information of the second voice command according to the obtained voice feature information to correct the second voice command.

 With reference to the first possible implementation manner of the second aspect, in a second possible implementation manner of the second aspect, the N, before and after the correction, based on the N first voice commands that have been corrected Determining the manner of correcting the second voice command further includes: if the N deviation values are less than the first threshold, discarding the correction of the second voice command.

 With reference to the first possible implementation manner of the second aspect, in a third possible implementation manner of the second aspect, the N of the pre-corrected and corrected voice feature information according to the N first voice commands that have been corrected And determining, by the deviation value, the manner of correcting the second voice command, further comprising: if the N deviation values are greater than the second threshold, re-determining the voice command set that needs to be corrected in the voice feature database .

 A third aspect of the embodiments of the present invention provides a voice command correction system, where the system includes: a determining unit, configured to obtain a voice command in a voice feature library that needs to be corrected; and a prompting unit, configured to prompt a user to input the voice a voice information corresponding to the command; a set unit for collecting voice information input by the user; an obtaining unit, configured to obtain voice feature information according to the collected voice information; and a correcting unit, configured to use the obtained voice feature information The voice feature information of the voice command to be corrected is updated to correct the voice command.

 In a first possible implementation manner of the third aspect, the determining unit is specifically configured to: receive the voice feature database; according to a specific voice command in the voice feature library preset by the system, or according to the voice The frequency of use of the voice command in the feature library, or the voice command selected in the voice feature library according to the input information of the user determines a voice command to be corrected, and the specific voice command is a voice that needs to be corrected after receiving the voice feature library each time. command.

 With reference to the third aspect, in a second possible implementation manner of the third aspect, the determining unit is specifically configured to:

Receiving the updated voice command and its voice feature information, and determining the updated voice command as a voice command that needs to be corrected. With reference to the third aspect, the first or the second possible implementation manner of the third aspect, in the third possible implementation manner of the third aspect, the correcting unit is specifically configured to: replace the obtained voice feature information with the The voice feature information of the voice command that needs to be corrected.

 With reference to the third aspect, the first, the second or the third possible implementation manner of the third aspect, in a fourth possible implementation manner of the third aspect, the prompting unit is specifically configured to: according to the voice The voice feature information of the command generates voice information and plays, or plays a machine sound corresponding to the voice command; or displays text or image information corresponding to the voice command.

 A fourth aspect of the embodiments of the present invention provides a voice command correction system, where the system includes: an obtaining unit, configured to obtain a voice command set that needs to be corrected in a voice feature database, where the voice command set includes M voice commands. The M is an integer greater than or equal to 2;

 a first correcting unit, configured to perform, for each of the first to Nth voice commands in the voice command set: prompting the user to input voice information corresponding to the first voice command, and collecting the user input And the voice information is obtained according to the collected voice information, and the sound feature information of the first voice command in the voice feature database is updated according to the obtained voice feature information to correct the first voice command, N is an integer smaller than the M is greater than 0, and the first voice command is any one of the first to Nth voice commands;

 a second correcting unit, configured to perform, for each of the N+1th to Mth voice commands in the voice command set: before and after correction according to the N first voice commands that have been corrected And determining, by the N deviation values of the following voice feature information, a manner of correcting the second voice command, and correcting the second voice command according to the determined correction mode, where the second voice command is the N+1 to any one of the M voice commands.

 With reference to the fourth aspect, in a first possible implementation manner of the fourth aspect, the second calibration unit is specifically configured to:

 If the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, the change rule is searched according to the N deviation values;

If the change rule is found, determining to update the voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining the second voice Command execution: prompting a user to input voice information corresponding to the second voice command, collecting voice information input by the user, obtaining voice feature information according to the collected voice information, and updating the voice information according to the obtained voice feature information Two voice command information of the voice command to correct the second voice command.

 With reference to the first possible implementation manner of the fourth aspect, in a second possible implementation manner of the fourth aspect, the second correcting unit is further configured to: if the N deviation values are smaller than the first threshold, The correction of the second voice command is then abandoned.

 With the first possible implementation of the fourth aspect, in a third possible implementation manner of the fourth aspect, the determining unit is further configured to: if the N deviation values are greater than the second threshold, Re-determining the set of voice commands in the speech feature library that need to be corrected.

 A fifth aspect of the embodiments of the present invention provides an electronic device, where the electronic device includes: a memory, configured to store a voice feature database, the voice feature library includes a plurality of voice commands, and a processor, configured to obtain a voice feature library. a voice command that needs to be corrected, obtains voice feature information according to the collected voice information, and updates voice feature information of the voice command to be corrected according to the obtained voice feature information to correct the voice command; And a method for prompting a user to input voice information corresponding to the voice command; and inputting means, configured to collect voice information input by the user.

 In a first possible implementation manner of the fifth aspect, when the electronic device receives the voice feature database, the processor is specifically configured to: according to a specific voice command in the voice feature library preset by the system Or determining a voice command to be corrected according to a frequency of use of the voice command in the voice feature library, or a voice command selected in the voice feature library according to input information of the user, where the specific voice command is a voice feature received each time. Voice commands that need to be corrected after the library.

 With reference to the fifth aspect, in a second possible implementation manner of the fifth aspect, when the electronic device receives the second voice feature database, the processor is further configured to: receive the updated voice command and the voice feature thereof Information, and determining the updated voice command as a voice command that needs to be corrected.

With reference to the fifth aspect, the first or the second possible implementation manner of the fifth aspect, in a third possible implementation manner of the fifth aspect, the processor is specifically configured to: obtain the obtained voice feature information The voice feature information of the voice command to be corrected is replaced. With reference to the fifth aspect, the first, the second or the third possible implementation manner of the fifth aspect, in a fourth possible implementation manner of the fifth aspect, the output device is specifically a voice player or a display screen; the voice player is configured to generate voice information according to the voice feature information of the voice command and play, or play the machine sound corresponding to the voice command; the display screen is used to display the text corresponding to the voice command or Image information.

 A sixth aspect of the embodiments of the present invention provides an electronic device, where the electronic device includes: a memory, configured to store a voice feature database, the voice feature library includes a plurality of voice commands, and a processor, configured to obtain the voice feature The set of voice commands to be corrected in the library is executed for each of the first to Nth voice commands in the voice command set: the voice feature information is obtained according to the collected voice information, and is updated according to the obtained voice feature information. The voice feature information of the first voice command in the voice feature library to correct the first voice command; each voice command in the N+1th to Mth voice commands in the voice command set Executing: determining, according to the N deviation values of the pre-corrected and corrected speech feature information, the N first deviations of the corrected N first voice commands, and correcting the second voice command according to the determined correction mode Corrected by the second voice command, where M is an integer greater than or equal to 2, and the N is an integer smaller than the M greater than 0, The first voice command is any one of the first to Nth voice commands, and the second voice command is any one of the N+1th to Mth voice commands;

 And an output device, configured to prompt the user to input voice information corresponding to the first voice command; and input means, configured to collect voice information input by the user.

 In a first possible implementation manner of the sixth aspect, the processor is specifically configured to:

 If the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, the change rule is searched according to the N deviation values;

If the change rule is found, determining to update the first voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining to perform the second voice command: prompting The user inputs voice information corresponding to the second voice command, collects voice information input by the user, obtains voice feature information according to the collected voice information, and obtains voice feature information according to the obtained voice information. The obtained voice feature information updates the voice feature information of the second voice command to correct the second voice command.

 With reference to the first possible implementation manner of the sixth aspect, in a second possible implementation manner of the sixth aspect, the processor is further configured to: if the N deviation values are smaller than the first threshold, The correction of the second voice command is then abandoned.

 With reference to the first possible implementation manner of the sixth aspect, in a third possible implementation manner of the sixth aspect, the processor is further configured to: if the N deviation values are greater than the second threshold, Then, the voice command set that needs to be corrected in the voice feature library is re-determined.

 The beneficial effects of the application are as follows:

 The method for correcting the voice command, the correction system, and the electronic device, by prompting the user to input voice information corresponding to the voice command, receiving voice information input by the user according to the prompt, acquiring voice feature information according to the voice information, and using the voice feature information Updating the original voice feature information of the voice command to be corrected; or, when the number of voice commands to be corrected is M, the first N voice commands of the M voice commands to be corrected use the input voice The voice feature information of the information replaces the corresponding voice feature information to correct the first N voice commands, and the back (MN) voice commands pass the original voice feature information of the corrected first N voice commands and the input voice feature information. The deviation value is selected to be corrected, so that the latter (MN) voice commands are corrected, so that the language feature information of the voice command can be adapted to the user terminal device after replacement or the voice feature library is upgraded, and the prior art is solved. After the user terminal device is replaced or the voice signature database is upgraded After that, the voice feature information of some or all voice commands of the voice feature library may be deviated from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized. DRAWINGS

1 is a flowchart of a method for correcting a voice command according to a first preferred embodiment of the present invention; FIG. 2 is a flowchart of a method for correcting a voice command according to a second preferred embodiment of the present application; A structural block diagram of a correction system for a voice command of a preferred embodiment; 4 is a structural block diagram of a system for correcting a voice command according to a fourth preferred embodiment of the present invention; FIG. 5 is a schematic structural diagram of an electronic device according to a fifth preferred embodiment of the present application;

 6 is a schematic structural diagram of an electronic device according to a sixth preferred embodiment of the present application. detailed description

 The embodiment of the present application solves the voice of some or all voice commands of the voice feature library after the user terminal device is replaced or the voice feature database is upgraded in the prior art by providing a voice command correction method, a correction system, and an electronic device. There is a deviation between the feature information and the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 The technical solution in the embodiment of the present application is to solve the above problem, and the general idea is as follows:

 Executing part or all of the voice commands in the voice feature library: prompting the user to input voice information corresponding to the voice command; collecting voice information input by the user; and obtaining voice feature information according to the collected voice information; And updating the voice feature information of the voice command to be corrected according to the obtained voice feature information to correct the voice command.

 Through the foregoing technical solution, the voice command is corrected, so that the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature database is upgraded, and the user equipment device is replaced or voiced in the prior art. After the feature library is upgraded, the voice feature information of some or all of the voice commands of the voice feature library may be deviated from the corresponding voice feature information of the device set, so that there is a technique of identifying the inaccurate or failing to recognize the input voice information during the voice command operation. problem.

 In order to better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the drawings and specific embodiments.

 Embodiment 1

 As shown in FIG. 1, a flow chart of a method for correcting a voice command according to a first preferred embodiment of the present application is shown. The method for correcting the voice command includes the following steps:

Step 110: Obtain a voice command in the voice feature library that needs to be corrected. Specifically, the voice feature library generally includes a plurality of voice commands, voice feature information corresponding to the plurality of voice commands, and further, may further include text information corresponding to the plurality of commands. The voice command in the voice feature library is usually a control instruction name, and when the feature value of the voice information collected by the electronic device through the audio input device matches the corresponding feature information of a voice command in the voice feature library, The electronic device performs a corresponding control operation according to the name of the control instruction corresponding to the voice command. The voice feature information, that is, the voice feature value, is a feature value extracted from the voice information generated when the specific voice input voice command is recognized, and the voice feature values in the voice information generated when different user voices input the same voice command may be different.

 Step 120: Prompt the user to input voice information corresponding to the voice command.

 Specifically, there may be multiple ways of prompting, the first one: playing the voice information corresponding to the voice command through an audio output device (such as an audio, a headphone, etc.) of the electronic device, such as voice feature information according to the voice command. Generating voice information and playing, or directly playing the machine sound corresponding to the voice command, for example, when the voice command corresponds to opening the camera control command, generating voice information according to the voice feature information or according to The machine sound corresponding to the voice command plays the voice message with "turn on the camera". The second type: displaying text or image information corresponding to the voice command, for example, when the voice command corresponds to opening a camera control instruction, displaying text information with the words "open camera" through a display screen or the like.

 Step 130: Collect voice information input by the user.

 Specifically, the voice when the user repeats according to the prompt can be collected by an audio input device (such as a microphone or the like) of the electronic device.

 The following description continues with the voice command corresponding and the camera control command being turned on. After the user receives the above prompt, the voice input has a voice message with "turn on the camera". The voice message with "turn on the camera" is collected by an input device such as a microphone.

 Step 140: Acquire voice feature information according to the collected voice information, that is, extract voice feature values of the voice information according to the collected voice information.

The speech feature information can be obtained by extracting the Mel Frequency Cepstrum Coefficient (MFCC) in the collected speech information: C(n) = ^(m) cos ~ 10≤n < M

 M )

 Where N is the number of points of the Fourier transform of the collected speech information, M is the number of filters of the filter bank used later, and S(m) is the logarithmic energy output by each filter bank. Different systems may use different order MFCC coefficients for speech feature extraction and recognition.

 The MFCC coefficient is a feature information parameter that is widely used at present. Extracting the Meir cepstral coefficient in the voice information of the collected set is the way to obtain the speech feature information of the present application, and can also be decomposed by empirical mode decomposition. The (EMD) technique or the like extracts speech feature information, and does not limit the use of MFCC coefficients.

 Step 150: Update voice feature information of the voice command to be corrected according to the obtained voice feature information to correct the voice command.

 There are at least the following two ways to update: First, the original voice feature information is directly replaced with the obtained voice feature information, and the original voice feature information is not retained, that is, the obtained voice feature information is replaced by the need. The voice feature information of the corrected voice command; secondly, adjusting the parameter value of the voice feature information of the voice command to be corrected according to the deviation value of the obtained voice feature information and the voice feature information of the voice command to be corrected, The voice feature information of the voice command is made applicable.

 The following is a specific implementation process of the method for correcting the voice command. First, a voice command to be corrected is obtained in the voice feature library, and the voice command is assumed to correspond to a control instruction of the control device to perform the camera operation; The eighth output has a voice message with "turn on the camera", or a prompt for text message; after the user receives the above prompt, the voice input has a voice message with "turn on the camera", and the band is collected by an input device such as a microphone. The voice information of "turning on the camera" is extracted; the voice feature information is extracted according to the voice information collected; finally, the voice feature information is updated with the voice feature information, for example, the original voice feature information is replaced, and the correction is completed.

The method for correcting the voice command receives the voice command that needs to be corrected, receives the voice information corresponding to the voice command input by the user, acquires the voice feature information according to the voice information, and updates the voice feature information to the corresponding original voice. Feature information, thereby making the language of the voice command The feature information can be adapted to the voice feature information of some or all voice commands of the voice feature library after the user terminal device is replaced or the voice feature library is upgraded in the prior art after the user terminal device is replaced or the voice feature database is upgraded. There may be deviations from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 The voice command to be corrected may be one of the voice feature library or a plurality of commands. When the voice command is multiple, the following voice commands may be executed for each of the plurality of voice commands. Steps 120, 130, 140, and 150, or performing the following steps one by one for the plurality of voice commands, that is, after the first voice command is corrected, performing the second voice command correction, and then performing the third voice command correction. .. ..., until the calibration is complete.

 Specifically, in step 110, how to obtain the voice command in the voice feature library is required to be corrected in the following two manners:

 First, this method can be used after the user terminal device is replaced, that is, when a voice feature library is newly installed on a terminal device. The method specifically includes the following steps:

 Receiving the voice feature library;

 Determining a need according to a specific voice command in the voice feature library preset by the system, or according to a frequency of use of the voice command in the voice feature library, or a voice command selected in the voice feature library according to input information of the user The corrected voice command is a voice command that needs to be corrected each time the voice feature library is received.

 Secondly, when the data is upgraded to the voice feature database, the method can be used. The method is specifically:

 The updated voice command and its voice feature information are received, and the updated voice command is determined to be a voice command that needs to be corrected.

The upgrade voice command set may include voice feature information corresponding to the original voice command in the voice feature library, or may be a newly added voice command and corresponding voice feature information. Therefore, it is determined that the voice command to be corrected may be the original voice command, or may be a newly added voice command. The method for correcting the voice command is configured to prompt the user to input the voice information corresponding to the voice command, receive the voice information input by the user according to the prompt, acquire the voice feature information according to the voice information, and update the voice feature information to the voice command that needs to be corrected. The voice feature information is such that the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature library is upgraded, and the voice feature database is solved after the user terminal device is replaced or the voice feature database is upgraded in the prior art. The voice feature information of some or all of the voice commands may deviate from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 Embodiment 2

 Based on the same inventive concept, the present application further provides a method for correcting a voice command, as shown in FIG. 2, which is a flowchart of a method for correcting a voice command according to a second preferred embodiment of the present application. The method for correcting the voice command includes the following steps:

 Step 210: Obtain a voice command set that needs to be corrected in the voice feature database, where the voice command set includes M voice commands, and the M is an integer greater than or equal to 2.

 Specifically, the voice feature library generally includes a plurality of voice commands, voice feature information corresponding to the plurality of voice commands, and further may further include text information corresponding to the plurality of commands. The voice command in the voice feature library is usually a control instruction name, and when the feature value of the voice information collected by the electronic device through the audio input device matches the corresponding feature information of a voice command in the voice feature library, The electronic device performs a corresponding control operation according to the name of the control instruction corresponding to the voice command. The voice feature information, that is, the voice feature value, is a feature value extracted from the voice information generated when the specific voice input voice command is recognized, and the voice feature values in the voice information generated when different user voices input the same voice command may be different.

Step 220: Perform, for each voice command in the first to Nth voice commands in the voice command set, prompting the user to input voice information corresponding to the first voice command, and collecting voice information input by the user, Acquiring the voice feature information according to the collected voice information, and updating the voice feature information of the first voice command in the voice feature database according to the obtained voice feature information to correct the first voice command, where the N is smaller than An integer whose M is greater than 0, the first voice command It is any one of the first to Nth voice commands.

 Performing for each of the first to Nth voice commands: prompting the user to input voice information corresponding to the first voice command, collecting voice information input by the user, and obtaining voice information according to the voice information collected by the user And the voice feature information is updated, and the voice feature information of the first voice command in the voice feature database is updated according to the obtained voice feature information to correct the first voice command, which can be implemented by: performing the first to the first Each of the N voice commands is simultaneously executed, or the first to the Nth voice commands are performed one by one, that is, the first voice command is corrected first, and then the second voice is corrected. The command is corrected, ... until the correction of the Nth voice command is completed.

 In addition, the voice feature information can be obtained by extracting the Mel Frequency Cepstrum Coefficient (MFCC) in the collected voice information:

C(n) = ^(m) cos ~ 10≤n < M

 M )

 Where N is the number of points of the Fourier transform of the collected speech information, M is the number of filters of the filter bank used later, and S(m) is the logarithmic energy output by each filter bank. Different systems may use different order MFCC coefficients for speech feature extraction and recognition.

 The MFCC coefficient is a feature information parameter that is widely used at present. Extracting the Meir cepstral coefficient in the voice information of the collected set is the way to obtain the speech feature information of the present application, and can also be decomposed by empirical mode decomposition. The (EMD) technique or the like extracts speech feature information, and does not limit the use of MFCC coefficients.

 Step 230: Perform, for each of the N+1th to Mth voice commands in the voice command set: before and after the correction of the N first voice commands that have been corrected And determining, by the N deviation values of the information, a manner of correcting the second voice command, and correcting the second voice command according to the determined correction manner, where the second voice command is the (N+1)th Go to any of the M voice commands.

That is to say, when the N deviation values satisfy different conditions, the manner of correcting each of the (MN) voice commands is different, and refer to the subsequent description. Each of the N deviation values may obtain the voice feature information by using the MFCC coefficient, and compare with the voice feature information of the corresponding first voice command to obtain a deviation value, where the deviation value may be the obtained voice feature information and The ratio of the voice feature information of the corresponding first voice command may also be the difference between the obtained voice feature information and the voice feature information of the corresponding first voice command.

 The method for correcting the voice command is to correct the N by updating the corresponding voice feature information by using the voice feature information of the input voice information from the first to the Nth voice commands of the M voice commands that need to be corrected. a voice command, and correcting the N+1th to Mth voice commands based on the N deviation values of the pre-corrected and corrected voice feature information according to the corrected N first voice commands, thereby completing the M The correction of the voice command enables the language feature information of the voice command in the voice feature library to be adapted to the user terminal device after the replacement or the voice feature library is upgraded, and solves the problem that the user terminal device is replaced or the voice feature in the prior art. After the library is upgraded, the voice feature information of some or all voice commands of the voice signature database may be deviated from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized. .

 Specifically, there are two ways for obtaining the voice command for obtaining the voice feature database that needs to correct the first voice feature information:

 First, this method can be used after the user terminal device is replaced, that is, when a voice feature library is newly installed on a terminal device. The method specifically includes the following steps:

 Receiving the voice feature library;

 Determining a need according to a specific voice command in the voice feature library preset by the system, or according to a frequency of use of the voice command in the voice feature library, or a voice command selected in the voice feature library according to input information of the user The corrected voice command is a voice command that needs to be corrected each time the voice feature library is received.

 Secondly, when the data is upgraded, the method may be used. The method includes the following steps:

Receiving an updated voice command and its voice feature information, determining the updated voice command as Voice commands that need to be corrected.

 The upgrade voice command set may include voice feature information corresponding to the original voice command in the voice feature library, or may be a newly added voice command and corresponding voice feature information. Therefore, the changed voice command may be the original voice command or a newly added voice command.

 Specifically, when the N deviation values are different, the following method is used to determine a manner in which each of the N+1th to Mth voice commands in the voice command set is corrected:

 If the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, the change rule is searched according to the N deviation values;

 If the change rule is found, determining to update the voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining to perform the second voice command: prompting user input And the voice information corresponding to the second voice command, the voice information input by the user, the voice feature information obtained according to the collected voice information, and the voice feature of the second voice command according to the obtained voice feature information Information to correct the second voice command;

 If the N deviation values are less than the first threshold, discarding the correction of the second voice command

 And if the N deviation values are greater than the second threshold, the voice command set of the voice feature database that needs to be corrected for the voice feature information is re-determined.

 Hereinafter, by way of example, when the N deviation values are different, the following method is used to determine the manner in which the first voice feature information of each of the (M-N) voice commands in the voice command set is corrected. 4, the first threshold is 5%, the second threshold is 20%, and 4 is said to be equal to 3.

In the first case, the first deviation value is 5.9%, the second deviation value is 6%, and the third deviation value is 6%. Since the difference between the three deviation values is only 0.1, three There is a variation law of the deviation value, that is, the difference between the first voice feature information and the required voice feature information is 6%, and the system can issue a prompt, such as displaying, by the display unit, "whether the (MN) voice needs to be adjusted according to the deviation value of 6%. The first voice feature information of each voice command in the command", when the user selects "yes", Then, the first voice feature information of each of the (MN) voice commands is corrected according to a deviation value of 6%.

 In the second case, if the first deviation value is 1%, the second deviation value is 2%, and the third deviation value is 4%, since the three deviation values are all less than 5%, therefore, the pair is discarded (MN Correction of each voice command in a voice command.

 In the third case, if the first deviation value is 15%, the second deviation value is 30%, and the third deviation value is 40%, because two of the three deviation values are greater than 20%, so Determining, in the voice feature library, a set of voice commands that need to be corrected for the first voice feature information, such as all voice commands that are not corrected by the voice feature library are listed as voice commands that need to correct the first voice feature information.

 The above-mentioned voice command correction method 200 corrects the first N voices by replacing the first and last N voice commands of the M voice commands that need to be corrected with the voice feature information of the input voice information by replacing the corresponding voice feature information. And the following (MN) voice commands select a correction mode by using the original voice feature information of the corrected N voice commands and the deviation value of the input voice feature information, thereby correcting the latter (MN) voice commands. Therefore, the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature library is upgraded, and the part or all of the voice feature database after the user terminal device is replaced or the voice feature library is upgraded in the prior art is solved. The voice feature information of the voice command may be deviated from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 Embodiment 3

 Based on the same inventive concept, the present application further provides a voice command correction system 300, as shown in FIG. 3, which is a structural block diagram of a voice command correction system 300 according to a third preferred embodiment of the present application. The correction system 300 includes: a determining unit 310, a prompting unit 320, a collecting unit 330, an obtaining unit 340, and a correcting unit 350.

 a determining unit 310, configured to obtain a voice command in the voice feature library that needs to be corrected;

The prompting unit 320 is configured to prompt the user to input voice information corresponding to the voice command, and the collecting unit 330 is configured to collect voice information input by the user; The obtaining unit 340 is configured to obtain the voice feature information according to the collected voice information;

 The correcting unit 350 is configured to update the voice feature information of the voice command to be corrected according to the obtained voice feature information to correct the voice command.

 In a specific implementation, the determining unit 310 is specifically configured to: receive the voice feature library; according to a specific voice command in the voice feature library preset by the system, or according to a voice command in the voice feature library The voice command to be corrected is determined using a frequency, or a voice command selected in the voice feature library according to the user's input information, and the specific voice command is a voice command that needs to be corrected each time the voice feature library is received.

 In a specific implementation, the determining unit 310 is specifically configured to: receive an updated voice command and voice feature information, and determine the updated voice command as a voice command that needs to be corrected.

 In a specific implementation, the correcting unit 350 is specifically configured to: replace the obtained voice feature information with the voice feature information of the voice command that needs to be corrected.

 In a specific implementation manner, the prompting unit 320 is specifically configured to: generate voice information according to the voice feature information of the voice command and play, or play a machine sound corresponding to the voice command; or display a text corresponding to the voice command Or image information.

 The above-mentioned voice command correction system 300 prompts the user to input the voice information corresponding to the first voice command through the prompting unit 320, and collects the voice information input by the user according to the prompt by the collecting unit 330, and the obtaining unit 340 according to the voice information. Acquiring the voice feature information, and the voice feature information is updated by the correcting unit 350 to the voice feature information of the voice command to be corrected, so that the language feature information of the voice command can be adapted after the user terminal device is replaced or the voice feature library is upgraded. After the replacement of the user terminal device or the upgrade of the voice feature library in the prior art, the voice feature information of some or all of the voice commands of the voice feature library may be deviated from the corresponding voice feature information of the device set, thereby operating in the voice command. There are technical problems in the identification of inaccurate or unrecognizable voice messages, packets.

 Embodiment 4

Based on the same inventive concept, the present application further provides a voice command correction system, as shown in FIG. 4, which is a structural block diagram of a voice command correction system 400 according to a fourth preferred embodiment of the present application. Place The correction system 400 includes an obtaining unit 410, a first correcting unit 420, and a second correcting unit 430. The obtaining unit 410 is configured to obtain a voice command set that needs to be corrected in the voice feature database, where the voice command set includes M voice commands, where the M is an integer greater than or equal to 2;

 The first correcting unit 420 is configured to perform, for each voice command, the first to the Nth voice commands in the voice command set: prompting the user to input voice information corresponding to the first voice command, and collecting the user Inputting voice information, obtaining voice feature information according to the collected voice information, updating sound feature information of the first voice command in the voice feature database according to the obtained voice feature information, to correct the first voice command, N is an integer smaller than the M is greater than 0, and the first voice command is any one of the first to Nth voice commands.

 a second correcting unit 430, configured to perform, for each of the N+1th to Mth voice commands in the voice command set: according to the corrected N first voice commands before and after the correction And determining, by the N deviation values of the corrected voice feature information, a manner of correcting the second voice command, and correcting the second voice command according to the determined correction mode, where the second voice command is Any one of the N+1th to Mth voice commands.

 In a specific implementation, the second correcting unit 430 is specifically configured to:

 If the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, the change rule is searched according to the N deviation values;

 If the change rule is found, determining to update the voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining to perform the second voice command: prompting user input And the voice information corresponding to the second voice command, the voice information input by the user, the voice feature information obtained according to the collected voice information, and the voice feature of the second voice command according to the obtained voice feature information Information to correct the second voice command.

 In a specific implementation, the second correcting unit 430 is further configured to: if the N deviation values are less than the first threshold, discard the correction of the second voice command.

In a specific implementation, the second correcting unit 430 is further configured to: if the N deviation values are greater than the second threshold, re-determine the voice command set in the voice feature database that needs to be corrected. The above-mentioned voice command correction system 400 updates the voice feature information of the input voice information to the corresponding voice feature by using the first to Nth voice commands of the M voice commands to be corrected by the first correction unit 420. The information is used to correct the N voice commands, and the second correction unit 430 determines the correction mode of the (N)th to Mth voice commands according to the N deviation values, and corrects the (MN) voice commands. Therefore, the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature library is upgraded, and the part or all of the voice feature database after the user terminal device is replaced or the voice feature library is upgraded in the prior art is solved. The voice feature information of the voice command may be deviated from the corresponding voice feature information of the device set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 Embodiment 5

 Based on the same inventive concept, the present application further provides an electronic device 500. As shown in FIG. 5, it is a schematic structural diagram of an electronic device 500 according to a fifth preferred embodiment of the present application. The electronic device 500 includes a memory 510, a processor 520, an output device 530, and an input device 540.

 The memory 510 is configured to store a voice feature library, where the voice feature library includes a plurality of voice commands. The memory 510 is a memory device in the terminal system for storing programs and data. The memory 510 can also be an external memory such as a magnetic medium or an optical disk, or an internal memory.

 The processor 520 is configured to obtain a voice command that needs to be corrected in the voice feature database, obtain voice feature information according to the collected voice information, and update voice feature information of the voice command that needs to be corrected according to the voice feature information, to Correct the voice command.

 The output device 530 is configured to prompt the user to input voice information corresponding to the voice command. The output device 530 expresses data in the form of numbers, characters, images, sounds, etc., such as a display, a voice output system such as a speaker, and the like.

 The input device 540 is configured to collect voice information input by the user. The input device 540 can be a device such as a microphone or a microphone.

In a specific implementation manner, when the electronic device 500 receives the voice feature database, the processor 520 is specifically configured to: according to a specific voice command in the voice feature library preset by the system, or according to The frequency of use of the voice command in the voice feature library, or according to the user's The voice command selected by the input information in the voice feature library determines a voice command to be corrected, and the specific voice command is a voice command that needs to be corrected each time the voice feature library is received.

 In a specific implementation, the processor 520 is further configured to: receive an updated voice command and voice feature information thereof, and determine the updated voice command as a voice command that needs to be corrected.

 In a specific implementation, the processor 520 is specifically configured to: replace the obtained voice feature information with the voice feature information of the voice command that needs to be corrected.

 In a specific implementation manner, the output device 530 is specifically a voice player or a display screen; the voice player is configured to generate voice information according to the voice feature information of the voice command, and play, or play the machine corresponding to the voice command. The display screen is used to display text or image information corresponding to the voice command.

 The electronic device 500 is configured to prompt the user to input voice information corresponding to the voice command through the output device 530, and receive voice information input by the user according to the prompt by the input device 540, and acquire the voice feature information according to the voice information by the processor 520. The voice feature information is updated to the voice feature information that needs to be corrected, so that the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature database is upgraded, and the prior art is replaced after the user terminal device is replaced. After the voice signature database is upgraded, the voice feature information of some or all of the voice commands of the voice feature library may be deviated from the corresponding voice feature information of the device set, so that the voice information may be inaccurately recognized or the input voice information may not be recognized during the voice command operation. Technical problem.

 Embodiment 6

 Based on the same inventive concept, the present application further provides an electronic device 600. As shown in FIG. 6, FIG. 6 is a schematic structural diagram of an electronic device 600 according to a sixth preferred embodiment of the present application. The electronic device 600 includes: a memory 610, a processor 620, an output device 630, and an input device 640.

 The memory 610 is configured to store a voice feature library, where the voice feature library includes a plurality of voice commands. The memory 610 is a memory device in the terminal system for storing programs and data. The memory 610 can also be an external memory such as a magnetic medium or an optical disk, or an internal memory.

The processor 620 is configured to obtain a voice command set that needs to be corrected in the voice feature database, and execute each voice command in the first to Nth voice commands in the voice command set: according to the collected words Acquiring information to obtain voice feature information, updating tone feature information of the first voice command in the voice feature library according to the obtained voice feature information to correct the first voice command; and N+ in the voice command set Performing each of the 1 to M voice commands: determining the second voice based on the N deviation values of the corrected and corrected voice feature information of the N first voice commands that have been corrected The command is used to perform the correction, and the second voice command is corrected according to the determined correction mode, where M is an integer greater than or equal to 2, and the N is an integer smaller than the M greater than 0, the first voice The command is any one of the first to Nth voice commands, and the second voice command is any one of the N+1th to Mth voice commands;

 The output device 630 is configured to prompt the user to input voice information corresponding to the first voice command. The output device 630 expresses data in the form of numbers, characters, images, sounds, etc., such as a display, a voice output system such as a speaker, and the like.

 The input device 640 is configured to collect voice information input by the user. The input device 640 can be a device such as a microphone or a microphone.

 In a specific implementation, the processor 620 is specifically configured to: if the N deviation values are between the first threshold and the second threshold, and the first threshold is less than the second threshold, Finding a change rule according to the N deviation values;

 If the change rule is found, determining to update the first voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, determining to perform the second voice command: prompting The user inputs voice information corresponding to the second voice command, collects voice information input by the user, obtains voice feature information according to the collected voice information, and updates the second voice command according to the obtained voice feature information. Voice feature information to correct the second voice command.

 In a specific implementation, the processor 620 is further configured to: if the N deviation values are less than the first threshold, discard the correction of the second voice command.

In a specific implementation, the processor 620 is further configured to: if the N deviation values are greater than the second threshold, re-determine the voice command set in the voice feature database that needs to be corrected. The electronic device 600 is configured to prompt the user to input voice information corresponding to the first voice command through the output device 630, and collect voice information input by the user according to the prompt by the input device 640, and extract the voice in the voice information by using the processor 620. Feature information, updating the extracted voice feature information to the voice feature information of the first voice command, thereby completing correction of the first to Nth voice commands of the M voice commands, and then passing the processor 620 And selecting, according to the N first deviation commands that have been corrected according to the N deviation values of the pre-corrected and corrected speech feature information, a manner of correcting the (MN) voice commands in the M voice commands, Correcting the (MN) voice commands according to the correction manner, so that the language feature information of the voice command can be adapted to the user terminal device after the replacement or the voice feature library is upgraded, and the prior art in the user terminal is solved. After the equipment is replaced or after the voice signature database is upgraded, the voice feature information and settings of some or all voice commands of the voice signature database are set. There is a deviation in the corresponding voice feature information of the backup set, so that there is a technical problem that the voice command operation is inaccurate or the input voice information cannot be recognized.

 Although the preferred embodiment of the invention has been described, it will be apparent to those skilled in the < Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and the modifications and modifications The spirit and scope of the invention. Thus, it is intended that the present invention cover the modifications and variations of the inventions

Claims

Rights request 1. A voice command correction method, characterized in that the method includes: Obtain the voice command that needs to be corrected in the voice feature library; Prompt the user to input voice information corresponding to the voice command; Collect the voice information input by the user; Obtain speech feature information based on the collected speech information; Update the voice feature information of the voice command that needs to be corrected according to the obtained voice feature information to correct the voice command. 2. The correction method according to claim 1, characterized in that said obtaining the voice commands that need to be corrected in the voice feature library specifically includes: Receive the speech feature library; The need is determined based on the specific voice command in the voice feature database preset by the system, or based on the frequency of use of the voice command in the voice feature database, or based on the voice command selected in the voice feature database based on the user's input information. Corrected voice command, the specific voice command is a voice command that needs to be corrected after receiving the voice feature library each time. 3. The correction method according to claim 1, characterized in that said obtaining the voice commands that need to be corrected in the voice feature library specifically includes: Receive the updated voice command and its voice characteristic information, and determine the updated voice command as a voice command that needs correction. 4. The correction method according to any one of claims 1-3, wherein the voice feature information of the voice command that needs to be corrected is updated according to the obtained voice feature information, specifically: The obtained voice feature information is replaced with the voice feature information of the voice command that needs to be corrected. 5. The correction method according to any one of claims 1 to 4, characterized in that the user is prompted to input voice information corresponding to the voice command, specifically: Generate voice information based on the voice feature information of the voice command and play it; or Play the machine tone corresponding to the voice command; or Display text or image information corresponding to the voice command. 6. A voice command correction method, characterized in that the method includes: Obtain a set of voice commands that need to be corrected in the voice feature library. The set of voice commands includes M voice commands, and the M is an integer greater than or equal to 2; Execute for each voice command from the first to the Nth voice command in the voice command set: prompt the user to input the voice information corresponding to the first voice command, collect the voice information input by the user, and collect the voice information according to the collection The obtained voice information obtains voice feature information, and the voice feature information of the first voice command in the voice feature library is updated according to the obtained voice feature information to correct the first voice command, and the N is less than the M and is greater than is an integer of 0, and the first voice command is any one of the first to Nth voice commands; Execute for each voice command from the N+1th to Mth voice commands in the voice command set: N based on the pre-correction and post-correction voice feature information of the N first voice commands that have been corrected deviation values, determine the method of correcting the second voice command, correct the second voice command according to the determined correction method, the second voice command is the N+1th to Mth Any one of the voice commands. 7. The correction method according to claim 6, characterized in that, based on the N deviation values of the corrected N first voice commands before and after the correction of the voice feature information, the determination of the first 2. Voice command correction methods include: If the N deviation values are between the first threshold and the second threshold, and the first threshold is smaller than the second threshold, then search for a change pattern based on the N deviation values; If the change rule is found, it is determined to update the voice feature information of the second voice command according to the change rule to correct the second voice command; otherwise, it is determined to execute: Prompt user input for the second voice command. Voice information corresponding to the second voice command, collecting voice information input by the user, obtaining voice feature information based on the collected voice information, and updating voice features of the second voice command based on the obtained voice feature information. information to correct the second voice command. 8. The correction method according to claim 7, characterized in that, based on the N deviation values of the corrected N first voice commands before and after the correction of the voice feature information, the determination of the correction of the first voice command is carried out. 2. Voice command correction methods also include: If the N deviation values are less than the first threshold, the correction of the second voice command is given up. 9. The correction method according to claim 7, characterized in that, based on the N deviation values of the corrected N first voice commands before and after the correction of the voice feature information, the determination of the corrected N first voice commands is determined. 2. Voice command correction methods also include: If the N deviation values are greater than the second threshold, then the set of voice commands in the voice feature database that need to be corrected is re-determined. 10. A voice command correction system, characterized in that the system includes: The determination unit is used to obtain the voice command that needs to be corrected in the voice feature library; The prompting unit is used to prompt the user to input the voice information corresponding to the voice command; the collection unit is used to collect the voice information input by the user; The obtaining unit is used to obtain speech feature information based on the collected speech information; A correction unit, configured to update the voice feature information of the voice command that needs to be corrected according to the obtained voice feature information to correct the voice command. 11. The correction system according to claim 10, characterized in that the determination unit is specifically used to: Receive the voice feature database; According to the specific voice command in the voice feature database preset by the system, or according to the frequency of use of voice commands in the voice feature database, or according to the user's input information in the voice feature database The voice command selected in determines the voice command that needs to be corrected, and the specific voice command is the voice command that needs to be corrected each time the voice feature library is received. 12. The correction system according to claim 10, characterized in that the determining unit is specifically used to: Receive the updated voice command and its voice characteristic information, and determine the updated voice command as a voice command that needs correction. 13. The correction system according to any one of claims 10 to 12, wherein the correction unit is specifically configured to: replace the voice feature information of the voice command that needs correction with the obtained voice feature information. 14. The correction system according to any one of claims 10 to 13, characterized in that the prompt unit is specifically used for: Generate voice information based on the voice feature information of the voice command and play it, or Play the machine tone corresponding to the voice command; or Display text or image information corresponding to the voice command. 15. A voice command correction system, characterized in that the system includes: The obtaining unit is used to obtain a set of voice commands that need to be corrected in the voice feature library. The set of voice commands includes M voice commands, and the M is an integer greater than or equal to 2; A first correction unit, configured to perform on each of the first to Nth voice commands in the voice command set: prompt the user to input voice information corresponding to the first voice command, and collect the user input voice information, obtain voice feature information according to the collected voice information, update the sound feature information of the first voice command in the voice feature library according to the obtained voice feature information to correct the first voice command, N is an integer smaller than M and greater than 0, and the first voice command is any one of the first to Nth voice commands; The second correction unit is configured to perform on each of the N+1th to Mth voice commands in the voice command set: before correction and correction according to the N first voice commands that have been corrected N deviation values of the final voice feature information, determine a correction method for the second voice command, correct the second voice command according to the determined correction method, and the second voice command is the third voice command. Any voice command from N+1 to Mth voice command. 16. The correction system according to claim 15, characterized in that the second correction unit is specifically used to: If the N deviation values are between the first threshold and the second threshold, and the first threshold is smaller than the second threshold, then search for a change pattern based on the N deviation values; If the change rule is found, determine the language of the second voice command according to the change rule. The voice characteristic information is updated to correct the second voice command; otherwise, determine to execute the second voice command: prompt the user to input voice information corresponding to the second voice command, collect the voice information input by the user, Obtain voice feature information according to the collected voice information, and update the voice feature information of the second voice command according to the obtained voice feature information to correct the second voice command. 17. The correction system of claim 16, wherein the second correction unit is further configured to: if the N deviation values are less than the first threshold, give up responding to the second voice command correction. 18. The correction system according to claim 16, wherein the determination unit is further configured to: if the N deviation values are greater than the second threshold, re-determine that the speech feature library needs to be corrected A collection of voice commands. 19. An electronic device, characterized in that the electronic device includes: A memory, used to store a voice feature library that includes multiple voice commands; a processor, used to obtain the voice commands that need to be corrected in the voice feature library, obtain voice feature information based on the collected voice information, and The obtained voice feature information updates the voice feature information of the voice command that needs to be corrected to correct the voice command; An output device is used to prompt the user to input voice information corresponding to the voice command; an input device is used to collect the voice information input by the user. 20. The electronic device according to claim 19, wherein when the electronic device receives the voice feature library, the processor is specifically configured to: select the voice feature library preset by the system. The specific voice command needs to be corrected, or the voice command that needs to be corrected is determined based on the frequency of use of the voice command in the voice feature database, or the voice command selected in the voice feature database according to the user's input information, and the specific voice command is for each The voice command that needs to be corrected after receiving the voice feature library for the first time. 21. The electronic device of claim 19, wherein the processor is further configured to: receive an updated voice command and its voice feature information, and determine the updated voice command as a voice command that needs correction. . 22. The electronic device according to any one of claims 19 to 21, wherein the processor is specifically configured to: replace the voice feature information of the voice command that needs correction with the obtained voice feature information. 23. The electronic device according to any one of claims 19 to 22, characterized in that the output device is specifically a voice player or a display screen; The voice player is used to generate voice information according to the voice feature information of the voice command and play it, or play the machine sound corresponding to the voice command; The display screen is used to display text or image information corresponding to the voice command. 24. An electronic device, characterized in that the electronic device includes: A memory, used to store a voice feature database, which includes multiple voice commands; a processor, used to obtain a set of voice commands that need to be corrected in the voice feature database, and perform the first to Nth set of voice commands in the voice feature database. Each voice command in the voice command executes: Obtain voice feature information according to the collected voice information, Update the voice feature information of the first voice command in the voice feature library according to the obtained voice feature information to correct the first voice command Voice command; Execute for each voice command in the N+1th to Mth voice commands in the voice command set: based on the voice characteristics of the N first voice commands that have been corrected before and after correction. N deviation values of the information, determine the method of correcting the second voice command, correct the second voice command according to the determined correction method, the M is an integer greater than or equal to 2, and the N is is less than the integer M is greater than 0, the first voice command is any one of the first to N-th voice commands, and the second voice command is the N+1-th to Any one of the M voice commands; An output device is used to prompt the user to input voice information corresponding to the first voice command; an input device is used to collect the voice information input by the user. 25. The electronic device of claim 24, wherein the processor is specifically configured to: if the N deviation values are between a first threshold and a second threshold, the first threshold is less than the second threshold value, then search for the change rule based on the N deviation values; If the change rule is found, determine the third voice command of the second voice command according to the change rule. A voice feature information is updated to correct the second voice command; otherwise, determine to execute the second voice command: prompt the user to input voice information corresponding to the second voice command, collect the voice input by the user information, obtain voice feature information according to the collected voice information, and update the voice feature information of the second voice command according to the obtained voice feature information to correct the second voice command. 26. The electronic device according to claim 25, wherein the processor is further configured to: if the N deviation values are less than the first threshold, give up processing of the second voice command. Correction. 27. The electronic device according to claim 25, wherein the processor is further configured to: if the N deviation values are greater than the second threshold, re-determine the required speech feature library. Corrected voice command collection.