A voice recording method, device, equipment and medium

Abstract

The present application discloses a voice note-taking method, device, equipment and medium, which are used to solve the problem that the existing voice note-taking operation is not flexible enough and the user experience is poor. Since the present application determines that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using the voice note, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making the target operation in the voice note more flexible and convenient, simplifying the process for determining each keyword contained in the text information of the voice information to be recognized, and improving the efficiency of the voice note-taking process.

Description

Technical Field

[0001] The present application relates to the technical field of natural language understanding, and in particular to a voice note-taking method, device, equipment and medium. Background Art

[0002] With the continuous breakthroughs in artificial intelligence technology and the increasing popularity of various smart devices, human-computer interaction appears more and more frequently in people's daily work and life. In daily life, people also use the voice note function on smart devices to record some important events in life for subsequent inquiries on important events.

[0003] In the prior art, the implementation method of voice memo is that after the smart device receives a trigger operation of storage, such as receiving a voice message containing keywords such as "storage" or "start storage", or the user selects the icon corresponding to the storage operation of voice memo on the smart device, it starts to collect and store voice information, and extracts each keyword contained in the text information of the stored voice information through a pre-trained keyword extraction model, and stores each keyword as a keyword set in a database. When the smart device subsequently receives a trigger operation of query, such as receiving a voice message containing keywords such as "query" or "start query", or the user selects the icon corresponding to the query operation of voice memo on the smart device, it starts to collect query voice information, extracts each keyword in the text information of the query voice information through a pre-trained keyword extraction model, searches for a keyword set containing each keyword in the text information of the query voice information in the database, generates reply information according to the keywords in the keyword set and outputs it.

[0004] In this method, the smart device can only start to execute the target operation of the voice note after receiving a trigger operation of a certain target operation explicitly requested by the user. Therefore, the process of the target operation of the voice note is not flexible enough and the user experience is poor. Summary of the invention

[0005] The present application provides a voice note-taking method, apparatus, device and medium, which are used to solve the problem that the existing operation of using voice notes for storage and query is not flexible enough and the user experience is poor.

[0006] In a first aspect, the present application provides a voice note-taking method, the method comprising:

[0007] Obtain text information of the voice information to be recognized;

[0008] Matching the text information with any pre-saved intent template;

[0009] If there is a matching intent template, and the intent corresponding to the matching intent template is the target operation of using voice notes, then each keyword contained in the text information is obtained according to the character interval where each keyword in the matching intent template is located, and the target operation is performed on each keyword.

[0010] Furthermore, the target operation is a storage operation or a query operation.

[0011] Furthermore, before performing the target operation on each keyword, the method further includes:

[0012] The voiceprint features of the received wake-up voice information are obtained through the pre-trained voiceprint model;

[0013] The voiceprint feature is matched with any pre-saved registered voiceprint feature. If there is a matching registered voiceprint feature, the subsequent step of performing the target operation on each keyword is performed.

[0014] Further, if the target operation is a storage operation, performing the target operation on each keyword includes:

[0015] Outputting prompt information for setting query permissions, and storing each keyword contained in the text information of the voice information to be recognized as a piece of information;

[0016] Each target identity identifier that is set to perform the storage information query is received, and each target identity identifier and the storage information are correspondingly saved.

[0017] Further, if the target operation is a query operation, performing the target operation on each keyword includes:

[0018] A query is performed according to each of the keywords. If target storage information containing each of the keywords is found, each target identity identifier corresponding to the target storage information is obtained. If the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, a reply message is generated and output according to each target keyword stored in the target storage information.

[0019] Further, the generating and outputting of reply information includes:

[0020] Obtaining the target location information of the user whose identity corresponds to the matched registered voiceprint feature;

[0021] According to the pre-stored position information of each playback device, the playback device whose position information is closest to the target position information is used as the target playback device, and the generated reply information is outputted through the target playback device.

[0022] Furthermore, the method further comprises:

[0023] If the identity corresponding to the matched registered voiceprint feature is not included in the target identity, the text information of the voice information to be recognized is filtered, and the preset rejection response reply information is output.

[0024] In a second aspect, the present application further provides a voice note taking device, the device comprising:

[0025] An acquisition unit, used for acquiring text information of the voice information to be recognized;

[0026] A matching unit, used to match the text information with any pre-saved intention template;

[0027] A processing unit is used to obtain each keyword contained in the text information according to the character interval where each keyword in the matching intent template is located, and perform the target operation on each keyword if there is a matching intent template and the intent corresponding to the matching intent template is a target operation of using voice notes.

[0028] In a third aspect, the present application further provides an electronic device, which includes at least a processor and a memory, and the processor is used to implement the steps of any of the above-mentioned voice note-taking methods when executing a computer program stored in the memory.

[0029] In a fourth aspect, the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned voice note-taking methods.

[0030] Since in the present application, when it is determined that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using voice notes, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making it more flexible and convenient to perform the target operation when using voice notes, and simplifying the process for determining each keyword contained in the text information of the voice information to be recognized, thereby improving the efficiency of the voice note-taking process. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 A schematic diagram of a voice note-taking process provided for some embodiments of the present application;

[0033] Figure 2 A specific voiceprint registration process provided for some embodiments of the present application;

[0034] Figure 3 A schematic diagram of an application scenario of an electronic device provided in some embodiments of the present application;

[0035] Figure 4 A flowchart of a specific storage operation using voice notes provided for some embodiments of the present application;

[0036] Figure 5 A flowchart of a specific query operation using voice notes provided for some embodiments of the present application;

[0037] Figure 6 A schematic diagram of the structure of a voice note-taking device provided in some embodiments of the present application;

[0038] Figure 7 A schematic diagram of the structure of an electronic device provided for some embodiments of the present application. DETAILED DESCRIPTION

[0039] In order to increase the flexibility of using voice notes to perform target operations and improve user experience, the present application provides a voice note method, device, equipment and medium.

[0040] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0041] In actual application, when a user wants to use voice notes through voice information, for example, to store the location of an item, query the password of an application, etc., the user can use smart devices such as smart speakers and smart TVs to start collecting voice information to be recognized after the user enters a trigger operation to use the voice notepad. The smart device that collects the voice information to be recognized can directly extract each keyword contained in the text information of the voice information to be recognized through a keyword extraction model locally, or send the voice information to be recognized or the text information of the voice information to be recognized to an electronic device, and the electronic device extracts each keyword contained in the text information of the voice information to be recognized through a keyword extraction model. After each keyword is obtained, some subsequent processing is performed to realize the storage or query of information through voice information.

[0042] Figure 1 A schematic diagram of a voice note-taking process provided for some embodiments of the present application, the process comprising:

[0043] S101: Acquire text information of voice information to be recognized.

[0044] The voice note-taking method provided in the present application is applied to an electronic device, which may be a smart device such as a smart speaker, a smart TV, or a smart housekeeper server, a cloud server, etc.

[0045] In the present application, the text information of the voice information to be recognized obtained by the electronic device may be text information generated based on the received voice information to be recognized, or may be text information of the voice information to be recognized received from other smart devices.

[0046] The voice information to be recognized received by the electronic device may be voice information sent by other smart devices or voice information collected by the electronic device itself, which is not specifically limited here.

[0047] S102: Match the text information with any pre-saved intent template.

[0048] Since electronic devices may perform many functions, such as controlling smart home devices to perform corresponding operations, playing music, etc., after obtaining the text information of the voice information to be recognized, the electronic device must perform intent recognition on the text information of the voice to be recognized. Determine the subsequent operation based on the intention of recognizing the voice information to be recognized. Therefore, in order to accurately determine the intention of the voice information to be recognized, in this application, an intent template is pre-saved. After the text information of the voice information to be recognized is obtained based on the above embodiment, the text information is matched with any pre-saved intent template, and the subsequent operation is determined based on the matching result.

[0049] S103: If there is a matching intent template, and the intent corresponding to the matching intent template is the target operation of using voice notes, then according to the character interval where each keyword in the matching intent template is located, each keyword contained in the text information is obtained, and the target operation is performed on each keyword.

[0050] In a specific implementation, the text information of the voice information to be recognized is matched with any pre-saved intent template. If the text information matches any pre-saved intent template, the target intent corresponding to the matched intent template is determined based on the correspondence between the pre-saved intent template and the intent. When there is a matching intent template, and the intent corresponding to the matched intent template is the target operation of using voice notes, it means that the current user wants to use the voice note function, then each keyword contained in the text information of the voice information to be recognized is obtained in the corresponding character interval according to the character interval where each keyword in the saved matched intent template is located.

[0051] For example, the text information is "Put 84 disinfectant in the bathroom cabinet", and the text information matches the pre-saved intention template "Put * in *". According to the correspondence between the pre-saved intention template and the intention, it is determined that the intention corresponding to the matched intention template "Put * in *" is the storage operation using voice notes. Then, according to the character interval where each keyword in the matched intention template "Put * in *", that is, the character interval between "put" and "put in", and the character interval after "put in", each keyword contained in the text information is obtained as "84 disinfectant" and "bathroom cabinet" respectively.

[0052] Among them, due to the different speaking habits of different users, invalid or meaningless characters or punctuation marks may exist in the text information of the voice information to be recognized obtained according to the received voice information to be recognized, for example, the text information is "Put, uh, that, 84 disinfectant in the bathroom cabinet", "Well, where is the alcohol placed", etc. Therefore, in order to facilitate the subsequent determination of whether the user uses the voice note function, the text information can be pre-processed first, and some invalid or meaningless characters or punctuation marks contained in the text information can be deleted according to the preset processing rules. For example, the text information is "Put, uh, that, 84 disinfectant in the bathroom cabinet". The text information is processed according to the preset processing rules, and the processed text information is "Put 84 disinfectant in the bathroom cabinet".

[0053] After the target operation is determined based on the above embodiment, and each keyword included in the text information of the voice information to be recognized is included, the target operation is performed on each keyword.

[0054] Since the present application determines that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using voice notes, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making it more flexible and convenient to perform the target operation when using voice notes, and simplifies the process for determining each keyword contained in the text information of the voice information to be recognized, thereby improving the efficiency of the voice note-taking process.

[0055] In order to improve the flexibility of using voice notes to perform target operations, based on the above embodiments, in this application, the target operation is a storage operation or a query operation.

[0056] Specifically, if the target operation is a storage operation, each keyword contained in the text information of the voice information to be recognized is taken as a storage information, and the storage information is saved. If the target operation is a query operation, the storage information of each keyword contained in the text information of the voice information to be recognized is searched, and the storage information of each keyword contained in the text information of the voice information to be recognized is taken as the target storage information, and then according to each target keyword saved in the target storage information and the preset reply template, the reply information is generated and output.

[0057] For example, the target operation is a storage operation, and the text information is "Put the alcohol in the TV cabinet in the living room". The keywords contained in the acquired text information are "alcohol" and "TV cabinet in the living room", then the keywords "alcohol" and "TV cabinet in the living room" are saved as one storage information.

[0058] The target operation is a query operation, the text information is "Where is the 84 disinfectant placed", and the keyword contained in the obtained text information is "84 disinfectant", then it is determined that the storage information "84 disinfectant bathroom cabinet" contains the keyword "84 disinfectant", and the "84 disinfectant bathroom cabinet" storage information is used as the target storage information, and based on the target keywords "84 disinfectant" and "bathroom cabinet" contained in the target storage information, and the preset reply information template, the reply information "84 disinfectant is placed in the bathroom cabinet" is generated and output.

[0059] The specific process of generating a reply message based on keywords and a preset reply message template belongs to the prior art and will not be described in detail here.

[0060] In order to protect user privacy, based on the above embodiments, in the present application, before performing the target operation on each keyword, the method further includes:

[0061] The voiceprint features of the received wake-up voice information are obtained through the pre-trained voiceprint model;

[0062] The voiceprint feature is matched with any pre-saved registered voiceprint feature. If there is a matching registered voiceprint feature, the subsequent step of performing the target operation on each keyword is performed.

[0063] In actual applications, users may record the storage locations of some valuable items that are not frequently used, or various private information such as account passwords, through the storage operation of electronic devices. Users generally do not want people other than family members to query the storage location of such private information. If anyone can query the storage information of the valuable items or passwords through voice information, the user's privacy cannot be guaranteed. Therefore, in order to protect user privacy, in this application, a voiceprint model is pre-trained to identify the identity of the user currently using the electronic device.

[0064] It should be noted that the training process of the voiceprint model belongs to the existing technology and will not be repeated here.

[0065] Specifically, before inputting the voice information to be recognized, the user can first wake up the electronic device through the wake-up voice information, such as "Hello, Harry", and then control the electronic device to perform the target operation. After receiving the wake-up voice information, the electronic device obtains the voiceprint features of the received wake-up voice information through the pre-trained voiceprint model. The voiceprint features are matched with any pre-saved registered voiceprint features. If there is a matching voiceprint feature, it means that the user who input the wake-up voice information has the operation authority for the electronic device, and the target operation can be performed on each keyword contained in the text information of the voice information to be recognized obtained after the wake-up voice information.

[0066] The registered voiceprint feature is the voiceprint feature of the wake-up voice information input by the user obtained by the electronic device through the pre-trained voiceprint model. The following is a specific implementation method to illustrate the process of voiceprint registration. Figure 2 The specific voiceprint registration process provided in some embodiments of the present application includes:

[0067] S201: The electronic device receives a trigger operation for voiceprint registration input by a user.

[0068] The trigger operation may be that the user clicks a button corresponding to voiceprint registration on the electronic device, or the user wakes up the electronic device and inputs a voice message or text message with content similar to "I want to register voiceprint".

[0069] S202: When the electronic device determines that the user is registering a voiceprint, it outputs a prompt message prompting the user to input a wake-up voice message, and determines whether the wake-up voice message is input within a set time period. If so, execute S203, otherwise, execute S204.

[0070] S203: The electronic device obtains the voiceprint feature of the wake-up voice information through a pre-trained voiceprint model, and saves the voiceprint feature as a registered voiceprint feature, and the electronic device outputs a prompt message indicating successful registration.

[0071] For example, the prompt information is sent to the user's smart terminal, and the prompt information of "voiceprint registration successful" is displayed in text format on the display interface of the smart terminal, and / or the prompt information of "voiceprint registration successful" is broadcast in audio format.

[0072] S204: Determine that the voiceprint registration of the user has failed. The electronic device may output a prompt message indicating whether to re-register if the registration has failed, and determine whether confirmation information for re-registration has been received. If so, execute S202; otherwise, execute S201.

[0073] For example, a prompt message is sent to the user's smart terminal, and a prompt message in text format "Voiceprint registration failed, do you want to re-register" is displayed on the display interface of the smart terminal, and / or a prompt message in audio format "Voiceprint registration failed, do you want to re-register" is broadcast.

[0074] In another possible implementation, some users will also record the storage locations of dangerous items such as 84 disinfectant, pesticides, etc. in the electronic device. Users generally do not want children other than adults in the family to know the storage locations of these dangerous items. If children in the family can also query the storage locations of these dangerous items through the electronic device, the possibility of children in the family being in danger due to contact with dangerous items will increase, and the life, health and safety of the children cannot be guaranteed.

[0075] Therefore, in order to prevent children from getting into danger after they find out where dangerous items are stored, in this application, when using the storage operation of voice memo, the user can set the query permission, so as to prevent children in the family from finding out where some dangerous items are stored. Specifically, if the target operation is a storage operation, performing the target operation on each keyword includes:

[0076] Outputting prompt information for setting query permissions, and storing each keyword contained in the text information of the voice information to be recognized as a piece of information;

[0077] Each target identity identifier that is set to perform the storage information query is received, and each target identity identifier and the storage information are correspondingly saved.

[0078] In the present application, the prompt information for setting the query permission can be outputted in the form of audio format, for example, the prompt information "Please set the query permission" for supplementing the key words of the access mode can be outputted in the form of text, for example, the prompt information "Please set the query permission: Dad, Mom, Grandma, Grandpa, Children" can be outputted in the form of audio format, or the prompt information can be displayed in the form of text on the display interface, for example, the prompt information for setting the query permission can be displayed in the form of text on the display interface. The two ways of outputting the prompt information can also be combined at the same time, that is, the prompt information in audio format can be broadcasted at the same time and the prompt information in text format can be displayed on the display interface.

[0079] The specific method of outputting the prompt information can be pre-set according to the user's preference, or can be selected according to the capabilities of the electronic device. For example, some electronic devices do not have a display interface that can display prompt information. For these electronic devices, when outputting prompt information, the prompt information can be broadcast in audio format.

[0080] In some possible embodiments, while outputting the prompt information for setting the query authority, the electronic device may also use each keyword contained in the text information of the voice information to be recognized obtained based on the above embodiment as a piece of storage information. After outputting the prompt information for setting the query authority, the electronic device may receive each target identity identifier set by the user that can query the storage information, and correspondingly save each target identity identifier and the storage information.

[0081] In order to facilitate the determination of the target identity, when voiceprint registration is performed, after the registration is successful, a prompt message prompting the user to enter the identity can be output, and the corresponding relationship between the voiceprint feature of the successful registration and the received identity can be saved. For example, after the user successfully registers the voiceprint, a prompt message prompting the user to enter the identity can be output, and a prompt message of "Please enter the identity tag" in text format can be displayed on the display interface, and / or a prompt message of "Please enter the identity tag" in audio format can be broadcast. The user can enter the identity "Mom" by voice or by entering text on the display interface, and the electronic device can correspondingly save the corresponding relationship between the voiceprint feature of the successful registration and the identity "Mom".

[0082] The following describes the process of performing a storage operation when the target operation is a storage operation with a specific implementation method.

[0083] First, a wake-up voice message of "Hello, Harry" said by the user is received, the voiceprint features of the user are obtained through a pre-trained voiceprint model, and the voiceprint features of the user are matched with any pre-saved registered voiceprint features to determine whether there is a matching registered voiceprint feature.

[0084] Receive voice information to be recognized "Put 84 disinfectant in the kitchen cabinet" input by the user, perform voice recognition on the voice information to be recognized "Put 84 disinfectant in the kitchen cabinet", obtain text information of the voice information to be recognized, and match the text information with a pre-saved intent template, determine that the matched intent template is "Put * in *", and the target operation is a storage operation, then according to the character interval where each keyword in the matched template is located, obtain each keyword "84 disinfectant" and "kitchen cabinet" contained in the text information; output a prompt message for setting query permissions, and use the keywords "84 disinfectant" and "kitchen cabinet" as one piece of storage information; receive each target identity identifier that can perform storage information query as "Dad", "Grandma", and "Mom", and correspondingly save each target identity identifier "Dad", "Grandma", "Mom" and the above-mentioned storage information.

[0085] After each target identity and stored information are saved, if any user wants to query the stored information later, the identity of the querying user must be in the target identity before the stored information can be queried. Specifically, if the target operation is a query operation, performing the target operation on each keyword includes:

[0086] A query is performed according to each of the keywords. If target storage information containing each of the keywords is found, each target identity identifier corresponding to the target storage information is obtained. If the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, a reply message is generated and output according to each target keyword stored in the target storage information.

[0087] In the present application, when the target operation is a query operation, each keyword contained in the text information of the voice information to be recognized obtained in the above embodiment is queried. When the target storage information containing each keyword contained in the text information of the voice information to be recognized is queried, it means that there is storage information to be queried, and each target identity identifier corresponding to the target storage information is obtained. It is determined whether the identity identifier corresponding to the registered voiceprint feature that matches the voiceprint feature of the wake-up voice information obtained by the voiceprint model pre-trained in the above embodiment is included in the target identity identifier, so as to determine whether the user who inputs the voice information to be recognized is a user who can query the target storage information.

[0088] Specifically, if the identity corresponding to the matched registered voiceprint feature is included in the target identity, it means that the user who inputs the voice information to be recognized can query the target storage information, then the reply information is generated and output according to each target keyword stored in the target storage information.

[0089] For example, when a wake-up voice message of "Hello, Harry" is received from the user, the voiceprint features of the user are obtained through the pre-trained voiceprint model, and the voiceprint features of the user are matched with any pre-saved registered voiceprint features to determine that there is a matching registered voiceprint feature, and the identity identifier "Dad" corresponding to the pre-saved registered voiceprint feature is obtained.

[0090] After receiving voice information to be recognized "Where is 84 disinfectant placed" input by the user, voice recognition is performed on the voice information to be recognized "Where is 84 disinfectant placed", text information of the voice information to be recognized is obtained, and the text information is matched with a pre-saved intent template, and the matched intent template is determined to be "*where to place", and the target operation is a query operation, then each keyword "84 disinfectant" contained in the text information is obtained according to the character interval where each keyword in the matched template is located; a query is performed according to the keyword "84 disinfectant", and target storage information containing the keyword "84 disinfectant" is found, then each target identity identifier corresponding to the target storage information is obtained as "Dad", "Grandma", and "Mom", respectively; it is determined that the identity identifier "Dad" is included in the target identity identifiers "Dad", "Grandma", and "Mom", then a reply message is generated and output according to each target keyword in the target storage information.

[0091] In some possible embodiments, in order to improve user experience, the method further includes:

[0092] If the identity corresponding to the matched registered voiceprint feature is not included in the target identity, the text information of the voice information to be recognized is filtered, and the preset rejection response reply information is output.

[0093] During the specific implementation process, if the identity identifier corresponding to the matched registered voiceprint feature is not included in the target identity identifier, it means that the user who inputs the voice information to be recognized cannot query the target storage information. In this case, the voice information to be recognized can be filtered and the response can be rejected. However, in order to improve the user experience, the preset rejection response reply information can be output.

[0094] Still taking the above example, if the identity identifier corresponding to the matched registered voiceprint feature is "child", and it is determined that the identity identifier "child" is not included in the target identity identifier, the voice information to be recognized is filtered, and the preset rejection response reply information "Sorry, I can't tell you" is output.

[0095] Among them, each target identity, storage information and text information can also be saved accordingly, so as to query the storage information later, and the corresponding reply information can be directly generated and output according to the text information, that is, the reply information in audio format is generated and output according to the text information, and / or the text information is directly used as the reply information and output. For example, the text information is "alcohol is placed in the living room cabinet", and the keywords contained in the text information are determined to be "alcohol" and "living room cabinet", then the text information "alcohol is placed in the living room cabinet", the keywords "alcohol" and "living room cabinet" are used as a storage information. When the voice information to be recognized is received later to query the storage location of alcohol, the reply information in audio format is directly generated and output according to the text information "alcohol is placed in the living room cabinet", and / or the text information "alcohol is placed in the living room cabinet" is directly used as the reply information and output.

[0096] Because in the present application, when executing the target operation, it is also necessary to obtain the voiceprint features of the wake-up voice information through the pre-trained voiceprint model, and then perform the target operation on each keyword contained in the text information of the voice information to be recognized based on the voiceprint features, thereby protecting the privacy of the user and effectively preventing children from getting into danger after inquiring about the storage location of dangerous items.

[0097] In order to further improve the user experience, based on the above embodiments, in this application, the generating and outputting of reply information includes:

[0098] Obtaining the target location information of the user whose identity corresponds to the matched registered voiceprint feature;

[0099] According to the pre-stored position information of each playback device, the playback device whose position information is closest to the target position information is used as the target playback device, and the generated reply information is outputted through the target playback device.

[0100] In actual application scenarios, it may happen that when the reply information is output, the user is not in the current room, resulting in the inability to receive the output reply information in time, resulting in a weak sense of interaction with the user. For example, the user is currently inputting the voice information to be recognized in the living room, and happens to remember that there is rice simmering in the kitchen and needs to check it, or the child in the study calls her, etc. The user is no longer in the living room. If the output reply information is still output in the living room, the user may not be able to hear or see it. Therefore, in this application, the electronic device can also receive the current location information of each user sent by the radio frequency base station. After determining the identity identifier corresponding to the currently matched registered voiceprint feature, the electronic device obtains the target location information of the user of the identity identifier, thereby determining the playback device responsible for outputting the reply information.

[0101] In order to accurately determine the playback device responsible for outputting the reply information, in this application, the electronic device stores the location information and IP address of each playback device, such as 192.168.4.41, 192.168.4.42, 192.168.4.43, 192.168.4.44, 192.168.4.45, and 192.168.4.46. After the target location information is determined based on the above embodiment, the distance between the location information of each playback device and the target location information is obtained respectively, and the playback device corresponding to the smallest distance, that is, the playback device whose location information is closest to the target location information, is used as the target playback device, and the IP address of the target playback device is obtained, and the generated reply information is sent to the target playback device with the IP address for output. Still taking the above as an example, if the current user's target location information is in the kitchen, the reply information is output through the playback device in the kitchen whose location information is closest to the target location information.

[0102] Among them, a radio frequency positioning base station is installed in each room, and each user wears a device corresponding to his or her identity. Whenever a user wearing the wearable device enters a room where a radio frequency positioning base station is installed, the radio frequency positioning base station in the room will detect the identity corresponding to the wearable device, determine the location information of the identity, and send the identity and location information to the electronic device. At the same time, in order to accurately update the location information of each user with an identity, the radio frequency positioning base station can also send the time information, identity and location information to the electronic device.

[0103] Figure 3 A schematic diagram of an application scenario of an electronic device provided in some embodiments of the present application, such as Figure 3 As shown, the family consists of a living room, kitchen, bedroom, study and entrance hall, each of which is equipped with a smart speaker and a radio frequency positioning base station; the IP addresses of the smart speakers in the living room, kitchen, bedroom 1, bedroom 2, study and entrance hall are 192.168.4.41, 192.168.4.42, 192.168.4.43, 192.168.4.44, 192.168.4.45 and 192.168.4.46 respectively, and each smart speaker is a playback device; each family member wears a wearable device.

[0104] When a family member enters any room, the radio frequency positioning base station in the room will send the identity, location information and current time of the family member to the electronic device. The electronic device performs voiceprint recognition on the wake-up voice information and determines that there is a registered voiceprint feature that matches the voiceprint feature of the wake-up voice information. The identity corresponding to the matched registered voiceprint feature is "Mom". When the electronic device determines the intent template that matches the text information of the voice information to be recognized, and the corresponding intent is a query operation using voice notes, each keyword contained in the text information is obtained according to the character interval of each keyword in the matched intent template, and a query is performed based on each keyword.

[0105] When the target storage information containing each keyword is queried, and the identity "Mom" is included in the target identity corresponding to the target storage information, the target location information of the identity "Mom" is obtained. According to the pre-saved location information of each smart speaker, the smart speaker whose location information is closest to the target location information is used as the target smart speaker, such as the smart speaker installed in the living room in the figure, whose IP address is 192.168.4.41. The reply information generated according to each keyword contained in the text information of the voice information to be recognized is sent to the target smart speaker with IP address 192.168.4.41 for output.

[0106] It should be noted that the time information is the time when the radio frequency positioning base station determines the location information of the user corresponding to the identity identifier.

[0107] Since the present application can determine the playback device whose location information is closest to the target location information as the target playback device based on the target location information of the user identified by the matching registered voiceprint feature and the pre-saved location information of each playback device, and output the generated reply information through the target playback device, the user experience is improved.

[0108] Table 1 shows the extraction of parameters when the intent template for text information matching provided by some embodiments of the present application is a target operation of using voice notes. See Table 1:

[0109]

[0110]

[0111] As shown in Table 1, taking the text information "Put alcohol in the kitchen" as an example, the matching intent template is "Put * in *" with an ID of 1, and ID is an identification value used to distinguish each intent template. The intent corresponding to the matching intent template is the storage operation of using voice notes. If the matching intent template is not a word segmentation processing template, the corresponding identification is "0" (if the matching intent template is a word segmentation processing template, the corresponding identification is "1"), wherein the word segmentation processing template refers to the character interval where the keywords in the matching template are located, and may contain at least two consecutive keywords in the same character interval. If it is included, the keywords in each character interval need to be segmented, and the intent template is also a word segmentation processing template. If it is not included, it is not a word segmentation processing template. According to the character interval of each keyword in the matching intent template, if there are characters in each character interval in the text information, it means that the content of the text information is complete, and the corresponding identifier is "1". (If there are no characters in the character interval in the text information, it means that the content of the text information is incomplete, and the corresponding identifier is "0"). The service module to which the target operation of using voice notes corresponding to the target intent belongs is Audio_note.<star index='1' / > Indicates the target keyword "alcohol" in the obtained text information.<star index='2' / > The target keyword "kitchen" in the obtained text information is converted into an aiml file according to the storage operation of using voice notes and the relevant content of the intent template of "*where to go" in the above table by the Audio_note service module. The Audio_note.aiml file is:

[0112]

[0113]

[0114] Taking the text message "Where is the alcohol placed" as an example, the matching intent template is "*where is it placed" with ID 6. The intent corresponding to this matching intent template is the query operation of using voice notes. This matching intent template is not a word segmentation processing template, so the corresponding identifier is "0". According to the character interval of each keyword in the matching intent template, if there are characters in each character interval in the text information, it means that the content of the text information is complete, and the corresponding identifier is "1". The service module to which the target operation of using voice notes corresponding to this target intent belongs is Audio_note.<star index='1' / > Indicates the target keyword "alcohol" in the acquired text information.

[0115] The Audio_note service module in the above table is converted into an aiml file according to the storage operation of using voice notes and the relevant content of the intent template of "*where to go", and the Audio_note.aiml file is:

[0116]

[0117] The following is the code for template matching based on the third-party Python dependency package aiml:

[0118] Audio_note = aiml.Kernel() #Create a kernel object

[0119] Audio_note.learn('semantic_cloud / Audio_note.aiml')#Configure the corpus of the intention template for the target operation of the learned voice note

[0120] Audio_note.respond(”.join(re_str))#Find_person.respond() is a function that can get the corresponding matching results. re_str is the processed string of the text ("Put 84 disinfectant in the kitchen cabinet", "Put 84 disinfectant in the kitchen cabinet", etc.).

[0121] The above code can realize the self-learning of aiml template matching algorithm. After the text information is successfully matched to the intent version, corresponding processing is performed to extract parameters, such as items, locations, target operations, etc., and the corresponding function is called according to the parameters and the processing results are returned.

[0122] Figure 4 A flowchart of a specific storage operation using voice notes provided in some embodiments of the present application includes:

[0123] S401: receiving a wake-up voice message.

[0124] S402: Acquire the voiceprint features of the received wake-up voice information through the pre-trained voiceprint model.

[0125] Among them, S402 only needs to be executed after S401 and before S406.

[0126] S403: Acquire text information of the voice information to be recognized.

[0127] S404: Match the text information of the voice information to be recognized with any pre-saved intention template to determine whether there is a matching intention template, and the intention corresponding to the matching intention template is the storage operation of using voice notes. If so, execute S405, otherwise, execute S401.

[0128] S405: Obtain each keyword included in the text information according to the character interval where each keyword is located in the matched intent template.

[0129] S406: Match the voiceprint feature obtained in S402 with any pre-saved registered voiceprint feature to determine whether there is a matching registered voiceprint feature. If so, execute S407; otherwise, execute S401.

[0130] S407: Output prompt information for setting query permissions, and save each keyword as a piece of storage information.

[0131] S408: receiving each target identity that is set to be able to perform storage information query, and correspondingly saving each target identity and storage information.

[0132] Figure 5 A flowchart of a specific query operation using voice notes provided in some embodiments of the present application includes:

[0133] S501: receiving a wake-up voice message.

[0134] S502: Acquire the voiceprint features of the received wake-up voice information through the pre-trained voiceprint model.

[0135] Among them, S502 only needs to be executed after S501 and before S506.

[0136] S503: Acquire text information of the voice information to be recognized.

[0137] S504: Match the text information of the voice information to be recognized with any pre-saved intent template to determine whether there is a matching intent template, and the intent corresponding to the matching intent template is a query operation using voice notes. If so, execute S505, otherwise, execute S501.

[0138] S505: Obtain each keyword included in the text information according to the character interval where each keyword is located in the matched intent template.

[0139] S506: Match the voiceprint feature with any pre-saved registered voiceprint feature to determine whether there is a matching registered voiceprint feature. If so, execute S507; otherwise, execute S501.

[0140] S507: Query according to each keyword to determine whether the target storage information containing each keyword is found. If so, execute S508; otherwise, execute S501.

[0141] S508: Obtain each target identity identifier corresponding to the target storage information, and determine whether the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier. If so, execute S510; otherwise, execute S509.

[0142] S509: Filter the text information of the voice information to be recognized, output the preset rejection response reply information, and then execute S501.

[0143] S510: Obtain the target location information of the user whose identity corresponds to the matched registered voiceprint feature.

[0144] S511: According to the pre-stored position information of each playback device, a playback device whose position information is closest to the target position information is used as a target playback device.

[0145] S512: According to each target keyword stored in the target storage information, the generated reply information is output through the target playback device, and then S501 is executed.

[0146] The present application also provides a voice note taking device, Figure 6 A schematic diagram of the structure of a voice note-taking device provided in some embodiments of the present application, the device comprising:

[0147] An acquisition unit 601 is used to acquire text information of the voice information to be recognized;

[0148] A matching unit 602, configured to match the text information with any pre-stored intent template;

[0149] Processing unit 603 is used to obtain each keyword contained in the text information according to the character interval where each keyword in the matching intent template is located, and perform the target operation on each keyword if there is a matching intent template and the intent corresponding to the matching intent template is the target operation of using voice notes.

[0150] In some embodiments, the processing unit 603 is also used to obtain the voiceprint features of the received wake-up voice information through a pre-trained voiceprint model before performing the target operation on each keyword; match the voiceprint features with any pre-saved registered voiceprint features, and if there is a matching registered voiceprint feature, execute the subsequent steps of performing the target operation on each keyword.

[0151] In some embodiments, the processing unit 603 is specifically used to output prompt information for setting query permissions if the target operation is a storage operation, and to treat each keyword contained in the text information of the voice information to be recognized as a storage information; receive each target identity identifier that is set to perform the storage information query, and correspondingly save each target identity identifier and the storage information.

[0152] In some embodiments, the processing unit 603 is specifically used to, if the target operation is a query operation, perform a query based on each of the keywords, and if the target storage information containing each keyword is queried, obtain each target identity identifier corresponding to the target storage information; if the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, generate and output reply information based on each target keyword stored in the target storage information.

[0153] In some embodiments, the processing unit 603 is specifically used to obtain the current target location information of the user whose identity corresponds to the matched registered voiceprint feature; based on the pre-saved location information of each playback device, the playback device whose location information is closest to the target location information is used as the target playback device, and the generated reply information is output through the target playback device.

[0154] In some embodiments, the processing unit 603 is further configured to filter the text information of the voice information to be recognized and output a preset rejection response reply message if the identity corresponding to the matched registered voiceprint feature is not included in the target identity.

[0155] For the concepts, explanations, detailed descriptions and other steps related to the technical solution provided by the embodiment of the present invention involved in the voice note taking device, please refer to the description of these contents in the aforementioned method or other embodiments, which will not be repeated here.

[0156] Since in the present application, when it is determined that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using voice notes, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making it more flexible and convenient to perform the target operation when using voice notes, and simplifying the process for determining each keyword contained in the text information of the voice information to be recognized, thereby improving the efficiency of the voice note-taking process.

[0157] like Figure 7 This is a schematic diagram of an electronic device structure provided by some embodiments of the present application. Based on the above embodiments, the present application also provides an electronic device, such as Figure 7As shown, it includes: a processor 701, a communication interface 702, a memory 703 and a communication bus 704, wherein the processor 701, the communication interface 702, and the memory 703 communicate with each other through the communication bus 704;

[0158] The memory 703 stores a computer program. When the program is executed by the processor 701, the processor 701 performs the following steps:

[0159] Obtain text information of the voice information to be recognized;

[0160] Matching the text information with any pre-saved intent template;

[0161] If there is a matching intent template, and the intent corresponding to the matching intent template is the target operation of using voice notes, then each keyword contained in the text information is obtained according to the character interval where each keyword in the matching intent template is located, and the target operation is performed on each keyword.

[0162] In some embodiments, the processor 701 is further used to obtain the voiceprint features of the received wake-up voice information through a pre-trained voiceprint model before performing the target operation on each keyword; match the voiceprint features with any pre-saved registered voiceprint features, and if there is a matching registered voiceprint feature, execute the subsequent steps of performing the target operation on each keyword.

[0163] In some embodiments, the processor 701 is specifically used to output prompt information for setting query permissions if the target operation is a storage operation, and to treat each keyword contained in the text information of the voice information to be recognized as a storage information; receive each target identity identifier that is set to perform the storage information query, and correspondingly save each target identity identifier and the storage information.

[0164] In some embodiments, the processor 701 is specifically used to, if the target operation is a query operation, perform a query according to each of the keywords, and if the target storage information containing each keyword is queried, obtain each target identity identifier corresponding to the target storage information; if the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, generate and output reply information according to each target keyword stored in the target storage information.

[0165] In some embodiments, the processor 701 is specifically used to obtain the current target location information of the user whose identity corresponds to the matched registered voiceprint feature; based on the pre-saved location information of each playback device, the playback device whose location information is closest to the target location information is used as the target playback device, and the generated reply information is output through the target playback device.

[0166] In some embodiments, the processor 701 is further configured to filter the text information of the voice information to be recognized and output a preset rejection response reply message if the identity corresponding to the matched registered voiceprint feature is not included in the target identity.

[0167] Since the principle of solving the problem by the above electronic device is similar to that of the voice note-taking method, the implementation of the above electronic device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0168] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0169] The communication interface 702 is used for communication between the electronic device and other devices.

[0170] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0171] The above-mentioned processor can be a general-purpose processor, including a central processing unit, a network processor (Network Processor, NP), etc.; it can also be a digital signal processing processor (Digital Signal Processing, DSP), an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc.

[0172] Since the present application determines that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using voice notes, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making it more flexible and convenient to perform the target operation when using voice notes, and simplifies the process for determining each keyword contained in the text information of the voice information to be recognized, thereby improving the efficiency of the voice note-taking process.

[0173] On the basis of the above embodiments, some embodiments of the present application further provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program executable by a processor, and when the program runs on the processor, the processor implements the following steps when executing:

[0174] Obtain text information of the voice information to be recognized;

[0175] Matching the text information with any pre-saved intent template;

[0176] If there is a matching intent template, and the intent corresponding to the matching intent template is the target operation of using voice notes, then each keyword contained in the text information is obtained according to the character interval where each keyword in the matching intent template is located, and the target operation is performed on each keyword.

[0177] In some embodiments, the target operation is a storage operation or a query operation.

[0178] In some embodiments, before performing the target operation on each keyword, the method further includes:

[0179] The voiceprint features of the received wake-up voice information are obtained through the pre-trained voiceprint model;

[0180] The voiceprint feature is matched with any pre-saved registered voiceprint feature. If there is a matching registered voiceprint feature, the subsequent step of performing the target operation on each keyword is performed.

[0181] In some embodiments, if the target operation is a storage operation, performing the target operation on each keyword includes:

[0182] Outputting prompt information for setting query permissions, and storing each keyword contained in the text information of the voice information to be recognized as a piece of information;

[0183] Each target identity identifier that is set to perform the storage information query is received, and each target identity identifier and the storage information are correspondingly saved.

[0184] In some embodiments, if the target operation is a query operation, performing the target operation on each keyword includes:

[0185] A query is performed according to each of the keywords. If target storage information containing each of the keywords is found, each target identity identifier corresponding to the target storage information is obtained. If the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, a reply message is generated and output according to each target keyword stored in the target storage information.

[0186] In some embodiments, generating and outputting reply information includes:

[0187] Obtaining the target location information of the user whose identity corresponds to the matched registered voiceprint feature;

[0188] According to the pre-stored position information of each playback device, the playback device whose position information is closest to the target position information is used as the target playback device, and the generated reply information is outputted through the target playback device.

[0189] In some embodiments, the method further comprises:

[0190] If the identity corresponding to the matched registered voiceprint feature is not included in the target identity, the text information of the voice information to be recognized is filtered, and the preset rejection response reply information is output.

[0191] Since the principle of solving the problem by the computer-readable medium provided above is similar to that of the voice note-taking method, after the processor executes the computer program in the computer-readable medium above, the implemented steps can refer to the implementation of the method, and the repeated parts will not be repeated.

[0192] Since the present application determines that there is an intention template that matches the text information of the voice information to be recognized, and the matched intention template is the target operation of using voice notes, each keyword contained in the text information of the voice information to be recognized can be determined according to the character interval where each keyword in the matched intention template is located, and the target operation is performed on each keyword, thereby making it more flexible and convenient to perform the target operation when using voice notes, and simplifies the process for determining each keyword contained in the text information of the voice information to be recognized, thereby improving the efficiency of the voice note-taking process.

[0193] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0194] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0195] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0196] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0197] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A voice note-taking method, characterized in that: The method comprises: Obtain text information of the voice information to be recognized; Matching the text information with any pre-saved intent template; If there is a matching intent template, and the intent corresponding to the matching intent template is a target operation of using voice notes, then according to the character interval where each keyword in the matching intent template is located, each keyword included in the text information is obtained in the corresponding character interval, and the target operation is performed on each keyword; Wherein, if the target operation is a storage operation, a prompt message for setting the query permission is output, and each keyword contained in the text information is used as a storage information, each target identity identifier that is set to be able to perform the storage information query is received, and each target identity identifier and the storage information are correspondingly saved; If the target operation is a query operation, the storage information containing each keyword in the text information is searched, and the storage information containing each keyword in the text information is used as the target storage information. According to each target keyword stored in the target storage information and the preset reply template, the reply information is generated and output.

2. The method according to claim 1, characterized in that: Before performing the target operation on each keyword, the method further includes: The voiceprint features of the received wake-up voice information are obtained through the pre-trained voiceprint model; The voiceprint feature is matched with any pre-saved registered voiceprint feature. If there is a matching registered voiceprint feature, the subsequent step of performing the target operation on each keyword is performed.

3. The method according to claim 2, characterized in that If the target operation is a query operation, performing the target operation on each keyword includes: A query is performed according to each of the keywords. If target storage information containing each of the keywords is found, each target identity identifier corresponding to the target storage information is obtained. If the identity identifier corresponding to the matched registered voiceprint feature is included in the target identity identifier, a reply message is generated and output according to each target keyword stored in the target storage information.

4. The method according to claim 3, characterized in that The generating and outputting of reply information includes: Obtaining the target location information of the user whose identity corresponds to the matched registered voiceprint feature; According to the pre-stored position information of each playback device, the playback device whose position information is closest to the target position information is used as the target playback device, and the generated reply information is outputted through the target playback device.

5. The method according to claim 3, characterized in that: The method further comprises: If the identity corresponding to the matched registered voiceprint feature is not included in the target identity, the text information of the voice information to be recognized is filtered, and the preset rejection response reply information is output.

6. A voice note taking device, characterized in that: The device comprises: An acquisition unit, used for acquiring text information of the voice information to be recognized; A matching unit, used to match the text information with any pre-saved intention template; a processing unit, configured to obtain each keyword included in the text information according to the character interval where each keyword in the matching intent template is located, and perform the target operation on each keyword if there is a matching intent template and the intent corresponding to the matching intent template is a target operation of taking a voice note; Wherein, if the target operation is a storage operation, a prompt message for setting the query permission is output, and each keyword contained in the text information is used as a storage information, each target identity identifier that is set to be able to perform the storage information query is received, and each target identity identifier and the storage information are correspondingly saved; If the target operation is a query operation, the storage information containing each keyword in the text information is searched, and the storage information containing each keyword in the text information is used as the target storage information. According to each target keyword stored in the target storage information and the preset reply template, the reply information is generated and output.

7. An electronic device, characterized in that: The electronic device comprises at least a processor and a memory, and the processor is used to implement the steps of the voice note-taking method as claimed in any one of claims 1 to 5 when executing a computer program stored in the memory.

8. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements the steps of the voice note-taking method according to any one of claims 1 to 5.

Images