TWI754816B - Method for processing voice command and system thereof - Google Patents

Abstract

A method for processing voice command and a system thereof are provided. The method is applied to a service host that receives voice message from a terminal device via a network. In the service host, the terminal device can be recognized according to a packet source, and a registered user correlated to the terminal device can be obtained. The voice message is compared with a voiceprint corresponding to the user so as to verify the user. A question and an answer are obtained for forming a question voice provided for the user to answer. A reply voice is then transmitted to the service host by the terminal device. The system verifies the user’s identity based on the voiceprint comparison. The system determines if the user is a real person according to the reply voice. The system provides a voice service according to the voice message since the system confirms a real person.

Description

Voice command processing method and system

The invention relates to a technology for controlling a remote service by voice, in particular to a method and a system for simultaneously judging whether it is a real person and authenticating the person when processing a voice command.

The technology of implementing a voice assistant through a specific computer device (such as a mobile device or a smart speaker) is becoming more and more mature. For example, a voice command can be sent directly to the computer device in the form of voice, which is converted into a voice packet by the computer device and sent to the server. At the end, the server performs semantic analysis to obtain the meaning of the voice. The software program in the server can compare the database to obtain the reply voice, or use artificial intelligence (Artificial Intelligence) to respond to the user's voice.

The general purpose of using a voice assistant is to obtain the required information through voice inquiry, such as inquiring about the weather, nearby restaurants, movie times, or inquiring network information. Actions that require identity verification require more rigorous authentication technology using voice, and there is no solution for effectively verifying identity and confirming whether the user is a real person (not a robot) in the prior art.

The disclosure discloses a voice command processing method and a system for implementing the method. The system includes a service host. The service host can provide corresponding services according to the voice commands generated by the user at the terminal, and reply to the user with voice, so that the user can Information is obtained without manual operation, and in particular, the entire process of voice command processing also implements user identity authentication and a judgment procedure to confirm that the user is a real person.

According to the embodiment of the voice command processing system, the system proposes a service host, which can be provided with a database, which mainly includes the voiceprint data of a plurality of users, and can record the question bank to be interacted with the users. The service host can be a computer host , cluster or servo system implementation, and the voice command processing method runs in the service host.

According to the embodiment of the voice command processing method, the service host receives a voice message from a terminal device, such as a voice message generated by a user through a smart speaker or a computer device. To confirm the user who sent the voice message, then one or more questions and answers can be generated according to the information provided by the user during registration, or from the preset question bank, to form a voice for one or more questions raised by the user .

Afterwards, the service host transmits the question voice to the terminal device through the network, and the terminal device broadcasts the question voice. When the user responds to each question voice, the service host receives the reply voice corresponding to each question voice. In this way, in the service host, the voiceprint comparison result can be obtained to determine whether it is the user who is logged in to the service host, to confirm the identity of the user, and to obtain the reply voice corresponding to each question voice, so as to determine whether it is the user himself. A real person.

Further, in the method, the source of a packet can be parsed from the voice message, such as the media access control address or the network address of the terminal device, so that the service host can obtain the information registered in the service host according to the source of the packet. The terminal device can be compared with one or more users registered in the service host. Once the voiceprint is compared, the result can be used to obtain the user who issued the question-and-answer voice.

Preferably, the questions and answers can be preset questions and answers in a question bank, or one or more questions and answers corresponding to the user are generated according to the data of the user when the user registers with the service host.

Preferably, the terminal device may be a smart speaker located near the user, or may be a computer device for processing voice messages, such as the user's mobile device.

Further, when the service host receives the voice message, it can perform a semantic analysis, and correspondingly obtain a service item requested by the user, and can provide a service voice corresponding to this service item. Similarly, the service voice will be available through Terminal device broadcast.

For a further understanding of the features and technical content of the present invention, please refer to the following detailed descriptions and drawings of the present invention. However, the drawings provided are only for reference and description, and are not intended to limit the present invention.

The following are specific embodiments to illustrate the embodiments of the present invention, and those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the concept of the present invention. In addition, the accompanying drawings of the present invention are merely schematic illustrations, and are not drawn according to the actual size, and are stated in advance. The following embodiments will further describe the related technical contents of the present invention in detail, but the disclosed contents are not intended to limit the protection scope of the present invention.

It should be understood that although terms such as "first", "second" and "third" may be used herein to describe various elements or signals, these elements or signals should not be limited by these terms. These terms are primarily used to distinguish one element from another, or a signal from another signal. In addition, the term "or", as used herein, should include any one or a combination of more of the associated listed items, as the case may be.

The disclosure discloses a technology for controlling remote services by voice. The proposed voice command processing method and system allow users to obtain services provided by the remote system by voice in a program with security verification, such as using voice to generate and query themselves. Financial account balances, transfers, data inquiries, message transmission, etc., in the process, especially the use of voice and voiceprint features to authenticate the user's identity, and can simultaneously determine whether it is a real person, not someone else's counterfeit or intelligent (AI) robot. voice. In this way, the voice command processing method can allow the system processing the program to simultaneously process whether it is a real person and verify the identity of the registered user according to the voice information, so that the system can provide the requested service accordingly.

Referring first to the schematic diagram of the application of the voice command processing system shown in FIG. 1, this figure shows a user 10, and a terminal device for processing voice commands is arranged nearby, which is used to process the voice messages sent by the user 10 and can be transmitted through the network. The voice message is sent to a specific destination and is also used to process the voice message received from the outside. The terminal device, such as the smart speaker 12 shown in the figure, can also be a mobile device or a personal computer operated by the user 10. , processing, broadcasting) computer equipment.

In this example, the user 10 is located within the range where the smart speaker 12 can receive voice messages, just like in the living room or the room, the smart speaker 12 plays the role of an Internet of Things (IoT) and is connected to the network at any time, so that the system uses One smart speaker 12 can process multiple voice services at the same time, that is, through the smart speaker 12, the system can provide corresponding voice services according to the voice messages of different users. This example is shown as being connected to an external service host 14 . Through the smart speaker 12 , the user 10 can use voice to generate an instruction to request a service to the service host 14 , and can also receive voice messages from the service host 14 through the smart speaker 12 .

According to the illustrated example, the smart speaker 12 can connect to a financial institution website to inquire about the account balance through a specific application programming interface (eg, OpenAPI), which is shown as the service host 14 in this example. When the user 10 makes a voice: "I want to check the account balance", the smart speaker 12 receives the voice message. If it is a voice command that can be processed internally, the smart speaker 12 can process the voice command according to its software program, for example Voice inquiries about the weather (connecting to a specific server), the user's personal calendar, inquiring information, playing music, etc.; if the voice command is about the service items processed by the specific service host 14, the voice message will be combined with the information of the smart speaker 12 (device information, network road information, etc.), or including user identification information and specific information, after the package processing, a voice packet is formed, which is transmitted to the service host 14 through the network, and then the service host 14 processes the voice message, including deriving the corresponding terminal from the voice packet. Device information, and request commands for specific services by semantically analyzing voice messages.

This example shows that the reply voice generated by the service host 14 is replied to the user 10 through the smart speaker 12, and the identity of the user 10 can also be confirmed by this, such as asking: "What is your ID number?", then the user 10. Answer: "Axxxxxxx". Similarly, the voice message is packaged and then sent to the service host 14 through the smart speaker 12, and the service host 14 authenticates the user's identity according to the data recorded in the database, which also includes voiceprint recognition. User identity. Finally, if the authentication is passed, the service host 14 generates a reply voice, and the smart speaker 12 replies: "The balance is XXXX".

The scenario is an implementation of a voice command processing system, and FIG. 2 then shows a schematic diagram of an embodiment of a system architecture for running the voice command processing method.

In this schematic diagram of the system architecture, the user 20 is provided with a terminal device 22, one terminal device 22 can serve one or more users 20, and the terminal device 22 can be a smart speaker or a computer device executing a specific voice assistant program. The device 22 is connected to the service host 26 through the network 24. The service host 26 can provide a variety of service items for a server, cluster or server system related to financial services, and can obtain the result after semantic analysis according to the voice message sent by the user 20. The requested service item.

The service host 26 is provided with a database 28, which can record a question bank and the voiceprint data corresponding to a plurality of users. The question bank is mainly used to generate one or more questions and answers. In addition to the question bank, it also includes random questions other than the question bank. , such as math problems, the service host 26 provides one or more questions raised by the user 20. With the question bank, the service host 26 can also generate a question corresponding to a certain user 20 according to the user data, so that the system can respond to the user 20. 20 A person who asks a question and can judge whether he is real or not based on the response. It should be mentioned here that the question asked by the service host 26 can be a preset question in the question bank, and has preset answers, such as asking the user 20 one or more math questions, historical questions, current affairs questions, scientific questions, etc. Let the user 20 answer a question, and the system can intelligently process the answer of the user 20 to determine whether it is a real person; the question asked by the service host 26 can also be a question related to the user 20, such as when the user 20 registers the service host 26. Fill in the birthday, address, contact person, educational background, etc., and then ask the user 20. Similarly, the system can also intelligently process the answer of the user 20 to determine whether it is a real person.

In the service host 26, the processor executes a specific software instruction set, so that the voice command processing method is executed. According to the function to be executed, a source identification unit 263 may be included, which can identify the source after parsing the information recorded in the voice packet. When the terminal device 22 registers When serving the host 26, a database list is created. If the network address (IP address) or media access control address (MAC address) of the terminal device 22 is obtained from the voice packet, it can be obtained by comparing the registration data corresponding one or more users.

The service host 26 includes a voiceprint comparison unit 265, which uses a software program to compare the received voice messages, sample audio signals from them, deal with signal noise problems to remove interference, gain control, and extract voiceprint features. The voiceprint data created when the user 20 registers with the service host 26 can be used to authenticate the identity of the user 20 .

The service host 26 includes a real question bank unit 267, which uses a software program to obtain questions and answers from the database 28, and forms a question and answer voice for asking the user 20, and broadcasts it to the user 20 through the terminal device 22. It can be judged whether it is issued by a real person. 's voice message. In addition, the voiceprint matching unit 265 can be used to sample the voice response from the user 20 to obtain the voiceprint feature, thereby authenticating the identity of the user 20 .

The service host 26 utilizes the identity authentication unit 269 to perform the identity authentication of the user 20. In the voice command processing system, one or more corresponding users can be identified first according to the information of the terminal device 22, which can be more efficiently compared. The voiceprint data of these users is used to authenticate the user's identity.

The service host 26 can also be a server that provides specific services at the same time, but it can also cover the implementation of providing services with another server. The service host 26 can provide corresponding services according to the voice commands generated by the user at the terminal, and Replying to the user with voice enables the user to obtain services without manual operation, and the entire voice command processing process can simultaneously perform user identity authentication and confirm that the user is a real person.

The voice command processing method performed in the service host can be referred to the flowchart shown in FIG. 3, wherein the described method is applied to the voice command processing system, and the system provides one or more questions and answers, in addition to the information provided by the service host. The question bank recorded in the database not only interacts with users, but also includes some random questions, and then uses the user's voiceprint data in the database to authenticate the user's identity. The method flow is as follows. The sequence of steps does not limit the implementation of the invention, but can cover Embodiments in which the order is changed to achieve the same object.

At the beginning, as in step S301, the service host receives a voice message from a certain terminal device. As in step S303, in the service host, a software program can be used to perform a voiceprint comparison to confirm the user who sent the voice message. At this time, In step S305, the software program running in the service host provides one or more questions and answers, including questions and answers obtained from a database question bank, or randomly formed questions and answers, to form a voice in response to the user's question . The selection of questions can be from question banks, or random questions such as mathematics, or they can form question voices based on the information associated with the user. until real results.

Next, in step S307, the question voice is converted into a network packet and sent to the terminal device, and the terminal device converts it into voice and broadcasts it. The user responds to the question voice and generates a reply voice, which is then sent back to the terminal device. The service host, in step S309, receives the reply voice corresponding to the question voice via the network by the service host.

In step S311, the voiceprint feature is parsed from the message of the reply voice, which can be used to confirm the user's identity information, and at the same time judge whether it is a real person according to the state of the reply voice (one or more).

Applying the above voice command processing method flow, the system can further allow the user to issue commands using voice, and let the service host perform semantic analysis to obtain the corresponding service items in the voice command, especially the service items that require identity authentication, so that the system will use voice. The message performs human judgment and user authentication.

FIG. 4 shows a flowchart of an embodiment of applying a voice command processing method.

The voice command processing method is applied to the service host, and starts to receive a voice message from a terminal device as in step S401. In this embodiment, as in step S403, the source of a packet can be parsed from the voice message, for example, the terminal device's The network hardware information (such as MAC address) is used to determine the object of the service. One way is to correspond to one or more users associated with the terminal device. In step S405, the associated information records that the terminal device is compared with a or multiple users registered in the service host, the user list can be established when the service host registers the service, so that the terminal device binds one or more users to limit the service objects from this device.

In step S407, the service host can also perform voiceprint comparison according to the received voice information, and after obtaining the voiceprint comparison result, it can be used to determine whether it is the user who logs in to the service host, so as to confirm the identity of the user. It is worth mentioning that, when the service host first obtains one or more users associated with the terminal device by the packet source, the object range of the voiceprint comparison can be limited, which makes the whole authentication process more efficient.

In step S409, after the voiceprint is compared, the comparison result is used to determine whether the received voice message is one of the users bound to the terminal device for identity authentication. In step S411, it is determined whether If the identity authentication is passed, if the identity authentication is not passed (No), as shown in step S413, the service host may generate an ending voice and notify the user through the terminal device.

If the identity authentication is passed (Yes), step S415 is executed to perform semantic analysis on the received voice message to determine the requested service item. In step S417, when the system judges that the service item to be provided needs to be further authenticated whether the user is a real person or not, a real person judgment procedure can be performed, and the question voice can be generated from the question bank, that is, according to the question bank recorded in the database, or randomly generated Non-question bank questions, generate one or more questions and answers, form one or more question voices to the user, and can transmit voice messages to the terminal device multiple times, the terminal device broadcasts the question voice, and the user answers, The next step is step S419, the service host receives the reply voice corresponding to each question voice, and can determine whether it is a real person (step S421), if it is judged that it is not a real person (No) according to the incoming and outgoing voice, that is, step S413 is executed, and the service host can generate an ending voice , and inform the user through the terminal device.

On the contrary, if it is confirmed that the source of the voice is a real person (yes), that is, as in step S423, the service host can provide services according to the service items judged in the above steps, for example, provide a query on the balance of a financial account, generate a service voice, and transmit it to the terminal through the network. device, the service voice is broadcast to the user through the terminal device.

It should be mentioned here that the sequence of comparing the voiceprints in step S415, step S417 and step S407 is not limited, and may be performed simultaneously.

FIG. 5 then describes an embodiment of a voice command processing method with multiple processes. The method runs on the user 51 , the terminal device 52 , the telecommunication server 53 and the service host 54 . Provide network service, connect with the service host 54, the service host 54 can be a computer host, or a system formed by a plurality of servers to authenticate the identity of the user 51, provide real person judgment, and provide services.

The process starts when the user 51 starts to issue a voice command, the generated voice message is received by the terminal device 52 (step S501 ), a voice packet is formed, the voice packet is sent to the telecommunication server 53 (step S502 ), and then the voice packet is forwarded to the service host 54 (step S503).

In the service host 54 , analyze the source according to the voice packets, perform semantic analysis, identity recognition, and generate question and answer questions, among which one or more questions and answers can be randomly generated or obtained from the question bank, and can be based on the user 51's original question. The user identity and service items are obtained from the voice message, so the questions and answers can also be generated by comparing the user information and service items (step S504). It is sent to the telecommunications server 53 (step S505 ), a message packet is formed and then forwarded to the terminal device 52 (step S506 ), and the terminal device 52 broadcasts it to the user 51 (step S507 ), and the user 51 replies to form a voice message. It will be received by the terminal device 52 (step S508 ), and transferred to the service host 54 through the telecommunication server 53 (step S509 ) (step S510 ).

In the service host 54, the reply voice generated by the user 51 will be verified, and if it is confirmed as a real person, the identity authentication will be completed from the above steps, and the service items can be further responded to (step S511), and a service message will be formed, which is passed through the telecommunication server. 53 (step S512 ) forwarding the service message to the terminal device (step S513 ), on the user side, the terminal device 52 broadcasts the service message (step S514 ), ie, the current voice service process is completed.

Next, according to an embodiment, within a certain time limit set by the system, the user can continuously send voice messages to obtain the service. In another embodiment, the software program in the service host runs continuously, and can continuously run the voice command processing method every time the user makes a voice, including continuously using the voice message to verify whether it is a real person and the identity of the user, and to provide semantic meaning in the voice The requested service item.

It should be mentioned here that the processes described in the above embodiments are not limited to the sequence of steps described in the embodiments, and all processes that achieve the same purpose through simple replacement are covered by the present invention, such as the real person judgment executed in the service host , voiceprint comparison and identity authentication and other steps can be modified according to the actual situation.

The contents disclosed above are only preferred feasible embodiments of the present invention, and are not intended to limit the scope of the present invention. Therefore, any equivalent technical changes made by using the contents of the description and drawings of the present invention are included in the application of the present invention. within the scope of the patent.

10: User 12: Smart Speakers 14: Service Host 20: User 22: Terminal device 24: Internet 26: Service Host 263: Source Identification Unit 265: Voiceprint comparison unit 267: Real Question Bank Unit 269: Authentication unit 28:Database 51: User 52: Terminal device 53: Telecom Server 54: Service Host Steps S301-S311: voice command processing flow Steps S401-S423: voice command application process Steps S501-S514: voice command processing flow

FIG. 1 is a schematic diagram showing a situation of applying a voice command processing system;

2 shows a schematic diagram of an embodiment of a system architecture for running a voice command processing method;

3 shows an embodiment flow chart of a method for processing voice commands for a service host;

FIG. 4 shows a flowchart of an embodiment of applying a voice command processing method;

FIG. 5 shows a flowchart of an embodiment of a voice command processing method.

10: User

12: Smart Speakers

14: Service Host

Claims (5)

A voice command processing method, applied to a service host, comprising: receiving a voice message from a terminal device, and performing voiceprint comparison with the voice message and the voiceprint data created when a plurality of users register the service host, To confirm and authenticate a user who sent the voice message, the voice message obtains a request for a service item after a semantic analysis, and provides a service voice corresponding to the service item, and the service voice is broadcast through the terminal device; from the The voice message parses out a packet source, and according to the packet source, the terminal device registered with the service host is obtained; the user who sent the voice message is obtained according to the result of the voiceprint comparison, and a message is provided for the user. or multiple questions and answers, forming one or more question voices posed by the user, the one or more questions and answers include questions and answers preset in a question bank of a database, or according to the user One or more questions and answers corresponding to the user are generated from the data when the service host is registered; the one or more questions are transmitted to the user's terminal device, and the one or more questions are broadcast by the terminal device. a question voice; and after the user responds to each question voice, receiving a reply voice corresponding to each question voice; wherein, when the voiceprint comparison result is obtained, it is determined whether it is the user who is logged in to the service host, so as to confirm The identity of the user, and the answering voice corresponding to each question voice is obtained, wherein the user is asked to answer a question with a preset answer in the question bank through the service host, or is generated when registering the service host and is related to the user. The question requires the user to answer, and then verify the reply voice generated by the user to determine whether it is a real person; wherein, if it is confirmed that the source of the user's voice is the real person, the service host responds to the service item to provide services, forming a service message, forwarding the service message to the terminal device through a telecommunication server, and broadcasting the service message by the terminal device The voice of the service is given to the user; wherein, the voice message received by the terminal device is that the user uses voice to issue a command, and the service host performs the semantic analysis to obtain the corresponding service item in the voice command. The voice command processing method according to claim 1, wherein a network address or a media access control address of the terminal device is obtained from the voice message to obtain the terminal device registered with the service host , and compare the registration data to obtain the corresponding one or more users. A voice command processing system, comprising: a service host with a database in which voiceprint data corresponding to a plurality of users is recorded, a voice command processing method running in the service host, the method comprising: receiving a voice command from a terminal device Voice message, compare the voice message with the voiceprint data created when the multiple users registered the service host to confirm and authenticate a user who sent the voice message, the voice message is subjected to a semantic analysis Then, a service item is requested, and a service voice corresponding to the service item is provided, and the service voice is broadcast through the terminal device; a packet source is parsed from the voice message, and according to the packet source, the service voice registered on the service host is obtained. The terminal device; obtains the user who sent the voice message based on the result of the voiceprint comparison, and provides one or more questions and answers for the user to form one or more question voices for the user , the one or more questions and answers include questions and answers preset in the question bank of the database, or one or more questions and answers corresponding to the user are generated according to the data of the user when the user registers with the service host answer; send the one or more question voices to the terminal device of the user, and the terminal device broadcasts the one or more question voices; and through the user responding to the question voices, receive the corresponding question voices Reply voice; wherein, when the voiceprint comparison result is obtained, it is determined whether it is the user who is logged in to the service host to confirm the identity of the user, and to obtain the answer voice corresponding to the voice of each question, which is obtained through the service host Asking the user to answer a question with a preset answer in the question bank, or asking the user to answer a question related to the user, and verifying the answering voice generated from the user to determine whether it is a real person; wherein , if it is confirmed that the user's voice source is the real person, the service host responds to the service item to provide services, forms a service message, and transmits the service message to the terminal device through a telecommunication server, and the terminal device broadcasts the service message The voice of the service is given to the user; wherein, the voice message received by the terminal device is that the user uses voice to issue a command, and the service host performs the semantic analysis to obtain the corresponding service item in the voice command. The voice command processing system of claim 3, wherein the terminal device is a smart speaker located near the user, or a computer device for processing the voice message. The voice command processing system of claim 3, wherein a network address or a media access control address of the terminal device is obtained from the voice message to obtain the terminal device registered with the service host , and compare the registration data to obtain the corresponding one or more users.

Images