JP4679919B2 - Distributed speaker verification device, distributed speaker verification terminal device, and computer program - Google Patents

Description

  The present invention relates to a distributed speaker verification device, a distributed speaker verification terminal device, and a computer program.

In recent years, it has become possible to access the Internet via a mobile phone network from a user terminal such as a mobile phone, and user authentication at the time of user access to a network data service in such a highly flexible communication mode However, the use of personal authentication technology (speaker verification technology) using voice, which is one of biometric information, is being studied. For example, a client / server type distributed speaker verification method has been proposed as a speaker verification method having excellent compatibility with data services. In the distributed speaker verification method, the user terminal converts the input voice into a voice feature amount suitable for speaker verification and transmits it to the speaker verification server. The speaker verification server uses the voice feature amount received from the user terminal to perform user voice registration processing and authentication processing. In addition, the speech feature (speech encoding method) conforms to the standard method (DSR: Distributed Speech Recognition) defined by the European Telecommunications Standards Institute (ETSI) and VoIP (Voice over Internet Protocol). It has been proposed to use the method.
ITU-T, "Question 15/16-Distributed Speech Recognition (DSR) and Distributed Speaker Verification (DSV)", [online], October 21, 2004, World Wide Web Consortium, [February 1, 2005 search ] Internet <URL: http://www.itu.int/ITU-T/2001-2004/com16/sg16-q15.html> Nuku Fukutsugi, Satoshi Kuroiwa, “Distributed Speech Recognition / Speaker Verification System”, [online], World Wide Web Consortium, [Search February 1, 2005], Internet <URL: http: //www.e.tokushima -u.ac.jp/News/memo/eng-fes/22-ren.pdf> JP 2001-268646 A

By the way, when voice authentication (speaker verification) is implemented as a service, interactive exchanges (multiple request / response steps) occur between the user and the system. Let's realize this using a normal Web server. Then, the following problems occur.
A normal Web server does not know the state of the terminal (browser) and returns a one-time response based on a one-time request from the terminal (stateless). Here, in the interactive exchange of speaker verification, if a single request from the terminal does not include all information necessary for processing of the speaker verification server, the Web server manages the state of each terminal. Need arises. When the Web server manages the status of individual terminals, the processing load on the Web server increases dramatically, leading to a decrease in the number of simultaneous registration / authentication requests and a decrease in processing speed. Problems such as the need for an increase in scale arise.

  The present invention has been made in consideration of such circumstances, and its purpose is to provide a distributed speaker verification service that can smoothly implement a speaker verification service without performing state management for each service request in the distributed speaker verification method. It is to provide a speaker verification device and a distributed speaker verification terminal device.

  Another object of the present invention is to provide a computer program for realizing the distributed speaker verification apparatus of the present invention using a computer.

  Another object of the present invention is to provide a computer program for realizing the distributed speaker verification terminal device of the present invention using a computer.

In order to solve the above problems, a distributed speaker verification apparatus according to the present invention is a speaker verification apparatus of a distributed speaker verification method, and has a speech feature corresponding to each of a plurality of utterance patterns of a registered user. Is stored in association with user information, and when receiving an authentication request from the terminal, receiving means for receiving user information together with the voice feature quantity of the voice input according to the utterance pattern , the received voice feature quantity and the user information, and the have line identity authentication process using the first speech features of registered users stored in the storage means, the authentication indicating the authentication success when determined to be the same person authentication is successful If the result is determined not to be the same person and the authentication fails, the authentication result indicating the authentication failure is obtained. If it is not possible to determine whether the person is the same person, the registered user stored in the storage means Including utterance pattern corresponding to the second audio feature, the authentication result indicating difficulties authentication, comprising an authentication processing means for outputting as the authentication result data, and transmitting means for returning the authentication result data to the terminal It is characterized by that.

  In the distributed speaker verification apparatus according to the present invention, the receiving means receives user information together with the voice feature amount of the inputted voice when a registration request is made from the terminal, and the authentication processing means receives the received voice. Using the feature amount and the user information, a user registration process for storing the voice feature amount of the user in the storage unit is performed.

A distributed speaker verification terminal device according to the present invention is a speaker verification terminal device of a distributed speaker verification system, and receives supply of utterance request data or authentication result data from the distributed speaker verification device, Voice input means for inputting voice uttered in accordance with a first utterance pattern included in the utterance request data or the authentication result data ; voice processing means for acquiring voice feature quantities of the inputted voice; and user information a user information obtaining means for obtaining, said during authentication request to a distributed speaker verification apparatus, and transmitting means for transmitting the user information together with audio feature said acquired, from the distributed speaker verification apparatus, the same An authentication result indicating successful authentication when it is determined that the person is authenticated and the authentication is successful, an authentication result indicating authentication failure when the authentication is failed when the person is determined not to be the same person, and Characterized by comprising a receiving means for receiving one of the authentication result data of the authentication result indicating a difficulty authentication including the second utterance patterns in the case where it can not be determined whether the person.

  In the distributed speaker verification terminal device according to the present invention, the transmission means transmits user information together with the acquired speech feature amount when a registration request is made to the distributed speaker verification device.

A computer program according to the present invention is a computer program for performing server processing in a distributed speaker verification method, and at the time of an authentication request from a terminal, user information together with a voice feature amount of a voice input according to a voice pattern. The first function of the registered user stored in the storage means for storing the function to receive, the received voice feature quantity and user information, and the voice feature quantity corresponding to each of the plurality of utterance patterns of the registered user in association with the user information . There line identity authentication process by using the audio feature of the authentication result indicating a successful authentication if determined to be the same person authentication is successful, determined authentication not the same person fails In the case of the authentication result indicating the authentication failure in the case, and in the case where it is not possible to determine whether or not they are the same person, the second of the registered user stored in the storage means Including utterance pattern corresponding to speech features, an authentication result indicating difficulties authentication, the features and function of outputting the authentication result data, that to realize a function for returning the authentication result data to the terminal to the computer To do.

In the computer program according to the present invention, at the time of registration request from the terminal, the storage means uses the function of receiving user information together with the voice feature amount of the input voice, and the received voice feature amount and user information. Further, the computer is further realized with a function of performing user registration processing for storing the voice feature amount of the user.
As a result, the distributed speaker verification device described above can be realized using a computer.

A computer program according to the present invention is a computer program for performing terminal processing in a distributed speaker verification method, and receives the utterance request data or authentication result data from the distributed speaker verification device, and the utterance A function of inputting voice uttered according to a first utterance pattern included in the request data or the authentication result data, a function of acquiring a voice feature amount of the input voice, a function of acquiring user information, when the authentication request to the distributed speaker verification device, a function of transmitting the user information together with audio feature said acquired, from the distributed speaker verification system, it is determined to be the same person authentication is successful The authentication result indicating the authentication success in the case where the authentication is successful and the authentication result indicating the authentication failure in the case where the identity authentication is determined to be not the same person, Characterized in that to realize the function of receiving one of the authentication result data to the computer of the second utterance pattern authentication result indicating that the difficulties authentication, including in the case can not be determined whether or not a human.

The computer program according to the present invention is characterized in that the computer further realizes a function of transmitting user information together with the acquired voice feature amount when a registration request is made to the distributed speaker verification device.
Thereby, the above-mentioned distributed speaker verification terminal device can be realized using a computer.

A computer program according to the present invention is a computer program for performing server processing in a distributed speaker verification method, and includes user information and voice of a registered user together with a voice feature amount of voice input when an authentication request is issued from a terminal. A function for receiving a feature value, a function for performing authentication processing using the received voice feature value of the input voice, user information and the voice feature value of the registered user, and returning the authentication result to the terminal It is characterized by having a computer realize the function to perform.
As a result, the distributed speaker verification device described above can be realized using a computer.

A computer program according to the present invention is a computer program for performing terminal processing in a distributed speaker verification method, and acquires a voice feature amount of voice input by voice input means and user information. A function for reading the registered user's voice feature quantity from a recording medium recording the registered user's voice feature quantity, user information together with the acquired voice feature quantity at the time of an authentication request to the distributed speaker verification device, and The computer is configured to realize a function of transmitting the read registered user's voice feature value and a function of receiving an authentication result from the distributed speaker verification device.
Thereby, the above-mentioned distributed speaker verification terminal device can be realized using a computer.

  According to the present invention, the distributed speaker verification terminal device transmits user information to the distributed speaker verification device together with the voice feature amount of the input voice at the time of registration request and authentication request of the speaker verification service. The distributed speaker verification terminal device can obtain user information necessary for service execution for each request. This makes it possible to smoothly implement a speaker verification service without performing state management for each service request in the distributed speaker verification method.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of a distributed speaker verification system according to the first embodiment of the present invention. In FIG. 1, the distributed speaker verification system includes a user terminal 1 and a speaker verification server 2. The Web server 3 provides a general Web application. For example, the Web server 3 provides an HTML document 32 to the user terminal 1. These user terminal 1, speaker verification server 2, and Web server 3 transmit and receive data via a communication network such as the Internet, but the communication data is encrypted.

  A user terminal 1 illustrated in FIG. 1 includes a microphone 11, a voice processing unit 12, a communication unit 13, a user information acquisition unit 14, and a login control unit 15. The microphone 11 is for inputting the voice uttered by the user. The voice processing unit acquires a voice feature amount from the input voice. The user information acquisition unit 14 acquires various user information including a user ID (identifier). The communication unit 13 transmits / receives data to / from the communication units 21 or 31 of the speaker verification server 2 or the Web server 3 via the communication network. The login control unit 15 controls the state of the user terminal based on information obtained through the communication unit 13.

The speaker verification server 2 includes a Web server function unit 20 and a speaker verification function unit 23. The Web server function unit 20 includes a communication unit 21 and a CGI (Common Gateway Interface) program unit 22. The speaker verification function unit 23 includes an authentication processing unit 24 and a storage unit 25.
The communication unit 21 transmits / receives data to / from the communication unit 13 of the user terminal 1 via the communication network. The CGI program unit 22 implements the CGI function by executing the CGI program. Based on the request from the user terminal 1 received from the communication unit 21, the CGI program unit 22 requests the user registration process or authentication process from the authentication processing unit 24. The CGI program unit 22 embeds the result of the registration process or authentication process received from the authentication processing unit 24 in the HTML document and returns a response to the communication unit 21. The authentication processing unit 24 performs user registration processing and authentication processing. The storage unit 25 stores the voice feature amount of the registered user. The voice features of the registered user are stored in a database.

  In the distributed speaker verification system of FIG. 1, the user registers his / her voice feature quantity in the speaker verification server 2 in advance. At the time of registration, the user terminal 1 sends a speaker verification service registration request to the speaker verification server 2. In response to this registration request, the speaker verification server 2 executes user registration processing. Thereby, the voice feature amount of the user is stored in the storage unit 25 of the speaker verification server 2.

  For example, when the user requests the Web application of the Web server 3 using the user terminal 1, the user terminal 1 makes an authentication request for the speaker verification service to the speaker verification server 2. In response to this authentication request, the speaker verification server 2 executes a user authentication process based on the registered user's voice feature value stored in the storage unit 25. As a result, the authentication result is returned from the speaker verification server 2 to the user terminal 1. The user terminal 1 controls login access to the Web server 3 based on the authentication result.

  The speaker verification server 2 receives and executes a registration request and an authentication request from the user terminal 1 by a CGI program.

The user terminal 1 transmits the user information acquired by the user information acquisition unit 14 to the speaker verification server 2 together with the voice feature amount of the voice inputted by the microphone at the time of the registration request and authentication request of the speaker verification service. The user information includes the following.
User ID: This is sent when registration is requested and when authentication is requested.
・ Voice pattern at the time of user registration (text data); sent when registration is requested.
-Number of utterances during user registration; sent when registration is requested.
・ Speech pattern (text data) at the time of user authentication; sent when authentication is requested.
Speaker verification criteria parameters such as judgment threshold at user authentication, gray zone width, etc .; sent when authentication is requested.
Web application name that uses speaker verification (information for returning to the Web application); sent when requesting authentication.
The above user information can be acquired in advance from the Web server 3, for example.

  Next, with reference to FIG. 2 and FIG. 3, the operation | movement which concerns on the distributed speaker verification system of this embodiment is demonstrated. FIG. 2 is an operation sequence diagram during user registration according to the distributed speaker verification system of the present embodiment. FIG. 3 is an operation sequence diagram during user authentication according to the distributed speaker verification system of the present embodiment. 2 and 3 exemplify data (a request to the speaker verification server and a response from the speaker verification server) transmitted and received between the user terminal 1 and the speaker verification server 2. Each data is given a code corresponding to the transmitted / received step.

First, the operation at the time of user registration will be described with reference to FIG.
In FIG. 2, the user terminal 1 displays a message for prompting the user to speak the user ID (S1). When the user utters “012345 (example)” according to this display (S2), the user terminal 1 transmits the voice feature amount and user information of the voice input to the microphone to the speaker verification server 2 (S3).

  Next, when the speaker verification server 2 receives the voice feature value and the user information from the user terminal 1, the speaker verification server 2 performs voice recognition using the received voice feature value, and if the recognition result is obtained normally, the recognition result The data “012345 (example)” is returned to the user terminal 1 (S4a). On the other hand, if the recognition result is not obtained normally, data without the recognition result is returned to the user terminal 1 (S4b). In addition, the user terminal 1 will repeat the process of said step S1, if the data without this recognition result are received.

  Next, when receiving the data of the recognition result “012345 (example)”, the user terminal 1 displays a message asking whether the received recognition result “012345 (example)” is a correct user ID (S5). When the user utters “Yes” to the effect of the display (S 6), the user terminal 1 sends the voice feature quantity and user information of the voice input to the speaker verification server 2 to the speaker verification server 2. Transmit (S7).

  Next, when the speaker verification server 2 receives the voice feature value and the user information from the user terminal 1, the speaker verification server 2 performs voice recognition using the received voice feature value, the recognition result is obtained normally, and the recognition result is “ If the voice is “yes” indicating “correct”, the data of the recognition result “yes” is returned to the user terminal 1 (S8a). On the other hand, if the recognition result is not obtained normally or if the recognition result is obtained normally but the voice is “No” indicating that the recognition result is “false”, the recognition result is not found or the recognition result is not obtained. “No” data is returned to the user terminal 1 (S8b). Note that the user terminal 1 repeatedly performs the process of step S <b> 1 when receiving the data with no recognition result or with the recognition result “No”.

  Next, when the data of the recognition result “Yes” is received, the user terminal 1 displays a message prompting to utter the utterance pattern “987654 (example)” at the time of user registration in the received data (S9). When the user utters “987654 (example)” according to this display (S10), the user terminal 1 acquires the voice feature amount of the voice input by the microphone and temporarily stores it. The user terminal 1 repeats the processes of S9 and S10 as many times as the number of utterances at the time of user registration in the user information. Thereby, in the example of FIG. 2, the voice feature amount for five times is acquired and temporarily stored. Next, the user terminal 1 transmits the voice feature amount and user information for five times to the speaker verification server 2 (S11).

  Next, when the speaker verification server 2 receives the voice feature value and the user information from the user terminal 1, the speaker verification server 2 determines whether or not the registration criterion is satisfied based on the received five voice feature values. As a result of this determination, if the registration is successful, the voice feature amount is stored in the storage unit 25 in association with the user ID. Further, registration success data is returned to the user terminal 1 (S12). On the other hand, in the case of registration failure, registration failure data is returned to the user terminal 1 (S12).

  Next, when receiving the registration result data, the user terminal 1 displays a message indicating registration success when the registration is successful (S13), while displaying a message indicating registration failure when the registration is failed. It is displayed (S14).

Next, the operation at the time of user authentication will be described with reference to FIG.
In FIG. 3, the user terminal 1 displays a message for prompting the user to utter a user ID (S21). When the user utters “012345 (example)” in accordance with this display (S22), the user terminal 1 transmits the voice feature amount of the voice input by the microphone and the user information to the speaker verification server 2 (S23).

  Next, when the speaker verification server 2 receives the voice feature value and the user information from the user terminal 1, the speaker verification server 2 performs voice recognition using the received voice feature value, and if the recognition result is obtained normally, the recognition result The data “012345 (example)” is returned to the user terminal 1 (S24a). On the other hand, if the recognition result is not obtained normally, data without the recognition result is returned to the user terminal 1 (S24b). In addition, the user terminal 1 will repeat the process of said step S21, if the data without this recognition result are received.

  Next, when the user terminal 1 receives the data of the recognition result “012345 (example)”, the user terminal 1 compares the received recognition result “012345 (example)” with the user ID in the user information. A message indicating a valid user ID is displayed on the screen (S5a). On the other hand, if they match, a message prompting to utter the utterance pattern “9854 (example)” at the time of user authentication in the received data is displayed on the screen (S25b). When the user utters “9854 (example)” according to this display (S26), the user terminal 1 transmits the voice feature amount and user information of the voice input to the microphone to the speaker verification server 2 (S27).

  Next, when the speaker verification server 2 receives the voice feature value and the user information from the user terminal 1, the speaker verification server 2 acquires the corresponding voice feature value from the storage unit 25 based on the user ID in the received user information. And based on the audio | voice feature-value acquired from this memory | storage part 25, and the received audio | voice feature-value, it is determined whether both speakers are the same person. As this determination condition, a speaker verification determination condition parameter in the received user information is used. As a result of the determination, if it can be determined that they are the same person and the personal authentication is successful, the authentication success data is returned to the user terminal 1 (S28a). On the other hand, if it can be determined that they are not the same person and the personal authentication fails, authentication failure data is returned to the user terminal 1 (S28b). On the other hand, if it is not possible to determine whether or not they are the same person, data that is difficult to authenticate is returned to the user terminal 1 (S28c). At this time, the utterance pattern “7654 (example)” at the time of user authentication different from the previous time is included in the data that is difficult to authenticate.

Next, when receiving the authentication result data, the user terminal 1 displays a message indicating the authentication success if the authentication is successful, and displays a message indicating the authentication failure if the authentication is failed. On the other hand, if authentication is difficult,
Again, a message prompting to utter the utterance pattern “7654 (example)” at the time of user authentication in the received data is displayed on the screen (S25b). In this case, the processes from step S25b to S28a, b, c are repeated.

  If the authentication is successful as a result of the authentication, the user terminal 1 executes login processing to the corresponding Web server 3 based on the Web application name in the user information. Thereby, the user can receive the Web application by the Web server 3 using the user terminal 1.

  As described above, according to the present embodiment, the user terminal transmits user information to the speaker verification server 2 together with the voice feature amount of the input voice at the time of requesting registration and authentication of the speaker verification service. The user verification server 2 can obtain user information necessary for service execution for each request. Therefore, the speaker verification server 2 does not need to perform state management for each service request from the user. This makes it possible to smoothly implement a speaker verification service without performing state management for each service request in the distributed speaker verification method.

Next, a second embodiment will be described.
FIG. 4 is a block diagram showing a configuration of a distributed speaker verification system according to the second embodiment of the present invention. In FIG. 4, parts corresponding to those in FIG. 1 are given the same reference numerals, and explanation thereof is omitted.
In 2nd Embodiment shown by FIG. 4, the memory | storage part 41 which memorize | stores the audio | voice feature-value of a registered user is provided in the user terminal 1 side. In the embodiment of FIG. 4, the storage unit 25 of FIG. 1 is deleted from the speaker verification server 2, but the storage unit that stores the voice feature amount of the registered user in the speaker verification server 2 as in FIG. 1. 25 may be provided.

  In FIG. 4, the storage unit 41 includes a recording medium that stores a registered user's voice feature amount. This recording medium is difficult to illegally rewrite or copy recorded data. More preferably, unauthorized rewriting or copying is virtually impossible. For example, an IC card having a built-in IC chip memory that has a unique ID and can read data without contact can be used. In the storage unit 41, the voice feature amount of the user is written in advance using another voice registration device.

  The user information acquisition unit 14 of the user terminal 1 has a function of reading the registered user's voice feature amount from the storage unit 41. The connection between the storage units 41 may be a contact type such as a cable connection or a non-contact type such as a wireless connection.

  At the time of user authentication, the user terminal 1 reads the voice feature amount of the registered user from the storage unit 41 and transmits it to the speaker verification server 2 together with the voice feature amount of the voice inputted by the microphone and the user information. Then, the speaker verification server 2 authenticates the speaker by using the registered user's voice feature value, the voice feature value of the voice input from the microphone, and user information received from the user terminal 1. This authentication process is the same as the process shown in FIG. 3 except that the voice feature amount of the registered user is provided from the user terminal 1 to the speaker verification server 2.

  According to the second embodiment, the user can manage the voice feature amount, which is important personal information of the user, and can contribute to the protection of the personal information. Moreover, since an arbitrary terminal can be used by using a portable recording medium (for example, IC card) as the storage unit 41, the convenience of the user is not impaired.

The following effects are obtained by the distributed speaker verification method according to each of the embodiments described above.
(1) Conventionally, in order for a user to use an arbitrary terminal, it has been necessary to register the user's voice with respect to each terminal, but this is not necessary and the convenience of the user is improved.
(2) Since the speaker verification server performs registration processing and authentication processing with a large amount of calculation, the waiting time until the processing is completed is shortened, and it becomes possible to cope with a Web application having a high real-time property.

Further, the program for realizing each step shown in FIG. 2 or FIG. 3 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into a computer system and executed, thereby being distributed. Server processing or terminal processing in the type speaker verification method may be performed. Here, the “computer system” may include an OS and hardware such as peripheral devices.
Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” means a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, etc. This is a storage device.

Further, the “computer-readable recording medium” means a volatile memory (for example, DRAM (Dynamic DRAM) in a computer system that becomes a server or a client when a program is transmitted through a network such as the Internet or a communication line such as a telephone line. Random Access Memory)), etc., which hold programs for a certain period of time.
The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design changes and the like within a scope not departing from the gist of the present invention.
For example, in the above-described embodiment, a speaker verification server is provided separately from a Web server that provides a general Web application. However, both servers may be integrated.

1 is a block diagram showing a configuration of a distributed speaker verification system according to a first embodiment of the present invention. It is an operation | movement sequence diagram at the time of user registration which concerns on the distributed speaker verification system of the embodiment. It is an operation | movement sequence diagram at the time of the user authentication which concerns on the distributed speaker verification system of the embodiment. It is a block diagram which shows the structure of the distributed speaker verification system which concerns on the 2nd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... User terminal, 2 ... Speaker collation server, 11 ... Microphone, 12 ... Voice processing part, 13, 21 ... Communication part, 14 ... User information acquisition part, 15 ... Login control part, 20 ... Web server function part, 22 ... CGI program part, 23 ... Speaker verification function part, 24 ... Authentication processing part, 25, 41 ... Storage part

Claims (8)

A speaker verification device using a distributed speaker verification method,
Storage means for storing voice feature amounts corresponding to each of a plurality of utterance patterns of a registered user in association with user information ;
Receiving means for receiving user information together with the voice feature amount of the voice inputted according to the utterance pattern at the time of authentication request from the terminal;
The received speech features and user information, and have rows authentication process using the first speech features of registered users stored in the storage means, determines that the same person authentication is successful The authentication result indicating that the authentication is successful, the authentication result indicating that the authentication is unsuccessful if the identity authentication is unsuccessful. An authentication processing means for outputting an authentication result indicating authentication difficulty as authentication result data, including an utterance pattern corresponding to the second voice feature amount of the registered user stored in the means;
Transmitting means for returning the authentication result data to the terminal;
A distributed speaker verification apparatus comprising: The receiving means receives user information together with a voice feature amount of the inputted voice at the time of registration request from the terminal;
The distribution according to claim 1, wherein the authentication processing unit performs user registration processing for storing the voice feature amount of the user in the storage unit, using the received voice feature amount and user information. Type speaker verification device. A speaker verification terminal device using a distributed speaker verification method,
Voice input means for receiving the utterance request data or the authentication result data from the distributed speaker verification device and inputting the uttered voice according to the first utterance pattern included in the utterance request data or the authentication result data When,
Voice processing means for acquiring voice feature values of the inputted voice;
User information acquisition means for acquiring user information;
When the authentication request to the distributed speaker verification apparatus, and transmitting means for transmitting the user information together with audio feature said acquired,
From the distributed speaker verification device , an authentication result indicating successful authentication when it is determined that they are the same person and successful authentication, and an authentication failure when it is determined that they are not the same person and authentication is failed Receiving means for receiving any authentication result data among the authentication result and an authentication result indicating authentication difficulty including the second utterance pattern when it is not possible to determine whether or not the same person ;
A distributed speaker verification terminal device characterized by comprising:   4. The distributed speaker verification terminal device according to claim 3, wherein the transmission unit transmits user information together with the acquired speech feature amount when a registration request is made to the distributed speaker verification device. A computer program for performing server processing in a distributed speaker verification method,
A function of receiving user information together with a voice feature amount of voice inputted according to the utterance pattern at the time of an authentication request from the terminal;
The received voice feature quantity and user information, and the first voice feature quantity of the registered user stored in storage means for storing the voice feature quantity corresponding to each of the plurality of utterance patterns of the registered user in association with the user information. There line identity authentication process using the authentication result indicating that the authentication success in the case where it is determined to be the same person authentication is successful, in the case where it is determined that it is not the same person authentication has failed authentication failure If it is not possible to determine whether or not the person is the same person, the authentication result including the utterance pattern corresponding to the second voice feature amount of the registered user stored in the storage means is shown. A function to output the result as authentication result data ;
A function of returning the authentication result data to the terminal;
A computer program for causing a computer to realize the above. A function of receiving user information together with a voice feature amount of input voice at the time of a registration request from a terminal;
A function of performing user registration processing for storing the voice feature amount of the user in the storage unit using the received voice feature amount and user information;
The computer program according to claim 5 , further comprising: A computer program for performing terminal processing in a distributed speaker verification method,
A function of receiving the utterance request data or the authentication result data from the distributed speaker verification device and inputting the uttered voice according to the first utterance pattern included in the utterance request data or the authentication result data;
The function of acquiring the speech features of the voice that is the input,
A function to acquire user information;
When the authentication request to the distributed speaker verification device, a function of transmitting the user information together with audio feature said acquired,
From the distributed speaker verification device , an authentication result indicating successful authentication when it is determined that they are the same person and successful authentication, and an authentication failure when it is determined that they are not the same person and authentication is failed A function of receiving authentication result data of any one of the authentication result and the authentication result indicating the difficulty of authentication including the second utterance pattern when it is not possible to determine whether or not the person is the same person ;
A computer program for causing a computer to realize the above. 8. The computer program according to claim 7 , further causing a computer to realize a function of transmitting user information together with the acquired voice feature amount when a registration request is made to the distributed speaker verification apparatus.

Images