JPH09127976A - Speaker recognition system and speaker recognition method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To cope with the case that a normal speaker recognition means cannot be used due to a cold of a speaker by collating the similarity between the feature of the inputted voice of the speaker and the voice feature corresponding to the specific information within the voice feature of the speaker stored in a speaker recognition information memory means. SOLUTION: A user inputs the specific information such as voice from a specific information input means 2. This voice is converted into the feature quantity and fed to a register section 6. The register section 6 registers the standard pattern of the voice of the user in a speaker recognition information memory section 5 correspondingly to the specific information inputted from the specific information input means 2. When a normal speaker recognition means cannot be utilized, the user inputs the prescribed specific information from the specific information input means 2. This information is converted into the feature quantity and fed to a speaker recognition section 7. The speaker recognition section 7 extracts the registered standard pattern from the speaker recognition information memory section 5 and collates this standard pattern with the feature pattern from a feature extraction section 4.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a speaker identification system and a speaker identification method for speaker identification.

[0002]

2. Description of the Related Art Conventionally, in a bank or the like, a user is required to input a personal identification number or the like in order to confirm his / her identity. Further, the computer confirms the identity of the user by allowing the user to enter a personal identification code string similar to a personal identification number, called a password. However, such confirmation by inputting the personal identification number or personal identification character string can be stolen without difficulty as long as another person knows the personal identification number or personal identification character string. Moreover, the PIN and PIN are the person who registered them (the person).
Many of them use the date of birth and anniversary, or phone number, spelling of name, etc., so it is not difficult for others to discover it.

In order to avoid such drawbacks of the personal identification number and personal identification character string, attention has recently been paid to speaker identification using speaker recognition, which determines whether or not the person is the person himself / herself by voice. Speaker identification using this speaker recognition is performed by checking whether or not the characteristic pattern of the voice uttered by a speaker matches the pre-registered standard voice pattern of this speaker. It is to determine whether or not. That is, the similarity between the feature amount (feature pattern) extracted from the speaker's voice and the voice standard pattern of this speaker is calculated, and it is determined whether or not the person is the person based on the degree of similarity. Since it utilizes the physical characteristics, it is difficult for other people to imitate the voice as compared with the personal identification number or the personal identification character string, and thus it is possible to more effectively prevent the theft of another person.

[0004]

However, in the conventional speaker identification system as described above, when the voice of the user suddenly changes due to a cold or the like, the speaker identification using the speaker recognition can be performed. There was a problem that I could not do it.

The present invention provides a speaker identification system and a speaker identification method capable of coping with such a situation even when the voice is suddenly changed due to a cold or the like and the speaker recognition cannot be used at all. It is intended to be provided.

[0006]

In order to achieve the above object, the invention according to claim 1 is a speaker identification information storage means for storing speaker identification information, and a user for specifying a user. Specific information input means for inputting specific information, voice input means for inputting the voice of the speaker, characteristics of the voice of the speaker input from the voice input means, and the speaker identification information storage Collating means for collating whether or not the speaker voice feature stored in the means is similar to the voice feature corresponding to the identifying information input from the identifying information inputting means; As a result, when it is determined that they are not similar to each other, a confirmation means for confirming the user is provided.

According to a second aspect of the present invention, in the speaker identification system according to the first aspect, the confirmation means stores associated information in association with identification information for identifying a user. An information storage unit is provided, and when it is determined that they are not similar as a result of the collation, the information storage unit is stored in the associated information storage unit in association with the identification information input from the identification information input unit. According to the accompanying information
It is characterized in that a notification for confirmation is given to the user, and whether the user is an authorized user or not is confirmed.

Further, the invention according to claim 3 is the speaker identification system according to claim 2, wherein the accompanying information storage means further includes, as additional information, a second identification information for identifying the user. Information is stored, and when the confirmation means determines that they are not similar as a result of the collation, the confirmation means instructs the user to input the second identification information, When the second specifying information is input, the confirming unit associates the second specifying information input by the user with the specifying information input from the specifying information input unit. It is characterized in that it is checked whether or not the user is an authorized user by collating with the second specifying information stored in the information storage means.

The invention according to claim 4 is the speaker identification system according to claim 1 or 2, wherein the confirmation means determines that the user is not similar as a result of the collation. On the other hand, a plurality of dummy identification information items including the correct second identification information item are presented, and the user is allowed to select one of them.

According to the invention of claim 5, in the speaker identification system according to claim 2, the telephone number of the user is stored as incidental information in the incidental information storage means, and the confirmation means is The user is called according to the telephone number.

According to a sixth aspect of the present invention, in the speaker identification system according to the fifth aspect, when the confirmation means calls the user by telephone, if the user is in a telephone call, the user can properly use the telephone. The feature is that it is judged as a person.

According to the invention of claim 7, in the speaker identification system according to claim 1 or claim 2, when it is not confirmed that the user is an authorized user as a result of the confirmation, the present invention is obtained. It is characterized in that a voice storage means for storing the voice of the user in a reproducible manner is further provided.

Further, according to the invention of claim 8, in the speaker identification system according to claim 1 or claim 2, when it is not confirmed that the user is an authorized user as a result of the confirmation, the present invention is obtained. It is characterized in that a video storage means for storing the video of the user in a reproducible manner is further provided.

According to the invention of claim 9, when the specifying information for specifying the user is input and the voice of the speaker is input, the feature of the input voice of the speaker is input. A check is performed to determine whether or not the pre-stored voice feature is similar to the identification information, and when the result of the comparison indicates that they are not similar, a confirmation is given to the user. Is characterized by.

The invention described in claim 10 is the same as the ninth invention.
In the speaker identification method described, when it is determined that they are not similar as a result of the collation, the user is instructed to input the second specifying information, and the user inputs the second specifying information. At this time, the second identification information input by the user is collated with the second identification information stored in advance corresponding to the identification information to determine whether the user is an authorized user. It is characterized by taking the confirmation of.

The invention according to claim 11 is the invention according to claim 9.
In the speaker identification system described above, when it is determined that they are not similar as a result of the collation, a plurality of dummy specifying information including the correct second specifying information is presented to the user to use. It is characterized by letting the person choose one of them.

[0017]

1 is a diagram showing a configuration example of a speaker identification system according to the present invention. Referring to FIG. 1, the speaker identification system is for confirming the identity of a person in a bank or the like by speaker recognition, and is a voice input unit (for example, a microphone) for inputting a voice of a user. 1, a specific information input unit (for example, a keyboard) 2 for inputting specific information for specifying a user, and only a voice part of the speaker from the signal input from the voice input unit 1. A voice section detector 3 for detecting a voice section, a feature extractor 4 for extracting a feature amount (feature pattern) from a voice signal in the voice section detected by the voice section detector 3, and a speaker recognition prior to the recognition. A registration unit 6 that pre-registers a standard feature amount (feature pattern) of a speaker's voice as a standard pattern in the speaker recognition information storage unit 5, and a user.
A speaker recognizing unit 7 that matches the feature amount (feature pattern) of the voice of (speaker) with the standard pattern registered in the speaker recognition information storage unit 5 and performs speaker recognition based on the degree of similarity. When,
It has a switching unit (for example, a switch) 8 for switching between a registration mode for registering a standard pattern and a recognition mode for speaker recognition.

Here, the feature extraction unit 4 may convert the voice signal into a spectrum as a feature amount (feature pattern) or an LPC cepstrum, and the type of the feature amount is not particularly limited. Not something to do. It should be noted that in order to convert to a spectrum, F to conversion of feature quantity
FT is preferably used, and LPC analysis or the like is preferably used for conversion into LPC cepstrum.

When the standard pattern is registered (in the registration mode), the registration unit 6 uses the feature quantity (feature pattern) extracted by the feature extraction unit 4 based on the voice uttered by a speaker as the standard pattern. When registering in the speaker recognition information storage unit 5, as shown in FIG. 2, the identification information input by the speaker from the identification information input means 2 (for example, the name or birth date of this speaker, Alternatively, the standard pattern can be registered in the speaker recognition information storage unit 5 in association with the speaker's personal identification number. In other words, the speaker recognition information storage unit 5 is registered with the speaker recognition information necessary for speaker identification, and the speaker recognition information storage unit 5 is also stored in the speaker recognition information storage unit 5. , Multiple speakers (eg users A, B,
C, D, ...) Speaker recognition information can be registered.

As a standard pattern of voices registered in the speaker recognition information storage unit 5, each user (speaker) is made to speak a word in advance according to the usage pattern of the speaker identification system. Alternatively, each user may freely utter a desired word.

Further, the speaker recognition unit 7 extracts a standard pattern corresponding to the current speaker from the standard patterns of a plurality of speakers registered in the speaker recognition information storage unit 5, and extracts the standard pattern as the standard pattern. Match the current speaker's characteristic pattern,
It is possible to use a speaker verification system that determines whether or not the current speaker is the regular speaker based on whether the degree of similarity is higher or lower than a predetermined reference value (threshold value).

When the speaker recognition unit 7 performs speaker recognition by the speaker verification method, the user (speaker) inputs from the specific information input means 2 in the registration mode at the time of speaker recognition. It is necessary to enter the same identifying information as the identifying information that was specified. As a result, the speaker recognition unit 7 can extract the standard pattern corresponding to the current speaker from the standard patterns of the plurality of speakers registered in the speaker recognition information storage unit 5, and the standard pattern can be extracted. It is possible to match the pattern with the characteristic pattern of the voice of the current speaker.

Further, the speaker recognizing unit 7 determines that the standard pattern of the voice registered in the speaker recognizing information storage unit 5 is for each user.
When the (speaker) has spoken a word in advance, the corresponding recognition can be performed, and the standard pattern of the voice registered in the speaker recognition information storage unit 5 can be used. Is a voice in which a desired word is freely uttered for each user, recognition corresponding to this can be performed. However, when each user (speaker) utters a predetermined word to perform speaker recognition, a similar criterion (threshold) can be set to a constant value for each speaker. However, when a desired word is uttered for each user to perform speaker recognition, a similar determination standard (threshold value) can be made different for each speaker.

When a user (for example, D) uses the speaker identification system having such a configuration for the first time, this user (speaker)
First, D needs to register his own voice as a standard pattern. Therefore, the user D operates the switching unit (for example, the switch) 8 to connect the feature extraction unit 4 to the registration unit 6 and set the registration mode.

Next, the user (speaker) D inputs predetermined identification information, for example (user D), from the identification information input means 2. Further, at this time, the user utters a predetermined specific word. This voice is input from the voice input means 1, and the voice section detection unit 3 and the feature extraction unit 4 input the feature amount.
(Feature pattern) and is given to the registration unit 6 as a standard pattern of the voice of the speaker.

As a result, the registration unit 6 associates the standard pattern of the voice of the user (speaker) D with the identification information input from the identification information input means 2, and the speaker recognition information storage unit. Register to 5. For example, in the past, a plurality of users (different users) A, B, and C have registered their own voices as standard patterns in the speaker recognition information storage unit 5, and the current user D is as described above. When the user's own voice is registered as a standard pattern, the standard pattern is stored (registered) in the speaker recognition information storage unit 5 as shown in FIG.

In this way, when the standard pattern of this voice is stored in the speaker recognition information storage unit 5, the user D
With this speaker identification system, the speaker recognition for the user D can be performed. That is, this user D can use this system to determine whether or not the user who is currently using is the user D himself / herself.

Specifically, when the user D subsequently uses this system, the user D operates the switching unit 8 to connect the feature extracting unit 4 to the speaker recognizing unit 7, and to use this system. Set to recognition mode.

Next, the user D inputs predetermined identification information, for example (user D), from the identification information input means 2. Further, at this time, the user D utters a predetermined specific word. This voice is input from the voice input means 1, and the voice section detection unit 3 and the feature extraction unit 4 input the feature amount.
It is converted into (feature pattern) and given to the speaker recognition unit 7.

As a result, the speaker recognition section 7 uses the standard pattern registered in association with the identification information (user D) input from the identification information input means 2 as the speaker recognition information storage section 5. From the standard pattern and feature extraction unit 4
And the similarity is calculated, and it is determined whether the similarity is higher or lower than a predetermined reference value. As a result, when it is determined that the similarity is low,
It is determined that the user is not the authorized speaker himself D, and the use by this user is rejected. On the other hand, when it is determined that the degree of similarity is high, it is determined that the user is the regular speaker himself D, and the use is permitted by the user. That is, the user is permitted to use the application (for example, processing such as deposit / withdrawal and balance inquiry).

By the way, in such a speaker identification system, as described above, when the voice of the user (speaker) suddenly changes due to a cold or the like, the voice of the user is not the true voice of the user. Therefore, the speaker identification cannot be performed.

In order to solve such an inconvenience, in the speaker identification system of FIG. 1, the characteristics of the speaker's voice input from the voice input means 1 and the speaker identification information storage unit 5 are stored. Information input means 2 for identification of the voice characteristics of the speaker
As a result of collating whether or not the voice feature corresponding to the identification information input from the user is similar, when it is determined that they are not similar, a confirmation means for further confirming with the user 11 is provided, and when the confirmation means 11 confirms that the user is the regular speaker himself, it is determined that the speakers are not similar as a result of the speaker verification in the speaker recognition unit 7. However, the current user is identified as the legitimate speaker himself.

FIG. 3 is a diagram showing an example of the structure of the confirmation means 11. In the example of FIG. 3, the confirmation unit 11 includes the incidental information storage unit 12 in which the incidental information is stored in association with the identification information for identifying the user, and the result of speaker verification in the speaker recognition unit 7, When it is determined that they are not similar, the user is notified for confirmation according to the accompanying information stored in the accompanying information storage unit 12 in association with the specifying information input from the specifying information input unit 2. The notifying unit 13 to perform, and when the notifying unit 13 notifies the user and a response for confirmation is given from the user, based on the response, it is determined whether or not the user is the proper speaker himself. It has a discriminating unit 14 for discriminating.

FIG. 4 is a diagram showing an example of the configuration of the incidental information storage unit 12. In the example of FIG. 4, the incidental information storage unit 12 stores, as incidental information, a method of notifying the user. It has become. For example, the telephone number of the authorized user, or the fact that guidance or the like is displayed on the display device of the speaker identification system is stored.

Further, in the example of FIG. 4, the accompanying information storage unit 1
The second identification information is associated with the identification information input from the identification information input unit 2 and stores second identification information for identifying the user as accompanying information.

Here, as the second specifying information, information different from the specifying information can be used. In the following, the identifying information will be referred to as the first identifying information in order to distinguish it from the second identifying information. In this case, as the first identification information, as described above, for example, the name of the user, the date of birth, or the personal identification number (hereinafter, referred to as the first personal identification number) can be used. , Second
As the identifying information of, for example, a second personal identification number different from the first personal identification number can be used.

It should be noted that such various accompanying information can be input together with the specifying information from the specifying information input means 2 when the user D newly registers the standard pattern of his or her voice. Accordingly, the accompanying information of the user D is registered in the accompanying information storage unit 12 in association with the identification information of the user D.

Further, as the notifying section 13, various types can be used according to the way of notifying the user stored as the accompanying information in the accompanying information storage section 12. For example, when the method of notifying the user is to call the user (when a telephone number is set as the accompanying information), the notification unit 13 may be a communication device (a telephone device or a telephone device). A terminal having a personal computer communication function) can be used. Further, in the case of notifying the user of this system by means of guidance or the like in this system, the notification unit 13 may be a display device, a voice synthesis output device, or the like provided in this speaker identification system. Can be used.

In this way, when a notification for confirmation is given from the notification section 13, the user inputs, for example, the second specifying information as a response to this notification,
Alternatively, it is possible to reply by voice or the like. In addition,
When the second specifying information is input, the second specifying information can be input, for example, by using the specifying information input means 2 (combined use), or alternatively, the specifying information input means 2 can be input. It can also be performed from other input means.

Further, when the second specifying information is input as the user's response to the notification from the notifying unit 13, the determining unit 14 determines the second specifying information and the specifying information input by the user. The identification information input from the information input means 2 is matched with the second identification information stored in the associated information storage section 12 to determine whether or not the user is an authorized user. Has become.

5 to 8 are views showing various usage examples of the speaker identification system of the present invention. The example of the usage pattern of FIG. 5 is the same as the configuration example of FIG. 3, except that the voice input unit 1, the specific information input unit 2, the voice section detection unit 3, the feature extraction unit 4, the speaker recognition information storage unit 5, and the registration unit 6 are included. The speaker recognition unit 7, the switching unit 8, the accompanying information storage unit 12, the notification unit 13, the determination unit 14, and the display device 16 are installed, for example, as a speaker recognition device unit 30 at a bank counter or the like. It has become a thing.

In the usage example of FIG. 5, in order to perform new registration, change or update of the standard pattern, and speaker recognition,
The user goes to the speaker recognition device unit 30 installed at, for example, a window of a bank, and uses the speaker recognition device unit 30 to perform a new registration operation of the standard pattern, a speaker recognition operation, and a change of the standard pattern. Alternatively, the update operation can be performed as described above. In addition,
When the speaker recognition device unit 30 has a standard pattern automatic updating function, the user automatically updates the standard pattern without changing or updating the standard pattern.

Further, in the usage pattern example of FIG. 5, when the user performs a new registration operation of the standard pattern by the speaker recognition unit 30, the user also inputs and inputs the accompanying information. The accompanying information can be stored in the accompanying information storage unit 12. That is, in this case, the accompanying information storage unit 12 displays guidance as the accompanying information, for example, together with the second specifying information for each user, and as a way of notifying the user on the display device 16, for example. The effect is stored.

In this way, the standard pattern is newly registered, changed or updated, and in the recognition mode, for example, the user D uses the identification information input means 2 for the first identification to recognize the speaker. When the information is input and the voice is input from the voice input unit 1, the speaker recognition unit 7 determines the feature pattern of the voice input from the voice input unit 1 and the first identification input from the identification information input unit 2. The standard pattern corresponding to the usage information (for example, the standard pattern of the user D) is collated to determine whether or not they are similar.

As a result, when it is determined that the characteristic pattern of the input voice and the standard pattern of the user D are similar, the user is identified as the authorized user D himself and this user is identified. For example, the use of an application for user D (application for deposit / withdrawal, balance inquiry, etc.) is permitted.

On the other hand, when it is determined that the characteristic pattern of the input voice and the standard pattern of the user D are not similar, it is confirmed whether or not the user D is a regular user.
That is, the notifying unit 13 reads out the method of notifying the user from the associated information storage unit 12, and if the method of notifying is the guidance display on the display device 16, for example, the second
Guidance for allowing the user to input the identification information of the speaker, for example, "Please input the second identification information" is displayed on the display device 1 of the speaker recognition device unit 30.
Display the screen on 6 and inform the user. In response to this, for example, when the user inputs the second specifying information from the specifying information input means 2, the discriminating unit 14 causes the discriminating unit 14 to input the second specifying information and the accompanying information storage unit 12 that have just been input. The second identification information of the user D stored in is compared. As a result, when they match, the user is identified as the authorized user D himself and, for this user, for example,
Allow the use of applications for user D (applications for deposit / withdrawal, balance inquiry, etc.).

On the other hand, as a result of the collation performed by the discriminating unit 14, if they do not match, the user is prohibited from using the application for the user D, for example.

As described above, even if the voice of the regular user D suddenly changes due to a cold, for example, and the characteristic pattern of the input voice becomes dissimilar to the standard pattern of the user D, the regular voice of the regular user D becomes normal. When the user D correctly inputs the second specifying information, the user is identified as the user D, and the user D can be permitted to use the application. Further, even if another person other than the user D, for example, E learns the first specifying information of the user D,
This stranger E cannot use the application for the user D unless he / she knows the second specifying information of the user D, so that the malicious user can use the application for the authorized user. It is possible to effectively prevent such a situation that it happens.

Further, in the usage example of FIG. 6, in the usage example of FIG. 5, the user is notified via the operation center 80, for example. In this case, the notification unit 13 has an access passive unit 24 installed in the operation center 80 and an access unit 23 for accessing the access passive unit 24. That is, in the usage example of FIG. 6, in the configuration example of FIG. 3, the voice input unit 1, the specific information input unit 2, the voice section detection unit 3, the feature extraction unit 4, the speaker recognition information storage unit 5, and the registration are performed. The unit 6, the speaker recognition unit 7, the switching unit 8, the access unit 23 of the notification unit 13, and the determination unit 14 are the same as in the usage example of FIG.
For example, although the speaker recognition device unit 30 is installed at a bank counter or the like, the access passive unit 24 of the notification unit 13 is used.
Is managed by the administrator of the operation center 80 as a telephone device, for example, and when the access passive unit 24 is accessed by the access unit 23, the administrator of the operation center 80 confirms separately with the user's mobile phone or the like. Is configured to make a telephone call or the like.
Further, in order to notify the user from the operation center 80 for confirmation, the accompanying information storage unit 12 is also provided on the operation center 80 side.

In the usage pattern example of FIG. 6, in the speaker recognition device unit 30, as a result of matching between the characteristic pattern of the voice input by the user and the standard pattern of the voice of the user D, for example, these are similar. When it is determined that there is not, the access unit 23 of the speaker recognition device unit 30 calls the access passive unit 24 of the operation center 80 by, for example, a telephone, and a voice guide such as "please check with the user D". Is transmitted to the manager of the operation center 80 from the handset of the access passive unit 24. Thereby, the administrator of the operation center 80 retrieves the associated information corresponding to the user D, for example, the telephone number of the user D from the associated information storage unit 12, and contacts the user D by telephone, for example. As a result, when it is confirmed that the user D himself / herself is performing the speaker recognition, the administrator informs the access passive unit 2
From the transmitter of No. 4, a message saying, for example, "I am a user D" is uttered. Alternatively, the user is informed that he / she is a user D by a predetermined function key of the access passive unit 24, for example, “*”. As a result, the access unit 23 receives this,
Allow the user to use the application for user D.

On the other hand, when the confirmation that the user D himself / herself is performing the speaker recognition cannot be obtained, the administrator of the operation center 80 uses the transmitter of the access passive unit 24 to, for example, “use”. It is not person D ”. Alternatively, "not user D" is notified by a predetermined function key of the access passive unit 24, for example, "#". As a result, the access unit 23 receives this and prohibits the user from using the application for the user D.

As described above, also in the usage pattern example of FIG. 6, as in the usage pattern example of FIG. 5, the voice of the authorized user D suddenly changes due to, for example, a cold, and a characteristic pattern of the input voice is obtained. Even if the standard pattern of the user D is not similar to this, the regular user D correctly identifies the user as the user D by correctly inputting the second specifying information, and The person D can be permitted to use the application. Also, even if another person other than the user D, for example E, knows the first specifying information of the user D, but does not know the second specifying information of the user D,
Since the stranger E cannot use the application for the user D, it is possible to effectively prevent the situation where the malicious user uses the application for the authorized user.

In the example of usage shown in FIG. 6, the administrator of the operation center 80 can directly inquire of the user D himself.
It is possible to directly confirm whether the user is the user D himself or herself without inputting the identification information. Therefore, in this case, the determination unit 14 may not be provided. However, even in this case, the user can further input the second specifying information, and at this time, the determination unit 14 needs to be provided. In this case, the speaker recognition unit 30
The side information storage unit 12 can also be provided on the side.

In the example of the usage pattern shown in FIG. 7, the user performs an operation such as speaker recognition on the terminal in the configuration example shown in FIG. That is, in the example of FIG. 7, the voice input unit 1, the specific information input unit 2, the voice section detection unit 3, and the feature extraction unit 4 are connected to a terminal 31 (for example, a personal computer or a personal computer) installed at the user's home or office. It is realized by a telephone device, etc., and includes a switching unit 8, a speaker recognition information storage unit 5, a registration unit 6, a speaker recognition unit 7, an accompanying information storage unit 12, and a determination unit 14.
Is realized by, for example, a speaker recognition device unit 32 installed at a bank counter or the like. Also, in this case,
The notification unit 13 is realized by the access passive unit 24 provided on the terminal 31 side and the access unit 23 provided on the speaker recognition device unit 32 and accessing the access passive unit 24 of the terminal 31.

In this case, the telephone number of the access passive unit 24 for each user is previously stored in the incidental information storage unit 12 as incidental information (method of notifying the user).
In addition, the terminal 31 on the user side and the speaker recognition device unit 32 installed in a bank or the like are adapted to exchange information with each other by communication means 33, for example, a communication line (wired) or wireless. There is.

In the example of FIG. 7, only one terminal 31 is connected to the speaker recognition device unit 32 through the communication means 33, but the speaker recognition device unit 32 is not shown. Not only one but also a plurality of terminals can be connected so as to be able to transmit and receive. Further, in FIG. 7, the voice input unit 1, the specific information input unit 2, and the access passive unit 24 are configured as an integrated unit (terminal), but they may be installed as separate devices.

In the use form example of FIG. 7, in order to perform new registration, change or update of the standard pattern, and speaker recognition,
The user operates the terminal 31 installed in the user's home or company to perform a new registration operation of the standard pattern for the speaker recognition device unit 32 installed in a bank counter, for example. The speaker recognition operation and the standard pattern change or update operation can be performed in the same manner as described above. However, in the usage example of FIG. 7, a switching instruction to switch between the registration mode and the recognition mode can be given from, for example, the identification information input means 2 of the terminal, and from the identification information input means 2 of the terminal. When an instruction to enter the registration mode or the recognition mode is transmitted via the communication means 33, the speaker recognition device unit 32 side controls the switching of the switching unit 8 according to the instruction. ing. When the speaker recognizing device unit 32 has a standard pattern automatic updating function, the user automatically updates the standard pattern without changing or updating the standard pattern.

In the use mode example of FIG. 7, in the recognition mode,
In the speaker recognizing unit 7 of the speaker recognizing device unit 32, the characteristic pattern of the voice of the input user and the normal user D
As a result of the comparison with the standard pattern of No. 2, if they are not similar, the access unit 23 of the speaker recognition device unit 32 reads the accompanying information (for example, telephone number) of the user D from the accompanying information storage unit 12, Accompanying information of this user D
Call the access passive unit 24 of the user D by (telephone number), for example, play a voice guide such as "Please enter the second identification information for confirmation" and use it from the handset of the access passive unit 24. Give to person D. In response to this, the user D uses the transmitter of the access passive unit 24 to
When the second specifying information is uttered, or when the second specifying information is notified by a push tone or the like by the key operation of the access passive unit 24 and the access unit 23 receives this, the determining unit 14 The received second specifying information is collated with the second specifying information of the user D stored in the associated information storage unit 12. As a result, if they match, the user is permitted to use the application for user D.

On the other hand, when the second specifying information input by the user does not match the second specifying information of the user D stored in the accompanying information storage unit 12, this user is identified. On the other hand, the use of the application for user D is prohibited.

As a result, similarly to the usage patterns shown in FIGS. 5 and 6, the voice of the regular user D suddenly changes due to a cold, for example, and the characteristic pattern of the input voice and the user D are changed.
Even if the standard pattern is not similar to this standard pattern, the normal user D correctly identifies the user as the user D by correctly inputting the second specifying information.
You can allow the use of applications for. In addition, another person other than the user D, for example E, is the user D.
Even if the first identification information of the user D is known, the stranger E cannot use the application for the user D unless the second identification information of the user D is known. It is also possible to effectively prevent a situation in which an application for a legitimate user is used by a certain other person.

In the example of the usage pattern shown in FIG. 7, the administrator of the speaker recognition device unit 30 can make a telephone call or the like for direct confirmation to the user D, and the user D can be directly inquired. At this time, even if the user D does not input the second specifying information, the user D
You can directly confirm whether you are the person or not. Therefore, in this case, the determination unit 14 may not be provided. However, even in this case, the user can further input the second specifying information, and at this time, the determination unit 14 needs to be provided. Further, in this case, the accompanying information storage unit 12 can be provided also on the speaker recognition unit 30 side.

In addition, in the usage example of FIG. 8, the access passive unit 24 is installed, for example, in the operation center 80 in the usage example of FIG. 7. This is performed in substantially the same manner as the usage example of No. 6.

Further, in the usage example of FIG. 7 (or FIG. 8), for example, as shown in FIG. 9, the voice input unit 1, the specific information input unit 2, and the access passive unit 24 are
It can also be shared as one telephone device (or personal computer communication device) 35. That is, as the telephone device (or personal computer communication device) 35, an existing one (for example, a touch-tone telephone) in the home or company of the user can be used. In this case, the handset transmission of the telephone device 35 is performed. Device is used as the voice input means 1, the handset receiver is used as the voice guide reception part in the access passive part 24, and the operation part (numeric keypad part) of the telephone device 35 is used as the specific information input part 2. be able to. In the access passive unit 24,
When the confirmation is transmitted by, for example, a voice message, the transmitter of the handset can be used as the confirmation transmitting unit of the access passive unit 24, and the second passive identification information is used in the access passive unit 24. When the call is transmitted by push tone, the operation unit (numeric keypad) of the telephone device 35 can be used as the confirmation transmission unit of the access passive unit 24.

Thus, for example, in the example of the use form of FIG. 7, the voice input means 1, the specific information input means 2, and the access passive unit 24 can be realized by one telephone device (or personal computer communication device) 35. Is possible, in this case
The user separately uses a device for recognizing a speaker (voice input means 1,
It is not necessary to prepare the specific information input means 2).

The voice input unit 1 and the access passive unit 2
When 4 is realized by one telephone device (or personal computer communication device) 35 in this way, when the user recognizes the speaker, the handset of the telephone device 35 is lifted, and the telephone device 35 becomes Therefore, when the access unit 23 accesses the access passive unit 24 to confirm the speaker, if the user is an authorized user (speaker himself), the access passive of the user destination The unit, that is, the telephone device 35 is in a call.

With this in mind, when the access unit 23 is in a call when accessing the access passive unit 24, it is determined that the user who is currently recognizing the speaker is the regular speaker himself. You can also check.

Further, in the configuration examples of FIGS. 7 and 8, the access unit 23 and the access passive unit 24 are provided, but it is also possible to realize the confirmation means 11 without providing them.

That is, it is necessary for the user to access the speaker recognition device unit installed in, for example, a bank by his / her own terminal (for example, telephone device or personal computer communication device) in order to identify the speaker. When a telephone number is entered and this telephone number is sent out as a digital signal from its own terminal, the speaker recognition device unit installed in a bank or the like uses the telephone number sent out as a digital signal from the user terminal, for example. It can also be configured to display.

In this case, the user accesses the speaker recognition device unit installed in a bank or the like, then inputs the identification information from the identification information input means 2 of the terminal, and also from the voice input means 1. At this time point, when a voice is generated and the characteristic pattern of the voice of the user input from the voice input means 1 is compared with the standard pattern of the user D and it is determined that they are not similar, Then, the operator on the speaker recognition device unit side (for example, a staff member of a bank or the like) stores in the incidental information storage unit 12 in association with the telephone number displayed as described above and the identification information input as described above. Check the phone number of a legitimate user registered in advance,
As a result, when they match, it can be confirmed that the user is an authorized user. On the other hand, when they do not match, it can be determined that the user is not an authorized user.

As described above, the confirmation can be performed without accessing the access passive unit 24 of the user from the speaker recognition device unit such as a bank.

Further, in each of the above configuration examples, when it is confirmed that the user is not the authorized user D but another person other than the user D, the history of who the other person is is also recorded. The more convenient it is. Voice feature patterns for speaker recognition (that is, speaker identification) include
Since the information of the voice of the user is included, it can be saved as a history, but the voice feature pattern is usually based on this because the data amount is compressed with respect to the original voice signal. It is difficult to judge who is.

Then, as a result of the confirmation by the confirmation means 11,
If it is confirmed that the user is not a legitimate user, the original voice of the current speaker can be reproducibly stored instead of the voice standard pattern of the current speaker.

FIG. 10 is a diagram showing a configuration example of a speaker identification system having a function of reproducibly storing the voice of the current speaker. Referring to FIG. 10, in this speaker identification system,
A voice storage unit (memory) for reproducibly storing the voice signal input from the voice input unit 1 or the voice signal after the voice section detection (voice signal within the voice section) in the recognition mode.
50 is further provided, and in the confirmation means 11,
When it is confirmed that the current speaker is the original speaker,
The voice signal stored in the voice storage means 50 is erased, for example, by control from the confirmation means 11, while the voice signal is stored in the voice storage means 50 when it is determined that the present speaker is not the regular speaker himself. The audio signal is saved as a history.

In the speaker identification system having such a configuration,
When the user utters a voice in the recognition mode, the input voice signal from the voice input means 1 is stored in the voice storage means 50. Thereafter, the confirmation means 11 confirms whether or not the present speaker is the regular speaker himself in various ways as described above, and when it is finally determined that the speaker is not the regular speaker himself, the voice memory is stored. By storing the voice signal now stored in the means 50 as a history and reproducing this voice later, it is possible to determine who impersonated the person and tried to use it.

In this configuration example, the voice signal from the voice input means 1 may be directly stored in the voice storage means 50, but in the case of saving the capacity of the voice storage means 50,
It is preferable to store the voice signal after detection of the voice section (voice signal within the voice section). Also, as a voice signal to be stored,
The amount of voice data is determined depending on whether PCM is used, ADPCM is used, or how much band is left. It is preferable that the voice storage unit 50 store the voice of the speaker with the best possible sound quality.

In the above example, when it is confirmed that the user is the proper speaker himself, the voice signal stored in the voice storage means 50 is erased in order to save the memory capacity. Even when it is confirmed that the speaker is the speaker himself, the voice signal stored in the voice storage means 50 is not erased but left as it is, for example, when the regular speaker himself uses it. You may make it overwrite this. Thus, even when the device mistakenly determines that the speaker is the proper speaker, it is possible to determine who used the speaker on behalf of the speaker based on the voice signal stored in the voice storage unit 50.

Further, in the configuration example of FIG. 10, the voice of the user is stored as the history, but the video of the user may be stored as the history. That is, as a result of the confirmation by the confirmation means 11, when it is confirmed that the user is not an authorized user, it is possible to save the image of the user.

FIG. 11 is a diagram showing an example of the configuration of a speaker identification system having a function of saving the image of the user. FIG.
With reference to, in this speaker identification system, an image pickup unit (for example, a camera) 52 for picking up an image of a user, and an A / D conversion unit 53 for A / D converting a video signal from the image pickup unit 52.
And a video storage means 54 for storing the video signal digitally converted by the A / D converter 53, and the confirmation means 11 confirms that the present speaker is the regular speaker himself. Sometimes, the video signal stored in the video storage means 54 is erased, for example, by the control of the confirmation means 11, while it is stored in the video storage means 54 when it is determined that the present speaker is not the regular speaker himself. The recorded video signal is saved as a history.

In the speaker identification system having such a configuration,
When the user performs an operation for recognition, the image pickup means 52
The video signal from is stored in the video storage means 54. After that, the confirmation means 11 confirms whether or not the present speaker is the regular speaker himself in various ways as described above, and when it is determined that the current speaker is not the regular speaker himself, the image storage means 54 stores the information. By storing the video signal that has just been stored as a history and playing this video later, it is possible to figure out who impersonated himself and tried to use it.

In the above example, when it is confirmed that the user is the legitimate speaker himself, in order to save the memory capacity,
Although it has been stated that the video signal accumulated in the video storage means 54 is deleted, even when it is confirmed that the person is a legitimate speaker,
The video signal stored in the video storage unit 54 may be left as it is without being erased, and may be overwritten on the next time the regular speaker himself uses it, for example. As a result, even when the device mistakenly determines that the speaker is the proper speaker himself, based on the video signal stored in the video storage means 54,
You can figure out who used it on behalf of the person.

In this configuration example, the image pickup means 52
May be for moving images or may be for still images. If necessary, the previous user's image can be viewed by viewing the image stored in the image storage means 54. .

By thus storing the user's voice and video in a reproducible manner, it is possible for another person to know later on. It should be noted that in the configuration examples of FIGS. 10 and 11, either one of the voice and the video is left as the history, but both the voice and the video are left as the history by combining FIG. 10 and FIG. 11. It can also be configured to.

Further, in each of the above-described configuration examples, since the voice change due to the user's cold or the like does not occur frequently, when the user inputs the second specifying information, the user specifies the second specifying information. You may not remember the identifying information exactly. In consideration of such a case, the usage corresponding to the characteristic pattern of the user's voice input from the voice input unit 1 and the identification information (first identification information) input from the identification information input unit 2 When it is determined that the standard pattern of the voice of the person (for example, user D) is not similar, the user is presented with a plurality of dummy identification information including the correct second identification information. You can also let the user choose one of them.

FIG. 12 shows an example of the configuration of a speaker identification system having a function of presenting a plurality of dummy specifying information including correct second specifying information to the user and allowing the user to select one of them. FIG. In the configuration example of FIG. 12, when the speaker recognition unit 7 compares the characteristic pattern of the input voice with the standard pattern and determines that they are not similar, the second pattern corresponding to the first specifying information is used. Of the identification information is extracted from the associated information storage unit, a plurality of dummy identification information (for example, about 20 pieces) including the extracted second identification information is generated, and these are displayed on the display device 16, for example. A dummy specifying information generator 29 is further provided.

In such a configuration, for example, the display device 16
The user is allowed to select any one of the plurality of pieces of dummy specifying information displayed in. When one piece of identifying information is selected by the user, the determining unit 14 determines the identifying information selected by the user and the second identifying information of the user D stored in the associated information storage unit 12. If they match, it is confirmed that the user is a legitimate user, and this user is permitted to use the application.
Thereby, even if the user does not remember the second identifying information correctly, the user can find the correct second identifying information from the presented plurality of identifying information and select it. it can.

If, for example, about 20 pieces of dummy identification information are displayed, the probability that another person will select the correct second identification information is only about 5%. Can effectively prevent the second selecting information from being selected. In addition, it is preferable that the plurality of pieces of dummy specifying information presented at this time have the same rule in the dummy specifying information generating section 29 and are always the same. That is, if the dummy specifying information other than the correct second specifying information is randomly generated, the specifying information other than the correct second specifying information changes every time. Information is displayed twice, and if there is identification information common to both, it is known to others that it is the correct second identification information.

As described above, in the configuration example of FIG. 12, even when the voice changes due to a cold or the like, and it is possible to correctly use the second specifying information without remembering it accurately.

In each of the above configuration examples, the accompanying information storage unit 12 is provided separately from the speaker recognition information storage unit 5. However, as shown in FIG. 13, for example, the accompanying information storage unit 12 functions as shown in FIG. The information can be stored in the speaker recognition information storage unit 5. In this case, the notification unit 13 and the determination unit 14 can read and use the notification method to the user, the second identification information, and the like from the speaker recognition information storage unit 5.

In the above configuration example, the feature extraction unit 4 is provided after the voice section detection unit 3, but instead of this, the feature extraction unit 4 is provided before the voice section detection unit 3. It may be.

Further, in the configuration examples of FIGS. 7 and 8, the voice section detection unit 3 and the feature extraction unit 4 are provided on the terminal side, but one or both of them are installed not at the terminal side but at a bank or the like. It is also possible to provide it on the side of the speaker recognition device unit that is used.

Further, in the configuration examples of FIGS. 7 and 8, the speaker recognition unit 7 is provided on the speaker recognition device unit side, but this is provided on the terminal side, not on the speaker recognition device unit side. It is also possible.

[0092]

As described above, according to the invention described in claims 1 to 11, the characteristics of the voice of the speaker input from the voice input means and the speaker identification information storage means are stored. When it is determined that the voice feature of the speaker who is not similar to the voice feature corresponding to the identification information input from the identification information input means is confirmed, the user is asked to confirm. The speaker can be identified even if the user's voice suddenly changes due to a cold or the like.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration example of a speaker identification system according to the present invention.

FIG. 2 is a diagram showing a configuration example of a speaker recognition information storage unit.

FIG. 3 is a diagram showing a configuration example of confirmation means.

FIG. 4 is a diagram showing a configuration example of an associated information storage unit.

FIG. 5 is a diagram showing an example of a usage pattern of the speaker identification system of the present invention.

FIG. 6 is a diagram showing an example of a usage pattern of the speaker identification system of the present invention.

FIG. 7 is a diagram showing an example of usage of the speaker identification system of the present invention.

FIG. 8 is a diagram showing an example of a usage pattern of the speaker identification system of the present invention.

FIG. 9 is a diagram showing an example of a usage pattern of the speaker identification system of the present invention.

FIG. 10 is a diagram showing a configuration example of a speaker identification system having a function of reproducibly storing a voice of a current speaker.

FIG. 11 is a diagram showing a configuration example of a speaker identification system having a function of saving a user's image.

FIG. 12 is a diagram showing another configuration example of the speaker identification system according to the present invention.

FIG. 13 is a diagram illustrating another configuration example of a speaker recognition information storage unit.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 voice input means 2 instruction means 3 voice section detection section 4 feature extraction section 5 speaker recognition information storage section 6 registration section 7 speaker recognition section 8 switching section 11 confirmation means 12 incidental information storage section 13 notification section 14 discrimination section 16 Display device 23 Access unit 24 Access passive unit 30 Speaker recognition device unit 31 Terminal 32 Speaker recognition device unit 33 Communication means 35 Telephone device (or personal computer communication device) 50 Voice storage means 52 Imaging means 53 A / D conversion section 54 Video Storage means 80 Operation Center

Claims (11)

[Claims] 1. A speaker identification information storage means for storing speaker identification information, an identification information input means for inputting identification information for identifying a user, and a speaker voice. A voice input means for inputting; a feature of the voice of the speaker input from the voice input means; and a voice feature of the speaker stored in the speaker identifying information storage means, from the specifying information input means Collating means for collating whether or not the inputted voice information corresponding to the identification information is similar, and for confirming to the user when it is determined that they are not similar as a result of the collation. Speaker identification system. 2. The speaker identification system according to claim 1, wherein the confirmation means includes an accompanying information storage means for storing accompanying information in association with identification information for identifying a user, When it is determined that they are not similar as a result of the collation, the user is confirmed according to the accompanying information stored in the accompanying information storage unit in association with the identifying information input from the identifying information input unit. The speaker identification system is characterized in that the user is notified and the user is confirmed whether or not the user is an authorized user. 3. The speaker identification system according to claim 2, wherein the accompanying information storage means further stores second specifying information for specifying a user as accompanying information. When the result of the collation indicates that they are not similar, the confirmation means instructs the user to input the second specifying information, and the user inputs the second specifying information. At this time, the confirmation means is stored in the associated information storage means in association with the second identification information input by the user and the identification information input from the identification information input means. A speaker identification system characterized by collating with second identification information to confirm whether or not the user is an authorized user. 4. The speaker identification system according to claim 1 or 2, wherein the confirmation means is a result of the collation,
When it is determined that they are not similar, the user is presented with a plurality of dummy identification information including the correct second identification information, and the user is allowed to select one of them. A speaker identification system characterized by being present. 5. The speaker identification system according to claim 2, wherein the attendant information storage means stores a user's telephone number as attendant information, and the confirmation means follows the user's telephone number. A speaker identification system characterized by making a telephone call. 6. The speaker identification system according to claim 5, wherein when the confirmation unit calls the user by telephone, if the user is in a call, it is determined that the user is an authorized user. A feature speaker identification system. 7. The speaker identification system according to claim 1, wherein the voice of the current user is reproducible when the result of the confirmation does not confirm that the user is an authorized user. A speaker identification system characterized in that a voice storage means for storing is further provided. 8. The speaker identification system according to claim 1 or 2, wherein if the result of the confirmation is that the user is not a legitimate user, the video of the current user can be reproduced. A speaker identification system, further comprising a video storage means for storing. 9. When the specifying information for specifying the user is input and the voice of the speaker is input, the characteristics of the input voice of the speaker and the input specific information are associated in advance. A speaker identification method characterized by performing matching whether stored voice characteristics are similar or not, and as a result of the matching, confirming to the user when it is determined that they are not similar. . 10. The speaker identification method according to claim 9, wherein when the result of the collation indicates that they are not similar, the user is instructed to input the second specifying information, and the user specifies When the second specifying information is input, the second specifying information input by the user is collated with the second specifying information stored in advance corresponding to the specifying information. , A speaker identification method characterized by confirming whether the user is an authorized user or not. 11. The speaker identification system according to claim 9, wherein when the result of the collation indicates that they are not similar to each other, a plurality of dummy identification information including correct second specifying information is presented to the user. A speaker identification method characterized by presenting identification information and allowing a user to select one of them.