JP2020205485A - Telephone system and server device - Google Patents

Abstract

To suppress special fraud.SOLUTION: A telephone system includes: a voiceprint database where a pre-registered speaker is associated with his/her voiceprint; a voiceprint authentication section; and a notification section. The voiceprint authentication section performs access to the voiceprint database based on voiceprint information acquired by analyzing voice of a caller of an incoming call, so as to authenticate the caller. The notification section notifies a responder of an incoming call of a message when a caller is not registered.SELECTED DRAWING: Figure 2

Description

Embodiments of the present invention relate to telephone systems and server devices.

Damage caused by special fraud known as "Oreore fraud" has become a social problem in recent years. One of the causes of this type of fraud is that the person receiving the call (the responder) cannot reliably identify the person to call. Some phones warn of incoming calls without number notification, while others do not have such a feature. Rather than making complicated settings on the phone, it is desirable to strengthen the support on the system side.

Japanese Unexamined Patent Publication No. 2011-135328 JP-A-2018-107769 Japanese Unexamined Patent Publication No. 2006-324715

Voiceprints are known as features for identifying individuals. The technology for analyzing voiceprints is also developing day by day. It is also being considered to apply this technology to deal with special fraud, but there is still much debate, such as in combination with the IoT (Internet of Things).

The purpose is to provide telephone systems and server devices capable of deterring special fraud.

According to the embodiment, the telephone system includes a voiceprint database in which a pre-registered speaker is associated with the voiceprint, and a voiceprint authentication unit and a notification unit. The voiceprint authentication unit accesses the voiceprint database based on the voiceprint information obtained by analyzing the voice of the callee of the incoming call, and authenticates the callee. The notification unit notifies the respondent of the incoming call of the message when the other party is not registered.

FIG. 1 is a diagram showing an example of a telephone system according to the first embodiment. FIG. 2 is a functional block diagram showing an example of the server device 200 shown in FIG. FIG. 3 is a sequence diagram showing an example of the processing procedure according to the first embodiment. FIG. 4 is a diagram showing an example of a telephone system according to the second embodiment. FIG. 5 is a functional block diagram showing an example of the server device 200 shown in FIG. FIG. 6 is a sequence diagram showing an example of the processing procedure according to the second embodiment. FIG. 7 is a diagram showing an example of a telephone system according to a modified example of the second embodiment. FIG. 8 is a functional block diagram showing an example of the server device 200 according to the third embodiment. FIG. 9 is a sequence diagram showing an example of the processing procedure according to the third embodiment.

The voiceprint is one of the characteristics of our voice. Voiceprint authentication is one of the so-called biometric authentication technologies, and is applied to, for example, operating a device such as an AI (Artificial Intelligence) speaker by voice or unlocking the device. In voiceprint authentication, acoustic features and linguistic features are often treated separately. The acoustic characteristics represent the frequency characteristics of the voice of the person who authenticates, and the linguistic characteristics represent the restrictions on the arrangement of phonemes.
One of the causes of special fraud using a telephone is that the person who answers the telephone cannot confirm the attributes of the other party (whether they are relatives or not, whether they are really bank employees, etc.). The following discloses technologies that can be a deterrent to this type of fraud.

[First Embodiment]
FIG. 1 is a diagram showing an example of a telephone system according to the first embodiment. The telephone system is formed with the server device 200 as the core. The server device 200 is connected to both a telephone network 2 such as (PSTN: Public Switched Telephone Networks) and a WAN (Wide Area Network) 5. WAN 5 is, for example, an IP (Internet Protocol) network or the like, and includes a database 100. The database 100 stores the voiceprint database 100a. The voiceprint database 100a is a database in which a pre-registered speaker is associated with the voiceprint.
In the first embodiment, it is considered that a person involved in the special fraud is associated with the voiceprint and registered in the voiceprint database 100a. A person involved in special fraud is, for example, a person who has been arrested as a criminal of "Oreore fraud" or a person who has been marked by the authorities as a suspect.

The telephone 4 placed in the subscriber's home is connected to the telephone network 2 and WAN 5 via, for example, a connection device 3 placed in the home. The call generated from the telephone 1 connected to the telephone network 2 arrives at the telephone 4 via the connecting device 3. When the respondent 20 of the telephone 4 picks up the handset, a call is set up with the other party 10 of the telephone 1. Here, the telephones 1 and 4 may be either a fixed telephone or a mobile telephone.

FIG. 2 is a functional block diagram showing an example of the server device 200 shown in FIG. The server device 200 includes a processor 250 such as a CPU (Central Processing Unit) and an MPU (Micro Processing Unit), a ROM (Read Only Memory) 220, a RAM (Random Access Memory) 230, a storage unit 240, and a communication unit 270. It is a computer.

The ROM 220 stores basic programs such as BIOS (Basic Input Output System) and UEFI (Unified Extensible Firmware Interface), and various setting data. The RAM 230 temporarily stores programs and data loaded from the storage unit 240.

The communication unit 270 is connected to the telephone network 2 and WAN5 and controls communication with each network. It should be noted that various programs executed by the server device 200 can be downloaded from a cloud server (not shown) via, for example, the communication unit 270 and installed in the storage unit 240. The latest program can be downloaded from the cloud server via the communication unit 270, and the installed program can be updated.
The storage unit 240 is, for example, a recording medium such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores a program 240a executed by the processor 250. The processor 250 executes an OS (Operating System) and various programs.

By the way, the processor 250 includes a voiceprint authentication unit 250a and a notification unit 250b as processing functions according to the embodiment. The voiceprint authentication unit 250a and the notification unit 250b are understood as a process generated by loading the program 240a stored in the storage unit 240 into the RAM 230 and executing arithmetic processing by the processor 250 as the program progresses. obtain.

The voiceprint authentication unit 250a accesses the voiceprint database 100a based on the voiceprint information obtained by analyzing the voice of the other party of the incoming call. If the voiceprint information is registered in the voiceprint database 100a, the voiceprint authentication unit 250a identifies the speaker corresponding to the voiceprint information as a call partner. This authenticates the other party.

The notification unit 250b notifies the responder 20 of the incoming call of a message when the callee is not registered in the voiceprint database 100a. Next, the operation in the above will be described.

FIG. 3 is a sequence diagram showing an example of the processing procedure according to the first embodiment. In FIG. 3, when the other party 10 makes a call, the telephone 4 is called via the telephone network 2 and the respondent 20 is called (step S1). When the responder 20 answers, the call is entered (step S2).

When the call is established, the telephone network 2 provides the call voice (voice information) to the server device 200 (voiceprint authentication unit 250a) (step S3). Here, if the telephone of the respondent 20 is a mobile telephone, it is possible to provide voice information to the server device 200 by the transfer function or the application function of the mobile telephone.

When the voiceprint authentication unit 250a of the server device 200 acquires the voice information, the voiceprint authentication unit 250a performs voiceprint authentication (step S4) and identifies the callee 10 with reference to the voiceprint database 100a. The result of the authentication is passed to the notification unit 250b (step S5).

After that, when the call ends, the call is disconnected, and the call end information (busy tone) from the telephone network 2 reaches the server device 200 (step S7). Then, the voiceprint authentication unit 250a ends the voiceprint authentication process and passes the call end information to the notification unit 250b (step S8).

The notification unit 250b determines whether or not there is a suspicion of fraud in the call based on the authentication result of the voiceprint authentication (step S9). That is, if a person having a criminal record of fraud is identified by voiceprint authentication, the notification unit 250b determines the suspicion of fraud (YES). Then, the notification unit 250b immediately makes an automatic transmission (step S10). That is, after the incoming call is completed, the notification unit 250b sets a new call with the telephone 4 via the telephone network 2 and notifies the respondent 20 of the message (step S11). Here, if the telephone 4 is a mobile telephone, it is possible to automatically notify a message by SMS (Short Mail Service) or an application linkage function.

The message may be in the form of a voice message such as "The previous call is suspected of fraud. Please be careful" or guidance such as "There is a risk of fraud, so please call 03-XXXX-XXXX". You will be notified.

As described above, in the first embodiment, the voiceprints of persons involved in the crime of special fraud are registered in advance in the voiceprint database 100a. When a call is generated, the voice of the other party 10 is sent to the voiceprint authentication unit 250a to perform voiceprint authentication. When it is found by the authentication that the call partner 10 is registered in the voiceprint database 100a, the respondent 20 is notified of the message after the call is completed.

That is, in the first embodiment, the call partner is specified by voiceprint authentication, and the respondent 20 is alerted. As a result, the respondent 20 can be alerted and special fraud such as "Oreore fraud" can be prevented. In addition, voiceprint authentication can be performed without being noticed by the criminal. From these facts, according to the first embodiment, it is possible to provide a telephone system and a server device capable of deterring special fraud.

[Second Embodiment]
In the first embodiment, the voiceprints of persons involved in special fraud are registered in the voiceprint database, and when a match is made by voiceprint authentication, the respondent is notified after the call. In the second embodiment, the technique of registering the voiceprint of the respondent's acquaintance in the voiceprint database will be described. An acquaintance is, in short, a person known to the respondent, and can be regarded as a concept that includes so-called relatives, family members, relatives, friends, and the like.

FIG. 4 is a diagram showing an example of a telephone system according to the second embodiment. In the following, the parts common to each figure of the first embodiment are designated by the same reference numerals, and only the parts different from those of the first embodiment will be described.
In FIG. 4, the AI speaker 6 is connected to the connecting device 3. Further, the voiceprint database (reference numeral 240b) is stored in the server device 200.

FIG. 5 is a functional block diagram showing an example of the server device 200 shown in FIG. The voiceprint database 240b is stored in the storage unit 240. The voiceprint database 240b is a database in which the acquaintances of the respondent 20 and their voiceprints are associated in advance.

The processor 250 includes a registration unit 250c as a processing function according to the embodiment. The registration unit 250c analyzes the voice of the acquaintance 30 (FIG. 4) collected by the AI speaker 6, extracts the voiceprint, associates the obtained voiceprint with the attribute of the acquaintance 30, and registers the obtained voiceprint in the voiceprint database 240b. For example, if the AI speaker 6 is activated by the voice of the same person every day, that person can be registered in the voiceprint database 240b as a family member of the respondent 20. Alternatively, you may ask your family in advance to register your voice using the AI speaker 6.

FIG. 6 is a sequence diagram showing an example of the processing procedure according to the second embodiment. In FIG. 6, when the call end information reaches the server device 200 after the call is disconnected (step S7), the voiceprint authentication unit 250a ends the voiceprint authentication process and passes the call end information to the notification unit 250b (step S8).

The notification unit 250b determines whether or not the other party 10 has been authenticated based on the authentication result of the voiceprint authentication (step S12). If the voiceprint information of the other party 10 exists in the voiceprint database 240b, the notification unit 250b authenticates. If the voiceprint information is not in the voiceprint database 240b, the notification unit 250b performs an automatic transmission process as authentication NG (No Good) in order to notify the respondent 20 that there is no registration (step S10). In step S10, the notification unit 250b notifies the respondent 20 of a message such as "there is no registration in the database".

The respondent 20 can know that the other party 10 is not registered in the database even though he / she claims to be a relative. Therefore, the respondent 20 can sense from this fact that the telephone may be a special fraud.

As described above, in the second embodiment, the voiceprints of acquaintances of the person who uses the telephone are registered in advance in the voiceprint database 100a. When a call is generated, the voice of the other party 10 is sent to the voiceprint authentication unit 250a to perform voiceprint authentication. When it is found by the authentication that the call partner 10 is not registered in the voiceprint database 100a, the respondent 20 is notified of the message after the call ends.

That is, in the second embodiment, when the callee cannot be identified by voiceprint authentication, the respondent 20 is alerted. This also alerts the respondent 20 and prevents special fraud such as "Oreore fraud". In addition, voiceprint authentication can be performed without being noticed by the criminal. From these facts, it is possible to deter special fraud by the second embodiment as well.

[Modified example of the second embodiment]
FIG. 7 is a diagram showing an example of a telephone system according to a modified example of the second embodiment. Instead of the AI speaker 6, the voiceprint information can be registered in the server device 200 by using the smartphone 7 possessed by the acquaintance 30. That is, the application for voiceprint registration is installed in advance on the smartphone 7, and the voice of the owner is collected. The obtained voice information is transferred to the server device 200 via the mobile communication network MN and passed to the registration unit 250c. The registration unit 250c analyzes the voice of the acquaintance 30 to extract the voiceprint, associates the obtained voiceprint with the attribute of the acquaintance 30, and registers the obtained voiceprint in the voiceprint database 240b.

[Third Embodiment]
In the second embodiment, the voiceprints of the respondent's acquaintances are registered in the voiceprint database, and when the voiceprint authentication does not match, the respondent is notified after the call. In the third embodiment, a technique for incorporating wording determination will be described.

FIG. 8 is a functional block diagram showing an example of the server device 200 according to the third embodiment. The processor 250 includes a wording identification unit 250d in addition to the voiceprint authentication unit 250a and the notification unit 250b as processing functions according to the embodiment. The wording specifying unit 250d specifies the wording included in the voice call between the other party 10 and the respondent 20. That is, the wording specifying unit 250d converts the call voice (voice information) in the call state into text and specifies the call content. At that time, the call contents may be classified into a plurality of classes by referring to a dedicated database that holds keyword information of various call contents. Classes include "normal conversation," "school contact network," "sales," and "special fraud." As a technology for converting call voice information into text, there are some that are already provided as a service such as a voice-text conversion function using a cloud server.

The notification unit 250b classifies the content of the voice call based on the wording specified by the wording specifying unit 250d. For classification, the score for each class is calculated. Then, when the score for classifying the call content into the special fraud is equal to or higher than the default value, the notification unit 250b notifies the responder 20 of a message suggesting that there is a possibility of the special fraud.

FIG. 9 is a sequence diagram showing an example of the processing procedure according to the third embodiment. In FIG. 9, the voiceprint authentication unit 250a performs voiceprint authentication (step S4). As a result, if it is found that the other party 10 is a person to be noted, the voiceprint authentication unit 250a notifies the wording identification unit 250d of the result of the voiceprint authentication and the voice information (step S13).

The wording specifying unit 250d converts the received call voice information into text and specifies the call content based on the keyword information or the like registered in advance (step S14). The specific result is passed to the notification unit 250b (step S15). It is possible to determine whether or not the target call is "Oreore fraud" by the combination of the person identified by the voiceprint authentication and the identified call content.

When the call ends, the call end information is sent from the telephone network 2 to the server device 200, and reaches the voiceprint authentication unit 250a, the wording identification unit 250d, and the notification unit 250b (steps S7, S8, S16). Then, the notification unit 250b classifies the contents of the call (step S17) and determines whether or not the classification score for special fraud exceeds a predetermined threshold value (step S18). If the score is equal to or higher than the threshold value, the notification unit 250b immediately makes an automatic transmission (step S10) and notifies the respondent 20 of a message suggesting suspicion of fraud (step S19).

As described above, in the third embodiment, the call partner 10 is authenticated by voiceprint authentication, and the content of the voice call converted into text is classified by the wording specifying unit 250d. Then, for a call determined to have a high classification score for special fraud, a message is notified to the respondent 20 after the call ends.

By combining voiceprint authentication and wording judgment, for example, even a phone call that requires money can be prevented from being judged as a special fraud if it is a relative. That is, the accuracy of the determination can be improved. Therefore, it is possible to deter special fraud by the third embodiment as well.

The present invention is not limited to the above embodiment. For example, in each of the above embodiments, a message is sent to the respondent 20 to call attention to the respondent 20. According to this, special fraud can be deterred without being noticed by the criminal. On the contrary, by notifying the other party 10 of the message, it is possible to more actively deter special fraud. In such a case, it is conceivable to notify a message such as "This call has been authenticated by voice".
In addition, voice call records and textualized data of call contents may be stored in the cloud as evidence. This information may be used as evidence in the arrest of the criminal.

Further, the functions of the voiceprint authentication unit, the notification unit, the registration unit, and the wording identification unit may be distributed to a plurality of servers without being integrated in the server device 200. For example, a voiceprint authentication system having a voiceprint authentication unit function, an automatic notification system having a notification unit function, a registration system having a registration unit function, and a wording identification system having a wording identification unit function are connected to WAN5, and each of them is connected. By exchanging information and data between the systems, it is possible to obtain the same effects as those of the first to third embodiments.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other embodiments, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are also included in the scope of the invention described in the claims and the equivalent scope thereof.

1 ... Telephone, 2 ... Telephone network, 3 ... Connection device, 4 ... Telephone, 6 ... AI speaker, 7 ... Smartphone, 10 ... Call partner, 20 ... Respondent, 30 ... Acquaintance, 100 ... Database, 100a ... Voice print database, 200 ... server device, 220 ... ROM, 230 ... RAM, 240 ... storage unit, 240a ... program, 240b ... voiceprint database, 250 ... processor, 250a ... voiceprint authentication unit, 250b ... notification unit, 250c ... registration unit, 250d ... wording Specific part, 270 ... Communication part.

Claims (10)

A voiceprint database that associates pre-registered speakers with their voiceprints,
A voiceprint authentication unit that accesses the voiceprint database based on the voiceprint information obtained by analyzing the voice of the other party of the incoming call and authenticates the other party.
A telephone system including a notification unit that notifies a respondent of an incoming call of a message when the other party is not registered. The voiceprint database is a database in which a person involved in a special fraud is associated with the voiceprint.
The telephone system according to claim 1, wherein the notification unit notifies the responder that the authenticated call partner is a party involved in the special fraud. The telephone system according to claim 1, wherein the voiceprint database is a database in which an acquaintance of the responder is associated with the voiceprint. The telephone system according to claim 3, further comprising a registration unit that analyzes the voice of the acquaintance, extracts the voiceprint, associates the voiceprint with the attribute of the acquaintance, and registers the voiceprint in the voiceprint database. A voiceprint database that associates pre-registered speakers with their voiceprints,
A voiceprint authentication unit that accesses the voiceprint database based on the voiceprint information obtained by analyzing the voice of the other party of the incoming call and authenticates the other party.
A wording specifying unit that specifies the wording included in the voice call between the other party and the respondent of the incoming call,
It is provided with a notification unit that classifies the content of the voice call based on the specified wording and notifies the responder of a message when the score for classifying the content as a special fraud is equal to or higher than a default value. Telephone system. The voiceprint database is a database in which a person involved in the special fraud is associated with the voiceprint.
The telephone system according to claim 5, wherein the notification unit notifies the responder that the authenticated call partner is a party involved in the special fraud. The voiceprint database is a database in which a person involved in the special fraud is associated with the voiceprint.
The telephone system according to claim 5, wherein the notification unit notifies the authenticated call partner of the message. The telephone system according to any one of claims 1 to 7, wherein the notification unit sets a new call after the end of the incoming call and notifies the message. In a server device that can connect to the telephone network
Based on the voiceprint information obtained by analyzing the voice of the other party of the incoming call generated in the telephone network, the voiceprint database that associates the pre-registered speaker with the voiceprint is accessed to access the voiceprint database of the other party. Voiceprint authentication department that certifies
A server device including a notification unit that notifies a respondent of an incoming call to a message via the telephone network when the other party is not registered. In a server device that can connect to the telephone network
Based on the voiceprint information obtained by analyzing the voice of the other party of the incoming call generated in the telephone network, the voiceprint database that associates the pre-registered speaker with the voiceprint is accessed to access the voiceprint database of the other party. Voiceprint authentication department that certifies
A wording specifying unit that specifies the wording included in the voice call between the other party and the respondent of the incoming call,
The content of the voice call is classified based on the specified wording, and when the score for classifying the content as a special fraud is equal to or higher than the default value, the respondent is notified of a message via the telephone network. A server device including a notification unit.