KR20110096329A - Method for construction personal phone book database using speech recognition, and automatic call connection service method and system using it - Google Patents

Abstract

PURPOSE: A personal telephone number unit data base estimating method, method for serving automatic call connection using the same, system thereof using voice recognition are provided to automatically connecting a call with a desired person by using a voice. CONSTITUTION: A voice and receiving telephone number about a name of a caller is inputted through a call network from a service user which a personal telephone number unit estimation is requested. The caller telephone number of a service user is obtained. A speaker authentication voice characteristics and speaker-dependent voice recognition name characteristics is extracted at the same time from name of a caller by using a voice recognition(306). By using the obtained caller number as a key, the speaker-dependent voice recognition characteristics per the extracted speaker authentication voice characteristics and the inputted calling number are stored(308).

Description

METHOOD FOR CONSTRUCTION PERSONAL PHONE BOOK DATABASE USING SPEECH RECOGNITION, AND AUTOMATIC CALL CONNECTION SERVICE METHOD AND SYSTEM USING IT}

The present invention, when a person wants to use a landline phone can not remember all the phone numbers of people who know the individual using a landline phone using the memory of the paper notebook or mobile phone, to solve the inconvenience of this process The present invention relates to a method of automatically connecting a telephone connection with a person using voice, and to a telephone book by adding a voice recognition function to a landline telephone of the prior art and a method of automatically connecting a telephone from a telephone and an exchange in a public telephone network. There is a significant difference in the network configuration, the technology of use element, and the method of using the service. In addition, the speaker-independent recognition rate of the speaker-independent large-capacity voice recognition method can not obtain the recognition rate of 95% or more necessary for commercial service. Possible recognition rates are difficult to achieve in the next five to ten years. In addition, in order to distinguish between users of the same terminal, an additional means such as a password input should be provided. Even if such additional means are provided, factors that inconvenience the user such as time and effort required for input remain.

Since the early 1990s, there have been many cases of domestic and international reports on the recognition of more than 97% of dozens of isolated words in the telephone environment. Therefore, in order to distinguish between the high recognition rate of the speaker and the users of the same terminal, there is a need for a means that can be used in a convenient way without providing additional means such as a password input. It is about a method of constructing an individual's personal phone book database using a telephone or the Internet in a wired communication network, and automatically connecting a call to a called phone number corresponding to the name by simply speaking the name registered in the established personal phone book database. It implements voice automatic connection service of personal phonebook by using voice technology of speaker authentication, voice recognition technology of speaker dependent, intelligent network technology supporting CCS-NO7 signal, packet network technology supporting real time service. The purpose of the present invention is to provide a method that can be used for automatic connection of a phone anytime, anywhere through a public telephone network.

To date, general landline telephone users often use a small notebook made of paper or a storage function of a mobile phone to remember a phone number associated with each individual, but often suffer from fluoride and loss. To solve this inconvenience, if a voice recognition chip is built into the wired telephone and the name is spoken by voice, the incoming telephone number corresponding to the name is found in the database built inside the telephone, and the MFC signal is received from the telephone to the local telephone exchange. Configuration 1 to make a call to connect, built-in voice recognition chip in a wired telephone, and speak the name by voice, find the incoming phone number corresponding to the name in the PC or database server connected to the telephone, the incoming call from the telephone to the local telephone exchange Configuration 2, which sends a number MFC signal to connect a telephone, built-in voice recognition, voice guidance device, and voice recognition database in the switchboard of public telephone network, and speaks the name in the landline telephone and finds the telephone number corresponding to the name. Configuration 3 to connect, but configuration 1 In addition to the cost of configuring the voice recognition module, inconvenience of personal database management, and the like, there is a considerable difference in the method of using the service. Configuration 2 is convenient for personal database management, but the cost of configuring a voice recognition module, a connection problem between a telephone and an information storage device, and the like, and there is a considerable difference from the present invention in terms of service usage. Configuration 3 is a method of implementing a personal phone number connection function by configuring a voice recognition module at the exchange. Since the speaker authentication process is not performed in the network configuration and service use, it is inconvenient for the service users to input the authentication process in the form of a password. The implementation for the service user has a significant difference in the implementation cost, network configuration, real-time service processing capability, convenience, etc. compared with the present invention. As compared with the present invention, the above-described configuration 1, configuration 2, and configuration 3 have considerable differences in service configuration technology, service configuration cost, service convenience, and the like.

The present invention, when a person wants to use a landline phone can not remember all the phone numbers of people who know the individual using a landline phone using the memory of the paper notebook or mobile phone, to solve the inconvenience of this process The present invention relates to a method of automatically connecting a telephone connection with a person using voice, and to a telephone book by adding a voice recognition function to a landline telephone of the prior art and a method of automatically connecting a telephone from a telephone and an exchange in a public telephone network. There is a significant difference in the network configuration, the technology of use element, and the method of using the service. In addition, the speaker-independent recognition rate of the speaker-independent large-capacity voice recognition method can not obtain the recognition rate of 95% or more necessary for commercial service. Possible recognition rates are difficult to achieve in the next five to ten years. In addition, in order to distinguish between users of the same terminal, an additional means such as a password input should be provided. Even if such additional means are provided, factors that inconvenience the user such as time and effort required for input remain.

Since the early 1990s, there have been many cases of domestic and international reports on the recognition of more than 97% of dozens of isolated words in the telephone environment. Accordingly, in order to distinguish between the high recognition rate of the speaker and the users of the same terminal, a means that can be used in a convenient way without additional means such as a password input is required.

The present invention devised to solve the problems described above, by using the telephone, each person's information file, the outgoing telephone number, name text (character), speaker authentication feature (feature) of the individual name, speaker dependent voice recognition name The personal phone book is composed of the called phone number corresponding to the feature and the name, and the voice phone is used to connect to the called phone number registered in the phone book by using voice recognition technology, intelligent network, packet network, and public telephone network. Its purpose is to provide a convenient voice connection anytime, anywhere.

The present invention simultaneously extracts two voice features (speaker authentication voice feature and speaker dependent voice recognition name feature) from the voice name of the person who wants to connect one at a time, the caller's telephone number, name text (text), called party How to register a phone book of individuals consisting of a phone number, etc. to build a database in the packet network, and to automatically connect the incoming phone number with one statement voice at a time using the built-in personal phone book of the database To provide. Voice recognition telephone is extracted without the user's effort or time wasted by simultaneously extracting two voice features of speaker authentication voice feature and speaker dependent voice recognition feature It provides a method that can use user specific voice recognition and speaker recognition with high recognition rate required for automatic connection. First, the terminal user is identified by a plurality of users who use the same terminal as a speaker authentication feature, and only the information registered in the personal phonebook of the specified user is recognized as a speaker-dependent voice recognition target, and then the recognized called party is called. Send the text of the name to the service user through the voice guidance device and pass the called party's phone number and the service user's outgoing phone number to the intelligent network service control point. By controlling the telephone exchange, the incoming and outgoing call is connected, making the telephone connection of the service user quick and convenient.

In addition, the present invention, in the automatic telephone connection service method using the voice recognition, an input step of receiving a caller's name by voice from the service user requesting the automatic telephone connection service through the telephone network, and obtaining the outgoing telephone number of the service user A voice feature extraction step of extracting a speaker authentication voice feature and a speaker dependent voice recognition statement feature from the voice input caller name; An authentication step of searching for a corresponding speaker authentication voice feature in a personal phonebook database using the obtained outgoing call number as a key and then performing authentication of the service user through comparison with the extracted speaker authentication voice feature; ; If the authentication is successful in the authentication step, the speaker-dependent voice recognition statement feature corresponding to the extracted speaker-dependent voice recognition statement feature is retrieved from the personal phone book database, and the called party corresponding to the retrieved speaker-dependent voice recognition statement feature A called number search step for obtaining a called number; And a telephone connection step of performing a telephone connection by using the obtained outgoing telephone number and the called telephone number.

In addition, the present invention, in the automatic telephone connection service system using the voice recognition, by receiving the caller's name by voice from the service user requesting the automatic telephone connection service through the telephone network to extract the speaker authentication voice feature and speaker dependent voice recognition statement feature A voice feature extraction device for acquiring an outgoing telephone number of the service user; And performing authentication on the service user through comparison of the corresponding speaker authentication voice feature retrieved from the personal phonebook database with the extracted caller authentication voice feature using the obtained outgoing call number as a key. And a voice recognition server for retrieving the corresponding incoming call number from the personal phonebook database using the speaker-dependent voice recognition statement feature and delivering the received call number to the intelligent network service control point (SCP).

Before describing the present invention in detail, all the processes of the present invention will be described. The present invention relates to a personal telephone book voice automatic connection service method in a public telephone network, wherein each service user corresponds to an individual telephone number. A first step of allocating a personal database; Sends the name text, speaker authentication voice feature, speaker dependent voice recognition feature, caller ID, etc. A second step of storing a telephone number as a key in the personal database; According to the request for connection of the called telephone number requested by the name of the called party, the service user calls the speaker authentication voice feature and speaker dependent voice recognition feature in the voice guidance and feature extraction device. Simultaneously extracting and sending the same to the voice recognition server / database such as an outgoing call number obtained by requesting the local exchange; A fourth step of authenticating a user with the speaker authentication voice feature in a voice recognition server / database and finding a name text and an incoming call number with a speaker-dependent voice recognition feature; Send the name text found in the above step to the voice guidance and voice feature extraction device 30, and search the syllable database stored in syllable units in the voice guidance and voice feature extraction device 30. A fifth step of combining the syllables into the syllables received from the voice recognition server / database and reproducing the voices to the service user so that the service user can confirm the name of the called party; A sixth step of passing the outgoing call number and the incoming call number obtained in the third and fourth steps from the voice recognition server / database to the intelligent network service control point after the checking step is finished; And a seventh step of controlling the telephone connection by the outgoing call number and the incoming call number received from the intelligent network service control point.

Here, the second step is the step of building the personal phone book database, specifically, an eighth step of requesting an outgoing local exchange from the voice guidance and voice feature extraction device to obtain an outgoing telephone number; A ninth step of simultaneously extracting a speaker authentication voice feature and a speaker dependent voice recognition feature from the name of one called party in the voice guidance and voice feature extraction device; The voice guidance and voice feature extraction device extracts the text of the name spoken by the service user with speaker independent voice recognition technology to obtain a text or a voice recognition feature through the confirmation of the device operator. A tenth step of collectively constructing a personal phone book database by allowing a service user to access a server on which software is installed; Outgoing phone number obtained in the eighth, ninth and tenth steps, speaker authentication voice feature (feature), speaker-dependent voice recognition feature, voice text and incoming call number An eleventh step of transmitting to a voice recognition server / database through a packet network (or an internet network) to a database; After performing the above steps, the voice recognition server / database sends the name text to the voice guidance and feature extraction device and sends the called phone number pressed by the service user to the service user from the voice guidance and feature extraction device. A twelfth step of sending by voice to confirm; After the process from the eighth to the twelfth step, the outgoing call number, name text, speaker authentication voice feature, speaker dependent voice recognition in the personal database with the outgoing phone number as a key. And a thirteenth step of storing a feature, an incoming call number, and the like as shown in FIGS. 4A and 4B.

A service user can pick up a phone configured as a dash line and wait for a few seconds, press a speed dial with a service identification number, or press the service identification number and call the name of the person who wants to make a call according to the announcement. In other words, the caller's telephone number is keyed by the caller's telephone number received from the voice guidance device of the packet network requested by the exchange, the speaker authentication voice feature extracted from the name spoken by the service user, the speaker-dependent voice recognition name feature, and the like. key) to search the preconfigured database and perform speaker authentication process, search for incoming phone number with speaker-dependent voice recognition feature, and then voice the name text corresponding to the called number. (feature) After sending to the extraction device (hereinafter referred to as voice guidance device) and the recipient of the service user confirmation process, the calling party By controlling the telephone connection from the intelligent network service control point in an intelligent network service control point neomgimyeon service users can conveniently and quickly call. In order to receive the same service over another person's phone and to charge the phone number in his / her own name, in addition to the above-described procedure, the service user may enter his / her phone number before speaking the name.

The process of building a personal phone book by the service user is to say the name two or three times according to the announcement by using the identification number for constructing the database different from the service use identification number, and press the telephone number corresponding to the name to the database of the network. Using the outgoing phone number as a key, the name text, the speaker authentication voice feature, the speaker-dependent voice recognition name feature, the called phone number, etc. are stored. For convenience, if you say the name once and press the called phone number, the speaker independent voice recognition feature of the name is added to the voice recognition server / database constructed in the network to build the database. Recognition rate can be obtained. In the above process, the name text is a speaker-independent speech recognition technology from the name spoken by the user of the service, and is registered in the voice recognition database after the voice to text process and the device operating agent.

The personal telephone book database may be constructed on the Internet to establish a server capable of performing the functions of the database construction voice recognition process, and store the data directly by a service user.

Voice recognition telephone automatic connection service users enter the service identification number of 1582, 1590, and then speak the name of the person who wants to connect the phone. The user authentication and voice recognition can be made without additional efforts by the user, and the speaker dependency can be obtained. Can use isolated word speech recognition technology. The present invention relates to a method of simultaneously extracting, storing, recognizing, and using two voice features (speaker authentication voice feature and speaker dependent voice recognition name feature) from a voice of a person to be connected one at a time. It is also applicable to many speech recognition applications that require dependent reliable word recognition rates. The present invention is a combination of various technologies and networks such that the service user can recognize the convenience and speed of the wired communication network voice call service that aims at convenience and speed. With real-time service processing capability, convenience, commercial service recognition rate and economic feasibility, it is possible to construct and operate voice automatic connection service of personal phone book.

1 is an exemplary diagram of a network to which the present invention is applied.
Figure 2 is a flow chart of an embodiment of the incoming call number calling process of the personal phone book voice automatic connection service method according to the present invention.
Figure 3 is a flow chart of an embodiment of the process of storing the incoming call number in the personal phone book voice automatic connection service method according to the present invention.
4A and 4B are diagrams illustrating one embodiment of contents and a data configuration method of a personal information file stored in a voice recognition server / database of a personal phone book voice automatic connection service method according to the present invention;

According to an aspect of the present invention, there is provided a personal phonebook voice automatic connection service method in a public telephone network, comprising: a first step of allocating a personal database corresponding to a personal telephone number for each service user; Sends the name text, speaker authentication voice feature, speaker dependent voice recognition feature, caller ID, etc. A second step of storing a telephone number as a key in the personal database; The name of one called party to be called by the service user, and the caller's voice feature and the speaker's dependent voice recognition feature are used by the voice guidance and voice feature extraction device according to the connection request. Simultaneously extracting and sending the same to the voice recognition server / database such as an outgoing call number obtained by requesting the local exchange; A fourth step of authenticating a user with the speaker authentication feature in a voice recognition server / database and finding a name text and an incoming call number with a speaker dependent voice recognition feature; Send the name text found in the above process to the voice guidance and feature extraction device 30 and search the syllable database stored in syllable units in the voice guidance and feature extraction device 30. A fifth step of combining the syllables into the syllables received from the voice recognition server / database and reproducing the voices to the service user so that the service user can confirm the name of the called party; A sixth step of transferring the outgoing call number and the incoming call number obtained in the third and fourth steps from the voice recognition server / database to the intelligent network service control point after the checking process is finished; And a seventh step of controlling the telephone connection by the outgoing call number and the incoming call number received from the intelligent network service control point.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating an organic configuration of a voice guidance and feature extraction apparatus, a voice recognition server / database, a packet network (or an internet network), an intelligent network, and the like on a public telephone network to which the present invention is applied. Here, the voice guidance and voice feature extraction device 30, the voice recognition server 40, and the personal phone book DB 41 may be referred to as "automatic telephone connection service system using voice recognition".

On the public telephone network to which the present invention is applied, a personal telephone directory of each individual constituted using a telephone or the Internet is configured in the voice recognition server / databases 40 and 41, and the personal telephone directory database 41 is registered. Incoming phone number is a service that allows the service user to conveniently connect the phone automatically from each phone connected to the public telephone network.

The personal phone book automatic connection service to be implemented by the present invention is a service user holding a phone and waiting for 5 seconds (dash line: warm line), entering a service identification number, or a speed dial number (2 digits registered with a service identification number). Press), press the star, listen to the voice prompt "tell me your name", listen to the announcement to confirm that the name is correct, and wait for a few seconds for the signal to be correct. You can hear the tone.

In the process, the service identification number entered by the service user or automatically inputted by the speed dial phone or the dash direct phone is translated by the outgoing local exchange 10 to extract the voice guidance and the voice feature connected to the voice call path 62. Open the voice call path 62 to the device (30).

The voice guidance and voice feature extraction device 30 sends a zero-started guide tone (“tell me a name”) to the subscriber after receiving and requesting an outgoing telephone number from the local exchange 10.

When the subscriber speaks his name, the voice guidance and voice feature extraction device 30 is a speaker authentication voice feature (named as a speaker feature for voice authentication, which may be simply referred to as a speaker authentication feature) and a speaker. It extracts the feature of the dependent voice recognition name and transmits it to the voice recognition server / database 40 and 41 through the packet network (Internet network) 66 as the calling number received by requesting from the outgoing local exchange 10. .

Then, the voice recognition server / database 40, 41 searches the database assigned to the number using the caller's number of the received data as a key, and the speaker recognizes the caller's number by the speaker's voice feature. The name text and the called telephone number having a feature corresponding to the speaker-dependent voice recognition feature are also verified.

The voice recognition server / database 40 and 41 sends the found name text back to the voice guidance and feature extraction device 30 to find the name in syllable units in the syllable syllable database stored in syllable units. After confirming that the name is correct for the service user, the incoming call number and the received outgoing call number obtained in the above process are sent to the service control point 50 of the intelligent network linked to the Internet 68.

The intelligent network service control point 50 controls the outgoing city exchange 10 and the incoming city exchange 20 through the intelligent network 64 with the received outgoing number and the incoming number and thus, the outgoing telephone 11 and the incoming telephone 21. Send a ring tone to

In addition, the intelligent network service control point 50 receives a call by the person receiving the telephone receiver 21 receives the call and connects the public telephone network 60 at the same time so that the phone is connected and manages the call start time.

After the call, the outgoing city exchange 10 reports the call termination time to the intelligent network service control point 50 so that the call charge data is generated at the intelligent network service control point 50.

Personal information storage of the personal phone book to be implemented by the present invention is to create a personal information database using the voice recognition feature extraction software supplied through the Internet, and then through the appropriate authentication process voice recognition server / database (40, 41) There is a method of registering in a batch and a service identification number for storage using a telephone and a name and a called number corresponding to the name according to the announcement. In the latter method, when a service user enters a storage service identification number, the user inputs a name two or three times according to the voice guidance and the feature extraction function 30 and then inputs an incoming call number corresponding to the name. When the voice guidance feature and the feature extraction unit 30 extracts the speaker authentication feature and the speaker dependent voice recognition name feature, and extracts the feature by the speaker independent voice recognition technology, the service operation personnel confirms the statement. The text (text), the calling number, and the called number are transmitted together with the voice recognition server / database 40 and 41 to be stored. Send the stored name text and the called number back to the voice guidance and feature extraction device 30 to find the name in syllable units in the syllable database stored in syllable units, and play the name in voice and send it to the service user. The statement is checked to see if it is correct. The above process is made as needed by the service user to establish a personal phone book database.

Figure 2 is a flow chart of an embodiment of the incoming call number calling process of the personal phone book service method according to the present invention.

As shown in Figure 2, the personal phone book voice automatic connection service according to the present invention, first to connect the incoming call number corresponding to the name of the service user information file stored in the voice recognition server / database 40, the service user If the user picks up the phone and waits for a few seconds (dash direct call), presses a speed dial to enter a service identification number, or presses the service identification number (200), the local exchange 10 receives a voice guidance and a feature extraction device. (30) and sends a voice announcement (204) to say the name. When the service user speaks the name by voice, the voice guidance and voice feature extraction device 30 extracts the speaker authentication feature of the name and the speaker-dependent voice recognition name feature (206) and transmits the call. It transmits to the voice recognition server / database 40, 41 together with the calling number obtained by requesting to the exchange 10. The voice recognition server / database 40 searches the database assigned to the service user by using the received caller ID as a key (208), performs speaker authentication with the speaker authentication feature, and then uses the speaker-dependent voice recognition feature ( The called party number and the name text corresponding to the feature are found (208). Sends the name text corresponding to the extracted called number to the voice guidance and voice feature extraction device 30 and finds the name text in syllable units in the syllable database stored in syllable units. Play to the service user to check whether the input service request name is correct (210), and send the extracted called number and the calling number to the intelligent network service control point 50 (212) intelligent network service control point 50 By controlling the call it connects the outgoing telephone and the incoming telephone (214).

That is, the voice recognition server / database converts the outgoing call number and the incoming call number found through the above process into an intelligent network service request message, and transmits it to the intelligent network service control point 50 through the packet network. Then, the intelligent network service control point 50 controls the outgoing time switch 10 using the outgoing call number received in the form of a service request message, and controls the outgoing time switch 20 using the incoming call number. The telephone 11 and the incoming telephone 21 are connected.

3 is a flowchart illustrating a process of storing a personal phone book database in a personal phone book service method according to the present invention.

As shown in FIG. 3, when the service user presses the service identification number 300 to build a database 41 assigned to the individual in the personal phone book service according to the present invention, the outgoing local exchange 10 receives voice guidance and A voice call path 62 connected to a feature extractor 30 is connected 302. Voice guidance and feature extraction device (30) is obtained by requesting the outgoing telephone number to the outgoing city switchboard (10) to send a service announcement "tell me your name" to the service user (304) to enter the name (306), and sends an announcement "Enter an incoming telephone number" (304) to enter the incoming telephone number (306). Extracts the speaker's authentication feature of the input name and the speaker's dependent voice recognition name feature and extracts the speaker's independence voice recognition technology. When sent to the voice recognition server / database (40, 41) (306), the voice recognition server / database (40, 41) through the name verification process of the called party 310, the personal phone book as shown in FIG. Build a database (308). The caller's name confirmation process 310 sends the extracted name text and the called number once again to the voice guidance and voice feature extraction device 30 and sends the name in syllable units in the syllable syllable database. Find and play the name by voice and send it to the service user with the called number to check whether it is correct (310), according to the result of the above process by requesting the input again or not, the announcement is good input (312) and Finish the save operation together. The process is repeated as needed by the service user.

The above process provides software having the functions of voice guidance and voice feature extraction device 30 and voice recognition servers / databases 40 and 41 on the Internet so that service users collectively make outgoing telephone numbers, It is possible to compose and store the name text, the speaker authentication feature, the speaker dependent voice recognition feature, and the called telephone number.

4A and 4B are diagrams illustrating an example of a storage form of elements constituting the personal phonebook database in the personal phonebook voice automatic connection service method according to the present invention.

As shown in Figs. 4A and 4B, a plurality of speaker authentication features (e.g., the number of family members) are included in one outgoing telephone number, and a plurality of statement texts are in one speaker authentication feature. multiple text and speaker-dependent speech recognition features (e.g., the number of personal information of one family member) in one text and speaker-dependent speech recognition features Incoming phone numbers (eg, home phone number, office phone number, mobile phone number, etc.) can be registered. That is, the personal telephone directory database uses the outgoing telephone number of the service user obtained from the local exchange as a key, and the outgoing telephone number, the speaker authentication voice feature, the name text, and the speaker dependent voice recognition name feature ( feature), called party's phone number, and so on.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawing.

* Explanation of symbols for the main parts of the drawings
10: caller's local exchange 11: caller's phone
* 20: Local Exchange 21: called phone
30: voice guidance and feature extraction device 40: voice recognition server
41: voice recognition database
50: intelligent network service control point (SCP) 60: public telephone network (PSTN)
62: R2-Trunk 64: Next Generation Intelligent Network
66: Packet Network / Internet
68: Dedicated communication line for interworking between packet network / internet and intelligent network

Claims (28)

In the method of building a personal phone book database using voice recognition,
An input step of receiving a 'voice for the name of the called party' and the 'called telephone number' from the service user who has requested the construction of the personal telephone directory, and obtaining the outgoing telephone number of the service user;
A voice feature extraction step of simultaneously extracting a speaker authentication voice feature and a speaker dependent voice recognition statement feature from the one voice input name using the voice recognition; And
A storing step of storing a speaker dependent voice recognition name feature and the received incoming call number for each extracted speaker authentication voice feature by using the obtained outgoing call number as a key;
Personal phone book database construction method comprising a. The method of claim 1,
The input step,
And providing a guidance message according to a service request of the service user, and receiving the name of the called party and the called telephone number through a telephone network in response to the provided guidance message. The method of claim 1,
The input step,
Connecting the voice call path to the voice guidance and voice feature extraction apparatus when the calling exchange receives the corresponding service identification number from the service user;
Obtaining, by the voice guidance and voice feature extracting device, an outgoing telephone number of the service user from the originating exchange; And
Receiving, by the voice guidance and voice feature extraction device, the called party's name from the service user by voice and the incoming call number through a button input;
Personal phone book database construction method comprising a. The method of claim 1,
The voice feature extraction step,
Performing a process of obtaining a caller name text from the voice input caller name using speaker independent speech recognition;
The storing step,
And additionally storing the obtained called party name text. The method of claim 4, wherein
Before performing the storing step, confirming the accuracy of receiving the called name text and the called telephone number reproduced by voice and providing to the service user to confirm the accuracy of the data to be stored.
Personal phone book database construction method further comprising a. 6. The method according to any one of claims 1 to 5,
The voice feature extraction step,
Performed by a voice guidance and voice feature extraction device connected to the calling exchange through a voice call path;
The voice guidance and voice feature extraction device may be configured to deliver the received incoming call number, the obtained outgoing call number, and the extracted speaker authentication voice feature and speaker dependent voice recognition name feature to a voice recognition server through a packet network. Characterized in the personal phone book database construction method. The method according to claim 6,
The storing step,
And the voice recognition server stores in a personal phone book database (DB) constructed on the packet network. The method of claim 7, wherein
The voice recognition server,
A method of constructing a personal phone book database, which is connected to an intelligent network service control point (SCP) for interworking with an intelligent network. In the automatic telephone connection service method using voice recognition,
An input step of receiving a name of a called party by voice from a service user who has requested an automatic telephone connection service through a telephone network and obtaining an outgoing telephone number of the service user;
A voice feature extraction step of simultaneously extracting a speaker authentication voice feature and a speaker dependent voice recognition statement feature from the one voice input name;
An authentication step of searching for a corresponding speaker authentication voice feature in a personal phonebook database using the obtained outgoing call number as a key and then performing authentication of the service user through comparison with the extracted speaker authentication voice feature; ;
If the authentication is successful in the authentication step, the speaker-dependent voice recognition statement feature corresponding to the extracted speaker-dependent voice recognition statement feature is retrieved from the personal phone book database, and the called party corresponding to the retrieved speaker-dependent voice recognition statement feature A called number search step for obtaining a called number; And
Telephone connection step of performing a telephone connection using the obtained outgoing call number and incoming call number
Automatic telephone connection service method comprising a. The method of claim 9,
The telephone connection step,
An intelligent telephone service control point (SCP) using the obtained outgoing call number and the incoming call number to control the corresponding call exchange and the call exchange to perform a call connection. The method of claim 10,
The input step,
And providing a guide message according to a service request of the service user, and receiving the name of the called party by voice in response to the guide message from the service user. The method of claim 11,
The service request,
By the service user holding the calling phone and waiting for a predetermined time, by the service user speed dialing, or by the service user directly inputting an automatic telephone connection service identification number. An automatic telephone connection service method. The method of claim 10,
Acquisition of the name text of the called party corresponding to the speaker-dependent voice recognition name feature retrieved in the called number search step, and reproduces the name of the called party by voice from the obtained caller's name text to the service user to match the called party; Steps to Confirm
Automatic telephone connection service method further comprising. The method of claim 10,
The input step,
Connecting the voice call path to the voice guidance and voice feature extraction apparatus when the calling exchange receives the corresponding service identification number from the service user;
Obtaining, by the voice guidance and voice feature extracting device, an outgoing telephone number of the service user from the originating exchange; And
Receiving, by the voice guidance and voice feature extraction device, the called party's name from the service user by voice;
Automatic telephone connection service method comprising a. The method of claim 14,
The voice feature extraction step,
An extracting step of extracting, by the voice guidance and voice feature extracting device, a speaker authentication voice feature and a speaker dependent voice recognition statement feature simultaneously from the single voice name inputted by voice recognition; And
The voice guidance and voice feature extraction device delivers the extracted speaker authentication voice feature and speaker dependent voice recognition name feature and the obtained outgoing call number to the voice recognition server on the packet network
Automatic telephone connection service method comprising a. The method of claim 15,
The called number search step,
The voice recognition server searches for a speaker dependent voice recognition name feature corresponding to the extracted speaker dependent voice recognition name feature in the personal phonebook database using the obtained outgoing call number, and the retrieved speaker dependent voice recognition name feature. Obtaining a called telephone number of a called party corresponding to the called party; And
Transmitting, by the voice recognition server, the obtained outgoing call number and incoming call number to the intelligent network service control point (SCP);
Automatic telephone connection service method comprising a. The method of claim 10,
The personal phone book database,
And the service user accesses a server having a voice recognition feature extraction function through the Internet and directly registers the service. The method of claim 10,
The personal phone book database,
And the service user is configured to perform voice input and button input using a telephone terminal. In the automatic telephone connection service system using voice recognition,
Voice feature for extracting the speaker authentication voice feature and speaker dependent voice recognition name feature simultaneously by receiving one caller's name through voice from the service user who requested the automatic telephone connection service, and obtaining the outgoing telephone number of the service user. Extraction device; And
Using the obtained outgoing call number as a key, the service user is authenticated through comparison of the corresponding speaker authentication voice feature retrieved from the personal phonebook database with the extracted speaker authentication voice feature, and the extracted Voice recognition server for retrieving the corresponding incoming phone number from the personal phonebook database using the speaker dependent voice recognition statement feature and delivering it to the intelligent network service control point (SCP) with the obtained outgoing call number
Automatic telephone connection service system comprising a. The method of claim 19,
The voice feature extraction device,
And transmitting the obtained outgoing call number and the extracted speaker authentication voice feature and speaker dependent voice recognition name feature to the voice recognition server through a packet network. The method of claim 19,
The voice feature extraction device,
And providing a guide message according to a service request of the service user, and receiving the name of the called party by voice in response to the guide message from the service user. The method of claim 19,
The voice recognition server,
Search for the speaker dependent voice recognition name feature corresponding to the extracted speaker dependent voice recognition name feature in the personal phonebook database using the obtained outgoing call number as a key, and search for the speaker dependent voice recognition name feature. Automatic telephone connection service system, characterized in that to obtain the corresponding incoming call number. The method of claim 22,
The voice recognition server,
Searching for the called party name text corresponding to the retrieved speaker dependent voice recognition name feature in the personal phonebook database and transmitting the retrieved name text to the voice feature extraction device;
The voice feature extraction device,
And reproducing the name of the called party by voice from the received name text of the called party and providing the service user with the service user to confirm whether the called party matches. The method of claim 19,
The voice recognition server,
And converting the obtained outgoing telephone number and the retrieved incoming telephone number into an intelligent network service request message and transmitting the converted outgoing telephone number to the intelligent network service control point (SCP). The method of claim 19,
The voice feature extraction device,
From the service user who requested the construction of the personal phone book, the voice of the called party's name and the called telephone number are inputted through the telephone network, and the speaker authentication voice feature and the speaker dependent voice recognition name feature are extracted from the voice input caller name. Performing a function of obtaining an outgoing telephone number of the service user;
The voice recognition server,
Storing the extracted speaker authentication voice feature and speaker dependent voice recognition name feature and the incoming call number by using the obtained outgoing call number as a key; Automatic telephone connection service system, characterized in that. The method of claim 19,
The voice feature extraction device,
Receives 'voice for caller's name' and 'called phone number' from the service user who requested the construction of personal phone book through the telephone network. Extracting the name text and further obtaining a caller's outgoing telephone number of the service user;
The voice recognition server,
Using the obtained outgoing call number as a key, the extracted speaker authentication voice feature, speaker dependent voice recognition name feature, and caller name text and the incoming call phone number are stored to build the personal phone book database. An automatic telephone connection service system, characterized in that it further performs a function. The method of claim 26,
The voice recognition server,
Before performing the construction of the personal phone book database by storing the extracted speaker authentication voice feature, speaker dependent voice recognition statement feature, and caller name text and the incoming call phone number, the called party name text and the called phone number Delivering to the voice feature extraction device;
The voice feature extraction device,
And reproducing the received caller name text and the called phone number by voice and providing the service user with the service user to confirm the accuracy of the data to be stored. The method of claim 19,
The personal phone book database,
And the service user accesses a server having a voice recognition feature extraction function through the Internet and directly registers the service.

Images