JPH0255448A - Voice dialing system - Google Patents

Abstract

PURPOSE:To attain prescribed incoming control when a call comes from an unspecified opposite party by controlling the permission of the incoming call or the storage of a message based on the characteristic quantity of a voice of a talker sent from the caller side. CONSTITUTION:A control section 5 sends a prescribed synthesis tone representing the urgency of a voice input of a caller at the incoming state to the caller. Then a characteristic extraction section 9 of the called party extracts a characteristic of the voice sent from the caller and collates the characteristic quantity with a talker collation pattern 10 registered in advance. As a result of collation, if the pattern is not correspondent with the talker collation pattern, the control section 5 discriminates the caller to be an unspecified talker and informs a synthesis tone representing the urgency of the message input desired to be informed to, e.g., the caller. Thus, when the caller sends a message, the control section 5 stores it to a storage section 13 and the result is outputted as a synthesis tone at a desired time.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a voice link system used in a telephone, and more particularly, to a voice link system used in a telephone.

This invention relates to a voice link system that uses voice recognition technology to make a call and control incoming calls when used on the receiving side.

[Conventional technology]

Conventionally, as disclosed in Japanese Patent Laid-Open No. 61-1.441.57, a voice system that recognizes a name entered by voice, reads out the Atayal number corresponding to the voice recognition result, and makes an Atayal call. It is known as the dial link system.

This type of voice link system is applied to telephones equipped with a transmitter and a handset, and when this telephone is used as a calling party, the name of the other party is entered by voice from the transmitter, and the telephone number is automatically transmitted to the other party. Since it is possible to make an Atayal call, there is no need to look up the Atayal number, perform dial operations, etc., and this has the advantage of improving operability.

[Problem to be solved by the invention]

In this way, in the conventional voice tire link system, convenience such as improved operability is obtained when the telephone is used as a calling party, but no consideration is given to the use of this telephone as a receiving party6. For example, if you receive a call while you are sleeping, the linker will operate unless you respond to the other party's call or the other party interrupts the call, and your sleep will be interrupted. If the other party's call is unnecessary, you may want the call to be automatically disconnected. In addition, if the call requires incoming calls, it may be desirable to control the incoming call in such a way that the details of the incoming call can be known later rather than at the time, but with conventional voice dialing systems, such incoming control is not possible. It wasn't.

SUMMARY OF THE INVENTION An object of the present invention is to provide a voice dialing system that allows predetermined call control to be performed when a telephone is used as a called party and receives a call from an unspecified party.

[Means to solve the problem]

In order to achieve the above object, the voice dialing method of the present invention sends a predetermined synthesized sound to the calling party when a call is received, and by receiving this synthesized sound, the voice dialing system of the present invention is configured to It compares the registered standard pattern for speaker verification, and if the feature amount of the speaker's voice does not correspond to the pattern for speaker verification, the message from the speaker is stored, and the message is sent when desired by the receiving party. The feature is that it is output as a synthesized sound.

[Effect]

In the above-described configuration, when a telephone is used on the receiving side, a predetermined synthesized sound is sent to the calling side to prompt the caller to input voice when the call arrives. The transmitting side receives this synthesized sound, outputs the voice, and sends it to the chain side. The receiving side extracts the features of the voice sent from the calling side, and compares the features with pre-registered patterns for speaker verification. As a result of the verification, if the pattern does not match the pattern for speaker verification, the receiving side determines that the calling party is an unspecified speaker, and notifies the caller of a synthesized voice prompting the caller to input the desired message, for example. As a result, when a caller sends a message, the @ chain stores the message. The message stored in this way is output as a synthesized voice when desired by the called party, so that the called party can later learn about messages from unspecified speakers without having to directly receive them over the phone. ,
You can select the speaker.

〔Example〕

Hereinafter, one embodiment of the present invention will be described based on the drawings.

FIG. 1 is a configuration diagram of a telephone to which a voice dialing system according to the present invention is applied, and this telephone includes a transmitter 1,
Handset 2 and call number I? 83, a line control unit 4, a control unit 5 that performs overall control regarding voice dialing,
An operation unit 6 such as a keyboard, a timer 7, and a transmission circuit 8
, a feature extraction section that extracts the feature amount of the voice input from the transmitter 1 or the voice input from the telephone call circuit 3 via the line control section 4 , and a standard in which the voice feature amount is registered in advance as a standard pattern. A pattern storage section 10, a number storage section 11 in which the dial number corresponding to the name input by voice is registered in advance, and a number storage section 11 in which the dial number corresponding to the name input by voice is registered in advance, and the feature amount extracted by the feature extraction section 9 and the standard pattern storage section 10 in which the dial number is registered in advance. a pattern matching unit 12 that performs pattern matching with the feature quantity, a synthesis standard pattern storage unit 13 in which standard patterns for speech synthesis are registered in advance;
and a speech synthesis section 14 that outputs the output to the speech synthesizer.

The operation unit 6 includes a registration button 20 for starting the process of registering the standard pattern etc. to the standard pattern storage unit 10 etc.
, a speaker verification button 21 for automatically performing predetermined incoming call control using a telephone on the incoming call side, and a message button 21 for playing back a message from an unspecified speaker stored in the incoming call control with synthesized voice. A reproduction button 22 and a dial 23 for inputting a dial number etc. are provided.

The standard pattern storage unit 10 also contains a recognition standard pattern unit 15 that is necessary when the telephone is used as a calling side, and a pattern that is necessary for performing predetermined call reception control when the telephone is used as an @chain side. A standard pattern section 16 for person verification is provided.

How to perform voice dialing in a telephone set as described above will be explained below with reference to FIGS. 2 and 3. Figure 2 is a flowchart showing the processing flow for inputting the name of the other party by voice and dialing the number corresponding to this name when using a telephone as the calling party. 12 is a flowchart illustrating a process flow for performing predetermined call reception control when the user is present at the person's side.

When a telephone is used as a caller and a dial call is made by voice input, the feature quantity is calculated for the name of the other party (for example, the name of the other party) inputted by voice from the handset 1. It is necessary to register the name in advance in the recognition standard pattern section 15 as a recognition standard pattern section, and also to register the dial number corresponding to the name in the number storage section 11 in advance.

Steps S1 to S6 in FIG. 2 show this registration process. In the registration process, the caller first presses the registration button 20 on the operation unit 6 (Step S1), and inputs the name of the other party by voice from the handset 1 (Step S2>. The input voice is used for feature extraction. Features are extracted in the section 9, and the feature amounts are registered in the standard pattern for recognition section 15 as a standard pattern for recognition, and further registered in the standard pattern for synthesis storage section 13 as a standard pattern for speech synthesis (
Step S3). Next, the speech synthesis section 14 synthesizes speech based on the standard pattern registered in the synthesis standard pattern storage section 13 and outputs the synthesized speech to the receiver 2 . As a result, the caller confirms his/her own voice input from the transmitter 1 using the receiver 2 (step S4). After confirming the synthesized voice, the caller inputs the dial number of the other party from the numeric keypad 23 of the operation unit 6 (step S5). The input dial number is registered in the number storage unit 11 in association with the standard recognition pattern registered in the standard recognition pattern unit 10 in step S3 (step S6).

After the standard pattern for recognition and the dial number are registered in advance in the standard pattern for recognition section 159 and the number storage section 11 in this way, steps S7 to S1. The dialing process shown in O is performed. That is, the caller inputs the name of the other party by voice using the transmitter 1 (step S7). The input voice is subjected to feature extraction in the feature extraction unit 9,
The feature amount is determined in step S in the pattern matching section 12.
The pattern is compared with a standard pattern for recognition registered in advance in the standard pattern for recognition section 15 in the processes from 1 to S6 (step S8). When a recognition standard pattern with the highest similarity to the input voice feature quantity is detected as a recognition result by pattern matching, the dial number of the other party corresponding to this recognition standard pattern is read out from the number storage unit 11. The signal is sent to the transmitting circuit 8 (step S9). As a result, the transmitting circuit 8 can output a dial tone signal to the central office line 24 via the line control unit 4 to make a dial call to the other party (step 510).

Next, when performing predetermined call control using a telephone as the receiving side, the caller's name (for example, the caller's name or It is necessary to register in advance the feature amount for the standard pattern for speaker verification in the standard pattern for speaker verification 16 as a standard pattern for speaker verification.

Steps TI to 'I'4 in FIG. 3 show this registration process. The registration process is started, for example, by pressing the registration button 20 on the operation unit 6 on the called party when the caller makes a dial call and is in a conversation state with either the called party or the calling party (step 1-1). . When the registration button 20 is pressed during a call, the control unit 5 controls the voice synthesis unit 1.
4, the voice synthesizer 14 outputs a synthesized voice prompting the caller to input the name of the caller, and the synthesized sound is sent to the receiver of the caller via the telephone call circuit 39 line control unit 41 central office line 24 (
Step T2). As a result, on the calling side, the caller inputs his or her name by voice from the telephone (step T3). The caller's voice is transmitted via the central office line 24 to the line control unit 4 on the receiving side.
1 is received by the communication circuit 3, the feature extraction section extracts the feature, and the feature amount is registered in the standard pattern for speaker verification 16 as a standard pattern for speaker verification (step T4).
.

In this way, after the standard pattern for speaker verification is registered in advance in the standard pattern for speaker verification section 16, the incoming call control processing shown in steps T5 to T1.7 can be performed. That is, at the time of arrival t8IlII, the speaker verification button 21 of the operation section 6 is kept depressed in order to automatically perform predetermined call reception control (step T5). When the caller receives a call from the caller while the speaker verification button 21 is pressed, the voice synthesizer 14 outputs a synthesized sound prompting the caller to input the name of the caller, and the call circuit 32 line control unit 41 and is sent to the receiver of the calling party via the central office line 24 (step T6). As a result, on the calling side, the caller enters his or her name manually through the telephone (step 1-7).

The caller's voice is received by the line control unit 41 on the receiving side via the office line 24, and the feature extraction unit extracts the features, and the characteristic quantities are extracted in the pattern matching unit 12 in steps T1 to T1. Standard pattern part 1 for speaker verification in the process of 4.
The pattern is matched with a speaker matching pattern registered in advance in step T8. As a result of pattern matching, if the caller's voice falls under one of the pre-registered speaker matching patterns, the caller is determined to be a specific speaker registered on the receiving side, and the control unit In step T 9 , the receiver performs an operation to notify that there is an incoming call (for example, operates a linker) and calls the called party (step T 9 ).

On the other hand, if the result of the pattern matching in step T8 is that the caller's B-voice does not correspond to any of the pre-registered speaker matching patterns, the caller is a non-registered user on the receiving side. It is determined that the speaker is a specific speaker, and the control unit 5 performs step T10 or step '
Perform the processing of steps I"11 to 1"17.In this case, proceed to step "I"10 or step T"
Depending on the request of the called party, the controller 5 is instructed in advance whether to proceed with the processing of steps 11 to TI-7 using a block diagram or the like.

When proceeding to step 110, the control section 5 rejects the incoming call and ends the incoming call control. As a result, the linker does not operate on the called side, so the called side does not have to receive calls from unspecified speakers.

However, on the receiving side, even if the call is from an unspecified speaker, the content of the call may be necessary depending on the situation.
There may be times when you want to memorize a message from an unspecified speaker for later reference. In this case, it is necessary to program the control unit 5 on the receiving side to proceed to steps Tll to T17 as described above. At this time, if it is determined that the caller is not the registered speaker, the receiving side accesses the voice synthesis unit 14 and
A synthesized sound prompting the input of the message to be conveyed is output to the calling side via the telephone call circuit 39, line control unit 41, and central office line 24 (
Step T11). Further, at the same time as outputting the synthesized sound to the transmitting side, the control section 5 activates the timer 7, and the timer 7 counts a predetermined period from the output of the synthesized sound (step T12).

On the calling side, the message to be conveyed is sent to the called side using the synthesized voice as a signal (step T13). On the receiving side, timer 7
The central office line 24 and the line control unit 41 within a predetermined period determined by
The feature extraction section 9 extracts the features of the message from the calling side inputted through the telephone call circuit 3, and the feature quantity is stored in the standard pattern storage section 13 for synthesis.Step T1.4>
, then terminates the incoming call operation (step T')
5). If you want to know the message stored in the standard pattern storage unit 13 for synthesis later, press the message playback button 22 on the operation unit 6 (step T”16).The stored message can then be synthesized into a voice by the voice synthesis unit 14. is output from the receiver 2 (step T17).
), the receiving party can receive messages from unspecified speakers. If the processes in steps Tll to T17 are selected in this way, the called party does not need to directly receive a call from an unspecified speaker and make a call, but can also receive a message from an unspecified speaker later. Therefore, it becomes possible to select the speaker and effectively prevent incoming prank calls.

〔Effect of the invention〕

As explained above, according to the present invention, when a call is received, a predetermined synthesized sound is sent to the calling side, and by receiving this synthesized sound, the characteristic amount of the speaker's voice sent from the calling side is registered in advance. The message from the speaker is stored when the standard pattern for speaker verification does not apply, and this message is output as a synthesized voice when desired by the receiving party. This eliminates the need to directly receive a call and make a call, and it also becomes possible to control incoming calls so that messages from unspecified callers can be learned later.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a telephone to which the voice dialing method according to the present invention is applied, and FIG. 2 is a flowchart showing the processing flow when the telephone shown in FIG. 1 is used on the calling side to make a call. FIG. 3 is a flowchart 1 showing a processing flow when predetermined call reception control is performed using the telephone set shown in FIG. 1 on the receiving side. 1... Transmitter, 2... Receiver, 3... Telephone circuit,
4... Line control section, 5... Control section, 6... Operation section, 7... Timer, 8... Transmission circuit, 9... Characteristic extraction section, 10... Standard pattern memory 11... Number storage unit, 2... Pattern matching unit, 3... Standard pattern storage unit for synthesis, 4... Speech synthesis unit, 5... Standard pattern unit for recognition, 6...・Standard pattern section for speaker verification, 0...Register button, 21...Button for speaker verification, 2
...Message playback button, 23...Numeric keypad 4.
...central line

Claims (1)

[Claims] When a call arrives, a predetermined synthesized voice is sent to the caller, and when the synthesized voice is received, the characteristics of the speaker's voice sent from the caller are compared with a pre-registered standard pattern for speaker verification, and the speaker is identified. A voice diagram characterized in that a message from the speaker is stored when the feature amount of the voice does not correspond to a pattern for speaker verification, and the message is output as a synthesized voice when desired by the receiving party. Ring method.