JPH042252A - Voice recognition data storage system - Google Patents

Abstract

PURPOSE:To relieve the load of a troublesome job such as listening and memo by a recipient by using a voice recognition means so as to process information inputted in voice from a communication line into a data and filing the data. CONSTITUTION:A voice recognition section 4 connects to a line connection section 1 and a guide voice is inputted to the voice recognition section 4 by operating a start switch of the voice recognition section d when the guidance voice is obtained. The voice recognition section 4 recognizes it as the voice and outputs the result of recognition as a character string data in, e.g. the JIS code form. The character string data is fed to a specific data extraction section 5, in which only a telephone number data is extracted. The extracted telephone number is sent to a storage section 6 comprising a RAM, in which the telephone number is stored, then information in voice from a communication line is filed as a data. Thus, a troublesome job load such as listening and memo by a recipient is relieved.

Description

[Detailed description of the invention] (a) Industrial application field Speech recognition data storage that enables voice recognition of information obtained by voice via a voice communication line, accumulation of recognized information, and creation of a database Regarding the system.

(b) Conventional technology In recent years, research on speech recognition technology has been actively conducted, with the aim of putting into practical use speech recognition devices with high recognition rates.

However, since there are various uncertainties in the speech uttered by humans, there are still many issues to be solved in order to realize a speech recognition device with a 100% recognition rate, that is, no false recognition. is the current situation.

Therefore, at present, speech recognition devices have been developed that are capable of recognizing speech with practical recognition rates for numeric speech and a limited number of syllable-based speech. The future challenge is to develop practical applications for voice response devices using this technology.

On the other hand, a typical voice response device is an information guidance system using a telephone line, and the application of voice recognition technology in this field is attracting attention.

For example, in conventional telephone directory assistance systems, the recipient of the inquiry listens to the voice of the telephone directory.
It was necessary to record this in a memo pad and call the recorded number again.

For this reason, there is a possibility that a person may mishear or record the information number incorrectly, and there is a fear that this may lead to calling the wrong party.

(c) Problems to be Solved by the Invention The present invention provides a voice recognition data storage system that can reliably store voice information obtained by the voice response device as described above. The purpose is to reduce the operational burden.

(d) Means for Solving the Problems The voice recognition data storage system of the present invention includes a means for connecting to a voice communication line, a voice recognition means for voice recognition and detection of information obtained by voice via the connecting means; information storage means for accumulating information obtained as a recognition result by the voice recognition means, the information voice obtained from the communication line via the connection means is voice recognized by the voice recognition means, and the information is The information is extracted and stored in a storage means.

(E) Function According to the voice recognition data storage system of the present invention, the information voice input from the communication line can be voice recognized by the voice recognition means and the recognition results can be stored, so that the information obtained by voice from the communication line can be converted into data. File ringing is possible.

Therefore, if this system is adopted for telephone number guidance from a telephone line, it is possible to construct a telephone number database, and furthermore, it becomes possible to perform automatic dialing as necessary.

(F) Embodiment FIG. 1 is a block diagram showing the configuration of one embodiment of the speech recognition data storage system of the present invention.

The line connection part (1) in the figure normally connects the telephone set (2) and the telephone line (3), and is connected to the network control circuit (NCU).
). Now, if you dial 104 from this telephone (2), tell the operator at the other end your desired name and address, and request telephone number assistance, NTT's voice response system will call you the desired telephone number with a synthesized voice. You will be guided. For example, the response "The telephone number for inquiries is 06-123-4567" is mechanically repeated three times.

Furthermore, the voice recognition section (4) in the figure is connected to the line connection section (
1), and when the above-mentioned guidance voice is obtained, the above-mentioned guidance voice is input to the voice recognition unit (4) by operating the activation switch of this voice recognition unit (4). In this voice recognition section (4), when the guidance voice is input, it is voice recognized and the recognition result is outputted as, for example, character string data in JIS code format.

In this way, the character string data output from the voice recognition section (4) is sent to the specific data extraction section (5), where only the telephone number data is extracted.

That is, only the numeric code needs to be collated and extracted, and in the above voice guidance example, the telephone number part molecule 061234567J is extracted. The extracted telephone number is sent to the storage section (6) and is accumulated and stored in the storage section (6) consisting of RAM11.

With the above configuration, the number information obtained from the voice response device, in this case, the telephone directory assistance system, can be mechanically stored in the storage section (6), and it is possible to save the inquirer's listening and memorandum work.

The telephone number information thus obtained can be displayed, for example, on a liquid crystal display (not shown), and the person making the inquiry can make a call to the desired party based on this display. The system employs an automatic dialing mechanism that automatically dials the telephone number stored in the storage section (6). This automatic dialing mechanism is similar to the speed dialing function of a normal telephone, in which the line connection unit (1) sends a dial signal to the telephone line (3) by the operation of the NCU.

Here, further regarding the specific data extraction unit (5),
This will be explained in detail with reference to FIG.

The character string that is the recognition result input to the specific data extraction unit (5) is extracted by the number extraction circuit (501) in the figure.
" is extracted and stored in the first buffer (502).Furthermore, the same processing is performed for the second and third voiced directory information, and the number extraction circuit (501) corresponding to the second voiced number is extracted and stored in the first buffer (502). ) is sent to the second buffer (503)
Then, the output of the number extraction circuit (501) corresponding to the third utterance is stored in the third buffer (504). Thereafter, the majority decision circuit (505) compares the contents of the first to third buffers (502) to (504) digit by digit, determines and outputs the telephone number by majority decision. By executing the majority decision in this manner, even if there is recognition due to line noise or the like during voice recognition, this can be corrected and the correct telephone number can be obtained.

In the above-described embodiment of the speech recognition data storage system, the speech recognition unit fil(4) recognizes synthesized speech of a speech response system with stable speech quality, so it is not necessary to use a large-scale computer device. There is no need to use an expensive speech recognition device for non-specific speakers that requires , and it is possible to use a relatively inexpensive speech recognition device for specific speakers. However, the present invention is not limited to this, and if a speech recognition device for unspecified and unspecified speakers with a practical speech recognition rate can be adopted, the speech uttered by the voice guidance operator himself can be recognized. This information can be stored and used.

Although the storage unit (6) in the above-described embodiment functions as a buffer memory, the storage capacity of this storage unit can be increased to store a plurality of telephone numbers, names, and addresses in association with each other, and to It is also possible to build a database of numbers. In this case, if necessary, a controller for controlling memory read/write, display, and dial functions may be attached, or a control section for controlling each of the above components (1) to (6), such as a microcomputer, may be used. Just do it. As a data entry means for constructing such a large number of telephone number databases, it is possible to input voice inquiries to NTT using synthesized voices from a voice synthesizer. This makes it possible to create a database of a large number of telephone numbers.

(G) Effects of the Invention According to the present invention, information input by voice from a communication line can be converted into data by voice recognition means and filed, so the complicated work of listening to the recipient and taking notes can be reduced. .

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the present invention, and FIG. 2 is a block diagram showing details of a specific data extraction section in one embodiment of the present invention. (1) is a line connection section, (2) is a telephone device, (3) is a telephone line, (4) is a voice recognition section, (5) is a specific data extraction section,
(6) is a storage section.

Claims (3)

[Claims] (1) A means for connecting to a voice communication line, a voice recognition means for recognizing and detecting information obtained by voice through the connection means, and an information storage for accumulating information obtained as a recognition result by the voice recognition means. and means for extracting information by voice recognition of the information voice obtained from the communication line via the connection means by the voice recognition means, and storing the information in the storage means. Speech recognition data storage system. (2) It is characterized by comprising specific information extraction means for extracting specific information from the information of the voice recognized by the speech recognition means, and storing the specific information extracted by the specific information extraction means in the storage means. The voice recognition data storage system according to claim 1. (3) a means for connecting to a telephone line; a voice recognition means for recognizing the number guidance voice of the voice guidance system via the connecting means;
Telephone information extraction means for extracting number information as a result of recognizing a telephone number voice from a recognition result by the voice recognition means; an information storage means for accumulating the number information extracted by the telephone information extraction means; A telephone number guide recognition system comprising automatic calling means for sending a dialing signal to the telephone line according to accumulated number information.