KR20030000314A - Telephone Message Memo System Using Automatic Speech Recognition - Google Patents

Abstract

PURPOSE: A system for taking notes of call contents using voice recognition is provided to recognize voice for the contents of a call transceived through a telephone set, and to output the contents with a printer or on video, so that taking a memo is not necessary. CONSTITUTION: A voice communication terminal(10) has a voice receiving part and a voice transmitting part. A voice recognition unit(20) recognizes voice signals from the voice transmitting part or voice signals from the voice receiving part, to convert recognized voice recognition data into character data. An output unit(30) outputs the character data acquired by the voice recognition unit(20). The output unit(30) includes a printer(30A) or a monitor(30B). The voice recognition by the voice recognition unit(20) is provided through Internet communication of a voice recognition server(50) having a voice recognition engine of the voice recognition unit(20). And a recognition switch(40) selectively switches on/off for the operation of the voice recognition by the voice recognition unit(20).

Description

Phone Message Memo System Using Automatic Speech Recognition}

(Technology)

The present invention relates to a phone call contents memo system using voice recognition, and more particularly, to recognize phone call contents transmitted and received through a telephone and output them to a printer or an image, thereby eliminating the need for a phone call memo using a pen. The present invention relates to a phone call memo system using voice recognition.

(Background)

The exchange of information between humans and machines is considerably improved due to the development of input means such as various switches, keyboards, mice, and touch screens, but it is still relatively inconvenient compared to the dialogue between humans.

Therefore, MMI (Man Machine Interface) for the exchange of information between humans and machines has been important, and since the mid-1970s, the research on voice recognition has been actively conducted, so MMI's technology goes beyond voice recognition. They have come to understand the meaning of the voice and perform the necessary functions.

In general, speech recognition refers to a technique of extracting a feature from a given speech signal and applying a pattern recognition algorithm to it to trace back which phoneme is a speech signal generated by the speaker.

Speech recognition technology can be classified into various types according to the standard, and can be classified into speaker dependent system, speaker independent system, and speaker adaptation system according to the speaker to be recognized.

The speaker dependent system is a system for recognizing the voice of a specific speaker and is typically a voice dial system used in a mobile phone. In a speaker-dependent system, a technique of comparing a pattern of an input voice with a pattern of a stored voice is generally used when storing and registering a user's voice and performing real recognition before using the system.

The speaker independent system is for recognizing the voice of unspecified majority speakers, and there is no hassle for the user to register the voice prior to the operation of the system as in the speaker dependent system. It involves the process of performing recognition using the trained model. Therefore, the characteristic characteristic of the speaker disappears and the characteristic common to each speaker is highlighted.

The speaker adaptation system is a system in which the speaker independence system is constructed while the speaker independence system is modified to suit the user's voice in actual use.

Speech recognition technology is also divided into small-capacity systems and large-capacity systems according to the number of words to be recognized (typically small-capacity systems have hundreds of words or less, and large-capacity systems have thousands of hundreds of thousands of recognized words. Depending on the type of pronunciation, each word is classified into two or more words, and it is also classified into an isolated language recognition system that assumes that there is a silent section of sufficient length between each word, and a continuous word recognition system that performs speech recognition on a sentence-by-sentence basis. .

Examples of voice recognition technologies currently applied include voice dialing mobile phones, voice-driven personal computers (voice interface technology with computers), voice recognition cars, stock information, telephone numbers, local information, and other information guide services, voice recognition unmanned exchanges. , Voice control of robots, automatic interpretation system, manuscript writing system that forms voice into text files, etc., and the spread of voice recognition technology is expected to bring great change to people's cultural life.

In using a telephone, which is one of the necessities of daily necessities used in daily life, it is often necessary to take note of necessary matters in the conversation with the other party. Such phone call memo is more essential when receiving a customer's order (delivery order) at various offices, including restaurants. However, until now, the telephone call memo has no other method than the method of summarizing the call with a pen on a paper (memo paper) or order form provided by the telephone.

SUMMARY OF THE INVENTION An object of the present invention is to eliminate the inconvenience associated with the above-mentioned phone call contents memo by allowing the phone call contents that need to be memo to be automatically memo by printing or image, without the handwriting of the caller. It is.

1 is an exemplary system configuration diagram of a phone call memo system using voice recognition according to the present invention;

2 is an exemplary block diagram of a system in accordance with the present invention.

<Description of Symbols for Main Parts of Drawings>

1: system 10 of the present invention: voice communication terminal

11: telephone receiver 12: receiver

20: voice recognition means 30: output means

30A: Printer 30B: Monitor

40: Recognition switch 41: Print execution button

42: Delete button 43: Main switch

50: voice recognition server

According to the present invention, there is provided a telephone call content memo system using voice recognition comprising a voice communication terminal, voice recognition means and output means.

Specifically, according to the present invention, a general voice communication terminal having a transmitter and a receiver, at least one of the voice signals from the transmitter and the receiver, recognizes the voice recognition data, and the character data corresponding to the recognized voice recognition data. There is provided a phone call content memo system using voice recognition, comprising voice recognition means for converting the voice recognition means, and an output means for outputting the text data obtained by the voice recognition means.

As the output means, at least one of a printer and a monitor may be used. Preferably, both a printer and a monitor are used. If necessary, a voice synthesizing means and a speaker may be added to output the voice recognition content as a synthesized voice. .

Preferably, the telephone call contents memo system of the present invention includes a recognition switch capable of freely selecting on / off of voice recognition execution of the voice recognition means, so that only the voice that needs to be memorized according to the caller's operation of the recognition switch is selected. It can be output through the output means.

As the speech recognition means, various speech recognition technologies developed so far can be applied to suit the purpose of the present invention. In addition, the voice recognition means is based on the integral inclusion in the system of the present invention, but the voice recognition engine, which is the core of the voice recognition means may be provided in the voice recognition server on the Internet to support text recognition through Internet communication have.

Hereinafter, with reference to the accompanying drawings will be described in detail a phone call content memo system using voice recognition according to the present invention. The following embodiments are merely illustrative of the telephone call content memo system using voice recognition according to the present invention, and are not intended to limit the scope of the present invention.

1 is an exemplary system configuration diagram of a phone call memo system using voice recognition according to the present invention, and FIG. 2 is an exemplary block diagram of a system according to the present invention.

As shown, the telephone call contents memo system 1 using the voice recognition according to the present invention basically includes a voice communication terminal 10, a voice recognition means 20 and an output means 30.

The voice communication terminal 10 is, as is well known, a receiver having a microphone 11 for converting a user's voice into an electrical signal and a speaker for converting a voice signal transmitted through a telephone line into voice. Means a conventional wired and wireless telephone 10 including (12), and is preferably a landline telephone 10 most commonly used in business offices and homes.

According to a feature of the invention, at least one of the voice signals of the talker 11 and the receiver 12 of the telephone 10 is connected to the voice recognition means 20 electrically connected to the telephone 10. Is converted into character data that can be recognized and output. The conversion of the voice signal into the text data by the voice recognition means 20 can be realized by applying a known voice recognition means in accordance with the present invention.

For example, the voice signal formed by the microphone of the transmitter 11 and the voice signal provided to the speaker of the receiver 12 are provided to the voice recognition engine 21 of the voice recognition means 20 so that a character corresponding to the voice is provided. It can be output as. The voice recognition engine 21 includes an analog-digital converter 22 for converting an analog voice signal into a digital voice signal, a feature extractor 23 for extracting the feature from the digital voice signal, and a pattern from the extracted feature. The voice pattern recognition unit 24 to recognize the voice, the voice pattern memory unit 25 storing the reference model related to the voice recognition, the information of the voice pattern recognition unit 24 and the information of the voice pattern memory unit 25 are compared. And a text data converter 27 for converting the recognized voice recognition data into text data.

The voice recognized by the telephone call content memo system 1 of the present invention is the voice of the user of the present invention (eg, the voice of a sales office caller) input into the microphone of the caller 11, and the other party output from the speaker of the receiver 12. May be the voice of a customer (eg, the voice of a customer ordering a product at a sales office), or both.

Since the contents requiring the memo are generally present in the contents of the call on both sides of the telephone call, it is most preferable to recognize and memoize the voices of both parties, but it has various voices, dialects and language habits. In order to reliably recognize a large number of other party's (customer) voices, a large amount of voice recognition means 20 is required, so that only the voice of the user of the present invention (for example, the voice of a sales office caller) is required for the simplification of the system 1. It may also be desirable to configure it to recognize.

In the case of recognizing only the voice of the latter office caller, the voice recognition rate is maintained high even when a relatively small voice recognition means 20 is used by memorizing and learning the voice of the office caller while employing a speaker dependent recognition method. It is preferable in that it can be. In this case, the voice of the other party (customer) is not recognized. Therefore, the sales caller shall repeat the contents of the call of the other party. Since the caller of the sales office receiving the actual product order generally checks the customer order contents in a manner such as "I have ordered one combination pizza", the system 1 of the present invention is checked by the counterpart ( Even if the customer's voice is constructed not to recognize, there is no big problem in performing essential functions.

The printable text data obtained by the voice recognition means 20 is output by the output means 30. The output means 30 used in the present invention is not particularly limited as long as it can replace a memo that has been performed using an existing pen and paper.

Specifically, the output means 30 may use a printer 30A and / or a monitor 30B for outputting text data to be recognized. Preferably, both the printer 30A and the monitor 30B are included. do. In addition, a voice synthesizer may be added to the voice recognition means 20 and a speaker may be mounted so that the memo contents are output as a synthesized voice.

The system 1 of the present invention may be constructed in a structure in which hardware and software of the voice recognition means 20 are attached to a widely used personal computer, and a printer 30A and a monitor 30B are connected to the personal computer. The voice recognition means 20, a predetermined printer 30A, and a monitor 30B such as a liquid crystal screen may be integrally built into the main body of the telephone 10.

In FIG. 1, not only a personal computer and a printer 30A and a monitor 30B connected thereto but also a printer 30A and a monitor 30B integrally mounted in the telephone 10 are shown together to show both of these configurations. It was.

In general, whether you're in a business office or at home, the notes you take during a call aren't the entire call, but most of them are important. Therefore, the system 1 of the present invention includes a recognition switch 40 which can freely selectively turn on / off the voice recognition means of the voice recognition means 20 in accordance with the on / off operation of the recognition switch 40. It is preferable that only the voice which needs to be memorized is output to the output means 30 selectively.

The recognition switch 40 may be electrically installed at any point between the talker 11 and the receiver 12 and the voice recognition means 20 of the telephone 10 to control the voice recognition, and the recognition switch 40 ) Is preferably a pushbutton type switch with other function keys of the phone 10 so as to be operated by lightly pressing with a finger while receiving a call.

In addition to the recognition switch 40, the telephone 10 executes the print execution button 41 and the recognition switch 40 for causing the printer 30A to perform printing, but the voice recognition is executed. If the useless and need to re-memo (for example, cancel the first order in the call process and place another order) convenient system of the present invention, such as the delete button 42 for deleting the contents already recognized voice recognition Various function buttons can be added for use. Reference numeral 43 denotes a main switch for completely stopping the operation of the voice recognition means 20.

In the above, the example in which the voice recognition engine 21, which is a core component of the voice recognition means 20, is physically provided in the system 1 of the present invention has been described, but the voice recognition engine 21 is the system of the present invention. It is also possible to remotely support text recognition through internet communication with the voice recognition server 50 having the voice recognition engine 21 without being integrally included in (1). In this case, when the caller operates the recognition switch 40, the corresponding voice signal is transmitted to the voice recognition server 50 through the Internet to perform voice recognition, and the text data formed by the voice recognition is transmitted to the caller again and output means ( The memo is executed in the manner outputted by 30).

According to the outsourcing of the voice recognition engine 21, various callers provide high quality voice recognition through the Internet from the voice recognition server 50 equipped with the voice recognition engine 21 having a large capacity, high precision, and high recognition rate. Since it can be received, it is preferable in that it can implement high quality voice recognition with a low system construction cost.

In the telephone call contents memo system 1 using the voice recognition of the present invention described above, the specific configuration of the voice recognition means 20, various software for realizing the same, and software for voice recognition are known in the art. It can be appropriately configured in accordance with the features of the present invention by the techniques described above, and the detailed description thereof will be omitted since these configurations themselves are not features of the present invention.

In addition, in the above description of the present invention, the system of the present invention has been described with reference to an example of being used in an office for receiving an order. However, the present invention is applied to the call content memo in a general home as well as the call content memo in an office. Naturally, it is also possible to apply various well-known OA programs to output the voice recognition memo content in a desired format.

According to the telephone call contents memo system using the voice recognition according to the present invention described above, the call contents that need to be memorized during the telephone call are automatically memorized by printing or an image through the voice recognition. It is not necessary to use a pen and a memo pad for the memo, and therefore, it is expected that the present invention will be more convenient in daily life and work activities related to the use of the phone and the use of the phone.

Claims (4)

A general voice communication terminal having a transmitter and a receiver, voice recognition means for recognizing at least one voice signal among the voice signals from the transmitter and the receiver and converting the recognized voice recognition data into corresponding text data; and And an output means for recognizably outputting the text data obtained by the voice recognition means. The telephone call contents memo system according to claim 1, wherein the output means comprises at least one of a printer and a monitor. The telephone call contents using voice recognition according to claim 1, wherein the voice recognition means by the voice recognition means is provided by a voice recognition server having a voice recognition engine of the voice recognition means through internet communication. Memo system. The apparatus of any one of claims 1 to 3, further comprising a recognition switch for selectively turning on / off the voice recognition performance of the voice recognition means so that only the voice that needs to be memorized in the contents of the call is recognized. Telephone call contents memo system using voice recognition.