JP2000067064A - Interaction recording system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an interaction recording system for accepting complaint or the like with a voice, and for transmitting this in a configuration to be used for the following processing to a processor. SOLUTION: This system is provided with a recording device 200 for recording interaction voice data and an information processor 100 for generating an identifier for identifying a specific part for voice data to be recorded, and allowing the recording device 200 to record this identifier. The information processor 100 accepts a request for the generation of the identifier and generates the identifier for the voice data to be recorded in the recording device 200, and the identifier is recorded in the recording device 200 so as to be made to correspond to the voice data to be recorded. Voice data 212 and identifier data 211 are recorded in the recording device 200, and text data 214 obtained by voice-recognizing the voice data by a voice recognizing part 330 are recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for recording the contents of a dialogue so that it can be used later, and is particularly preferably used when receiving voices of complaints, inquiries, reports of malfunctions, reports of accidents, and the like. The present invention relates to a dialog recording system and a reception processing system.

[0002]

2. Description of the Related Art Generally, inquiries, complaints and the like concerning products are made via telephone. For example, if a failure occurs in a computer system or information communication system,
When a customer of the system receives a phone call reporting a failure,
The maintenance technician takes notes while listening to the equipment to be delivered, the status and contents of the failure, and accepts the failure and confirms the contents of the report. Then, when the acceptance of the failure report is completed, one or more maintenance personnel go to the customer site and perform processing based on the received report content.

However, in recent years, from the viewpoint of efficient staffing, labor saving, and responding to various inquiries,
From the viewpoint of the necessity of 24-hour response, a reception center is provided for intensively covering a wide service area, and is receiving complaints, inquiries, accident reports, trouble reports, and the like. For example, there is a tendency that a small number of reception centers cover a wide area, such as one location in East Japan and one location in West Japan.

[0004]

When such a centralized reception center is provided, the reception center arranges reception specialists, receives many kinds of notices and announcements, and creates memos. Is transmitted to the processing staff, and the specific processing is performed by the processing staff while looking at the memo. Such a division of labor system has the advantage that even if reception is available for 24 hours, it can be handled efficiently with a small number of personnel.

[0005] However, the memo is different from the real voice of the declarant and the reporter, and the information is organized and transmitted. On the other hand, since the real voice of the declarant and the reporter is not transmitted to the processing person, part of the information is not transmitted. Is easily lost.

[0006] On the other hand, it is conceivable that a receptionist asks a reporter, a filer, and the like to make a medical examination, and inputs the contents of the dialogue between them to a computer to create a report. However, in this method, while interacting by telephone or the like,
It is necessary to perform key input, and it is not always easy to perform an accurate input, and it is not common.

It is also conceivable that the contents of the dialogue at the time of reception are recorded, and the processing staff listens to this and deals with it. In this case, the receptionist can concentrate on the medical consultation so as to retrieve necessary items from the other party. However, listening to long conversations and picking out what is needed is time consuming and inefficient. In particular, when urgency is required, the problem of this point becomes remarkable.

The present invention has been made to solve such a problem, and can receive complaints, inquiries, reports of failures, reports of accidents, and the like by voice, and can process the received information in a later process. It is an object of the present invention to provide a dialogue recording system capable of transmitting a dialogue to a processor in a form usable for a user and a reception processing system using the system.

[0009]

According to a first aspect of the present invention, there is provided a dialog recording system for recording the contents of at least one of dialogs.
A recording device for recording voice data of the dialogue, and an information processing device for generating an identifier for identifying a specific portion with respect to the voice data to be recorded, and performing a process of causing the recording device to record the information; The processing device receives a request for generation of an identifier for identifying a specific portion of the audio data recorded in the recording device, generates an identifier, and associates the identifier with audio data to be recorded. A dialog recording system is provided which performs a process of recording in the recording device.

According to another aspect of the present invention, there is provided a reception processing system including a telephone device for receiving a call from outside and making a call, and the above-described dialogue recording system.
The telephone device includes a microphone and a speaker for talking, and causes a received audio signal and an audio signal of the microphone to be input to the information processing device, wherein the information processing device receives the input audio signal. Is recorded in the recording device.

In the above-mentioned dialog recording system and reception processing system, the following various aspects can be appropriately adopted singly or in combination.

A) Volume detection means for detecting the rising edge of the voice after the interruption of the voice based on the volume of the input voice, and a request signal for generating an identifier for identifying the voice data input at that time Means for outputting an identifier, wherein the information processing apparatus generates the identifier in response to the identifier generation request signal.

B) further comprising an input device provided with a key for requesting generation of an identifier;
Upon receiving an identifier generation request signal from the input device,
Generating the identifier.

C) An identifier generation request signal from the input device identifies an area recorded in the future in time from the position of the identifier and an identifier generation request signal in the past in time from the position of the identifier. The input device has a function of inputting each of them.

D) The information processing apparatus uses time information at the time of receiving the identifier generation request signal as the identifier.

E) The information processing apparatus uses, as the identifier, an address used when the corresponding audio data is recorded in the recording device when the identifier generation request signal is received.

F) The apparatus further comprises voice recognition means for performing voice recognition on the digitized voice data and converting it into a character string, wherein the information processing apparatus converts the converted character string in correspondence with the original voice data. Recording on a recording device as text data.

G) The character string converted from the voice by the voice recognizing means is examined to determine whether it contains a predetermined keyword. If the character string is included, a means for changing the display mode of a corresponding portion of the character string is further provided. Be prepared.

H) The information processing device records each voice of the speaker of the dialogue in the recording device in different areas.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. In the following example, a complainant, an inquiry, a failure report, a failure reporter, etc., a reporter, a reporter, and the like are representatively assumed to be a customer C, and a description is given assuming one reception center A and one processor B. I do. Of course, the present invention is not limited to this mode.

FIG. 1 shows an outline of an environment to which the voice reception processing system of the present invention is applied. That is, in FIG. 1, first, some abnormality occurs in the system of the customer C, and the person in charge of the customer C (hereinafter, simply referred to as a customer) communicates with the communication network M
A call is made to the reception center A via the Internet, and the receptionist of the reception center A makes an inquiry and records the conversation. Then, at that time, processing is performed by a voice reception processing system using the dialog recording system of the present invention. After that, the processor B who takes measures such as a failure receives the content received by voice,
The necessary items are extracted, and necessary processes are executed for the customer system C based on the extracted items. As a result, measures are taken against various reports and declarations received by voice.

FIG. 2 shows an outline of information transfer in the reception center A. As shown in FIG. 2, the reception center A includes:
A telephone device 500 for constituting a voice reception processing system for conducting a conversation with the customer C by telephone and a conversation recording system are installed. The dialogue recording system is described in detail in FIG.
As shown in FIG. 1, the apparatus includes at least the information processing apparatus 100, the recording apparatus 200, the input apparatus 410, the display apparatus 420 and the like.
The telephone device 500 accepts a telephone call from the customer C, and interacts with the receptionist at the center A. The content of the dialogue is
It is recorded on the recording device 200. At this time, the information processing apparatus gives an identifier to the recorded content. In the present embodiment, the beginning part of the conversation in the customer conversation (the beginning part)
A function of automatically assigning an identifier and a function of assigning an identifier by an input device as needed by a receptionist.
Of course, only one of them may be used. Thereby, the conversation content of the customer is recorded in the recording device 200 as customer voice data together with the identifier. The recorded customer voice data can be sent to the processor via communication means.

FIG. 3 shows an example of the configuration of a hardware system used in the present embodiment. This system includes a telephone device 500, an information processing device 100, a recording device 200, and peripheral devices 310 to 430.

Telephone device 500 includes a line control unit 510 for connecting to a telephone line, and a transmitting / receiving device 520 for transmitting and receiving. In this transmission / reception device 520, the received voice signal of the customer is taken out,
Send the voice signal of the receptionist. The transmission / reception device 520 includes:
A speaker 531 that converts the received voice signal of the customer into an acoustic signal and outputs the signal, and a microphone 532 that inputs the voice of the receiver as a voice signal as an electric signal to the transmitter / receiver are connected. When there is an incoming call, line controller 510 notifies the receiver of the incoming call with a ring tone, and sends a signal to information processing apparatus 100 when the receiver answers the telephone, that is, goes off-hook. The line control unit 510 may be mounted on a private branch exchange or the like.

The information processing apparatus 100 has a central processing unit (CPU) 110 for executing various processes, a memory 120 for storing programs executed by the CPU 110, data used for the programs, and an interface 130. CP
Although not shown, U110 has a built-in memory and the like. Further, the CPU 110 has a calendar function. As a result, a time stamp described later is realized.
Note that a timer may be provided.

Various devices are connected to the information processing apparatus 100 via an interface 130. That is,
Recording device 200, analog / digital converter (A / D)
310, an analog-to-digital converter (A / D) 320,
Voice recognition section 330, volume detection section 340, first identifier 35
0 and the third identifier automatic generation unit 360 are connected in a form incorporated in the information processing apparatus 100. Of course, each of these devices may be an external device. Further, the input device 410, the display device 420, and the communication control device 430 are connected.

The recording device 200 is composed of, for example, a hard disk device. The recording device 200 is provided with a customer voice file 210 for each customer who makes a call, and customer data 290 including the customer's address, telephone number, various attribute data, and the like. The customer voice file 210 includes
Customer voice data 212 which is a digitized voice signal of the customer, receiver voice data 213 which is a digitized voice signal of the receiver, text data 214 obtained by voice recognition, and identifier data to be described later. 211 is recorded.

In the present embodiment and other embodiments described later, the identifier data 211 and the customer voice data 2
12. Recipient data 213 and text data 214
Is stored in the same file. However, the present invention is not limited to this. Some or all of the data may be stored in an independent form without being collected for each customer. For example, the identifier data 211 is stored in the customer data 29
You may make it store by itself like 0. Further, the identifier data 211, the customer voice data 212, the acceptor data 213, and the text data 214 may all be stored independently.

The A / D 310 converts a customer's voice signal into a digital signal. The A / D 320 converts the voice signal of the recipient into a digital signal. A / D310 and A
/ D320 both have an output buffer (not shown),
The digitized audio signal is temporarily held. The A / D 310 and the A / D 320 output a recording request signal for audio data to the information output device 100,
When the acceptance is permitted, the digitized audio signal stored in the output buffer is sent to the information processing device 100. The sent voice data is input to the information processing apparatus 100 via the interface 130, and then recorded in the customer voice data file 210 in the recording device 200. The audio data recorded in the audio data file is read out and converted into an analog signal by a digital-to-analog converter, so that it can be reproduced as audio. Note that the data may be compressed during recording. The A / D 320 includes an amplifier (not shown) for amplifying an input from the microphone 532.

When the telephone line is a digital line and voice is transmitted as digital data, the A / D 31
0 can be omitted. However, the information processing device 1
It is preferable to provide a buffer memory for temporarily holding data to be input to 00.

The voice recognition section 330 reads a digitized voice signal, converts the voice into a corresponding character or character string, and generates text data. In the present embodiment, they are connected as independent devices. Of course, you may make it process by CPU110. In the present embodiment, a speech recognition unit 300 that supports an unspecified speaker is used.

The volume detector 340 has a function of detecting a point where the level sharply increases after the sound is interrupted. That is, the sound volume detection unit 340 detects the level of the voice signal of the customer, determines whether the level is equal to or higher than a predetermined reference value, and determines that the speech is interrupted when the level is lower than the predetermined reference value. When the value exceeds the value, it is determined that the speech has begun, and a cueing signal is output. FIG. 4 shows an example. For example, before time t4, the volume level remains low for a while (silence state), and at time t4, the volume level sharply increases. Volume detector 3
40 detects this state and outputs the head signal. This head signal is input to the third identifier automatic generation unit 360. In the present embodiment, only the voice of the customer is detected, but the volume of the receiver may also be detected.

The first identifier automatic generator 350 receives the output signal of the volume detector 340 and generates the first identifier at that time. This identifier will be described later.
The generated identifiers are stored in the corresponding customer voice data files 210, respectively.

[0034] The third identifier generation unit 360 is provided with the speech recognition unit 3.
It is determined whether the voice text data converted into the text data in 30 includes a preset keyword, and if so, a third identifier is generated.

The input device 410 comprises, for example, a keyboard. Although only a keyboard is shown in FIG. 3, a mouse, a touch panel, and the like may be provided in addition to the keyboard. A key for inputting the second identifier is defined on this keyboard. This key assigns a specific key on the keyboard. For example, a key “>” is defined as a key (hereinafter referred to as a forward key) indicating that attention should be paid to contents recorded in a temporally future area immediately after the key is pressed, and “<"Can be defined as a key (referred to as reverse key) indicating that attention should be paid to the contents recorded in the past area immediately before the key is pressed. Alternatively, a symbol may be displayed on a display screen of a display device described later, and the identifier may be input by clicking the symbol with a mouse or the like. Further, a message may be input before and after the key input described above. In particular, if you add an identifier after the story of the customer, the content can be specified,
By adding a special message, code, or the like representing the content, it is possible to facilitate later retrieval. Note that “>” and “<” indicate attributes of the identifier. It is recorded together with a later-described time stamp, address, and the like for indicating the position of the identified data.

The display device 420 displays the screens necessary for the above-described operations, an inquiry guide screen for the receptionist, a customer data display screen, a voice recognition result data, and a screen for confirming the recognition result. It is used for displaying various screens such as a screen for displaying the text data obtained.

The communication control device 430 is a device for performing data communication and the like by the information processing device. The communication control device 430, the communication network M, and the processor's communication control device 93
0 and the processor's information processing device 800 can be used to perform communication for transferring a telephone customer voice data file or the like to the processor. Also, data from the processor can be received. When the processor's communication control device 930 is a mobile phone and the information processing device 800 is a mobile terminal, the processor can obtain necessary information on site.

Next, the above-mentioned identifier will be described.
The identifier is for making it easy to obtain the target information when referring to the contents of the dialog later. Therefore, first, an identifier that functions to indicate where information is located, second, an identifier that functions to indicate that there is a matter to be particularly referred to, and third, a specific content. There are identifiers that serve to indicate presence.

As the first identifier indicating the position, the above-described identification indicating the head portion is typical. In this case, since it is attached regardless of the content, it is easy to automatically attach it.

The second identifier indicating what the content should be referred to is, for example, when a question about a specific matter is obtained and an answer is obtained, and when the next matter to be spoken can be predicted, the result already spoken is obtained. This is added when the conversation content of the speaker can be predicted or specified to some extent, such as when the content is known. Therefore, it is suitable when the receptionist manually determines the generation timing. Of course, in the case of a question-and-answer conversation such as an interview, it would be easy to automatically generate the identifier.

As the third identifier indicating that the specific content exists, a specific keyword is set in advance, and when the keyword is specified in the conversation, a character string matching the keyword is displayed. Change the mode. For example,
Change the display color.

Next, a description will be given of how an identifier is specifically generated and a corresponding portion (position) of target data is specified.

The first method is a time stamp method. In this method, an identifier is generated as time information indicating a time when the identifier is generated. That is, when a specific event occurs, a method of generating time information indicating the time point, using the elapsed time from a specific starting point, specifying the elapsed time of the event occurrence, and stamping the time information as if And the like. As the time information, for example, calendar information generated in the CPU 110 can be used. Further, a timer may be set to obtain time information indicating the elapsed time.

The time information is, for example, as shown in FIG.
At time t2, when an input instruction of the forward key is made from input device 410, CPU 110 generates a time stamp for that time. At time t3, when a reverse key is input from input device 410, CPU 110 generates a time stamp for that time. Further, after the state in which the volume level is low for a certain period of time, at time t4, when the volume detection unit 340 detects that the volume level has increased, in response to this,
First identifier automatic generation section 350 outputs a first identifier automatic generation signal at time t4. This signal is an identifier generation request signal and has the attribute of a forward key.

The second method is a method of identifying the information based on the information indicating the storage address. This is to generate an identifier as address information indicating a storage address of information stored at the time when the identifier is generated. For example, at a certain point in time when audio data is sequentially stored in the recording device 200, if there is an identifier generation request, a storage address at that time of an address counter built in the CPU 110 is acquired, and this is identified as an identifier. It is assumed that.

As shown in FIG. 4, for example, as shown in FIG. 4, at time t2, when an input instruction of the forward key is made from the input device 410, the CPU 110 stores the audio data at that time. 200 is generated as an identifier. At time t3, when a reverse key is input from the input device 410, the CPU 110 similarly generates, as an identifier, the address of the recording device 200 in which the audio data is stored at that time. Further, after the volume level has been kept low for a certain period of time, at time t4, when the volume level detection unit 340 detects that the volume level has increased, the first identifier automatic generation unit 350 receives this at time t4. It outputs a first identifier automatic generation signal. This signal is an identifier generation request signal and has the attribute of a forward key.

The third method is a keyword method. This associates the keyword with audio data containing the keyword, and specifically associates the keyword with a storage address of audio data corresponding to the keyword. That is, when the input voice data includes a set keyword, the storage address of the voice data is associated with the keyword, and the keyword is stored in the identifier data 2.
11 is stored.

The first identifier and the second identifier generated as described above correspond to the corresponding customer voice data file 2
10 is stored as identifier data 211. In the case of the time stamp, the storage address of the data indicated by each time stamp is obtained by using the storage start time and the storage end time of the audio data, the data amount of the audio data, or the start address and the tail address. , The storage position of the corresponding recorded content can be known. When an address is used as an identifier, the address is used as it is. Further, the time stamp and the address can be used together. In this case, the search can be performed by using both the time stamp and the address.

In any case, since the time when the identifier is generated and the time when the corresponding audio data is stored do not always match exactly, a shift may occur. However, the content to be searched is not instantaneous data, but content of conversation that is continued for a certain period of time. Therefore, even if there is some deviation, there is no problem in practical use.

It should be noted that the time stamp can also be used to arrange the order of speech of a plurality of speakers, that is, the parties involved in the dialogue. In general, in a telephone conversation, there are many cases where statements of parties involved in the conversation are duplicated. When the dialogue is performed in such an overlapping state, if the dialogue is reproduced later, it is difficult to distinguish the utterance contents of each speaker. Therefore, in the present invention, the head part of each speaker's utterance can be searched using the time stamp, and a group of utterances of each speaker can be arranged and reproduced in a time series so as not to overlap. . Specifically, for example, a reading order table is created, and in this table, according to the order of the time stamp,
A group of remarks of each speaker is registered. Then, at the time of reading, the data is sequentially read according to the reading order table.
As a result, it is possible to separate redundant statements and to arrange the flow of the dialogue. It should be noted that, in addition to the recording of voice data, the customer text 214 and the recipient text data 215 can be similarly organized.

Next, generation of the third identifier will be described. The third identifier is different from the first identifier and the second identifier and is not generated for audio data.
Generate for text data. That is, the presence / absence of a preset keyword is determined for the text data converted by the voice recognition unit 310, and if there is a corresponding keyword, the third identifier is generated. For this reason, the generation of the identifier may be performed once after recording.

Next, the operation of this embodiment will be described with reference to the flowcharts of FIGS.

In this embodiment, the following processing is performed prior to the processing by the information processing apparatus. First, the line control unit 5
When there is an incoming call, the sound device 10 sounds a sound device (not shown) to notify the receptionist of the incoming call. When the acceptor goes off-hook, communication is enabled via the transmission / reception device 520, and the speaker 531 and the microphone 532 are connected to the line via the transmission / reception device. When the receiver goes off-hook, an off-hook signal is sent to the information processing device 10. Also, when communication is interrupted, a signal indicating the interruption is sent to CPU 1.
Send to 10.

The information processing apparatus 110 monitors the input of the off-hook signal (step 1001), and when there is an incoming call, the display device 420 specifies items such as the customer's name, telephone number, address, person in charge, etc. Is displayed on the display screen to alert the receptionist (step 100).
2). At this stage, in the case of a system that can capture the telephone number of the other party, the captured telephone number is sent to the CPU 110 and displayed on the display screen to request confirmation.

Next, the receptionist accepts the input of the result of the inquiry about the item for specifying the customer displayed on the screen by the incoming call to the customer, and newly creates the customer voice data file 210. (Step 100
3). For example, an input about a customer name, a department name, a person in charge, a telephone number, an address, a customer code, and the like is received.
These inputs are received by the input device 410. Since the customer specifying data is very important data, in the present embodiment, a form in which the input is manually input is adopted, but, of course, the present invention is not limited to this. For example, from this stage, it may be recorded as audio data. The customer voice data file 210 is provided on the working memory of the CPU 110 for the moment. The data is stored in the storage device 200 when it becomes both to some extent and when the storage is completed.

Further, when the input of a part of the information of the above-mentioned medical inquiry is received, the customer data file is searched for the presence or absence of the candidate of the customer concerned. May be displayed on the display screen of the display device 420. By doing so, it is only necessary to check the items displayed on the display screen,
It is possible to reduce the input burden of the receptionist.

Next, after input of necessary specific items is completed,
When the acceptor inputs a start instruction via the input device 410, the acceptor receives the instruction and starts a voice input process (step 1004). That is, it waits for a voice data input request from either A / D 310 or A / D 320 (step 1005).

When there is a voice data input request,
It is determined which of the input request is / D310 and A / D320, and if it is not a customer, it is instructed to store it as voice data 213 in the voice data file 210 (step 1007). On the other hand, in the case of the customer's voice, it is instructed to store it as the customer voice data 212 in the customer voice data file 210 (step 1008). Here, the distinction between the input voice and the voice of the receiver is made in the present embodiment in a state where the transmitting and receiving device is separated into the transmitting side and the receiving side, and the received signal and the transmitting side are separated. By extracting the signal, it is distinguished between the voice of the receptionist and the voice of the customer. of course,
The distinction between the customer and the receptionist is not limited to this. For example,
It is possible to separate the customer and the receiver by using the difference in the voice level, and to separate the customer and the receiver by using the difference in the width of the frequency band.

Thereafter, it is checked whether the input of the voice data has been completed (step 1009). Specifically, it is checked whether the off-hook signal has been disconnected, whether or not the input device 410 has input an end instruction, and the like. When these are not input, it is assumed that the input is continued.

It is checked whether there is an identifier generation request (step 1010). This is done by checking the identifier input from the first identifier automatic generation unit 350 and the presence or absence of an identifier (second identifier) generation request from the input device 410.

Upon detecting the identifier generation request signal from the first automatic identifier generation unit 350, the CPU 110 obtains the time (including the date) from the built-in calendar, and further records the time at that time. The storage destination addresses of the audio data to be performed are obtained, and the addresses are associated with each other, and the identifier data is stored in the customer audio data file 210 together with the preset attribute of the identifier, in this case, the attribute information indicating the forward key. It is stored as 211 (step 1011). That is, in this embodiment, a time stamp and an address are used together as an identifier.

When an identifier generation request is received from the input device 410, the identifier generation request is acquired from the built-in calendar in the same manner as described above.
Further, at that time, the storage destination address of the audio data to be recorded is obtained, and these are associated with each other, together with the attribute of the specified identifier, the attribute information indicating the forward key or the reverse key, and the customer audio data file. 21
0 is stored as identifier data 211 (step 10).
11).

Next, similarly to step 1009 described above, it is determined whether the input of the voice data has been completed (step 1012). If not, the process returns to step 1006. On the other hand, in the case of termination (including the case of step 1009), a screen indicating that acceptance of the complaint from the customer by telephone is terminated is displayed (step 1013).

At this time, the audio data file 210
Is displayed together with a question as to whether or not to transfer to the processor B, and when a transfer destination is specified and an instruction to transfer is received,
The CPU 110 transmits the audio data file 210 to the processor B from the communication control device 430 via the communication network M. Note that the content may be sent in response to an inquiry from the processor B without transferring the audio data file 210.

Next, when it is necessary to convert the audio data recorded in the recording device 200 into text data, upon receiving an instruction from the input device 410, the CPU 1
Reference numeral 10 denotes customer voice data 21 recorded in the voice data file 210 for the designated customer from the voice data file 210 stored in the recording device 200.
2 is transferred to the voice recognition unit 330 by the length of the processing. The voice recognition unit 330 performs a recognition process on the transferred voice data using a voice pattern dictionary prepared in advance, and converts the recognition result into a character code. Then, a character string for the transferred portion is generated. The generated character string is transmitted to the audio data file 210 by the CPU 110.
Is stored as text data 214. At this time, a character string may be corrected via the input device 410 while being displayed on the display device 420 so that the receptionist can confirm the conversion state. In this way, the unclear voice / character conversion can be converted into accurate conversion content. If a kana-kanji conversion program is installed as a program for the CPU 110, the
With the PU 110, kana-kanji conversion can be further performed on the character string subjected to the text conversion. In this case, the kana-kanji mixed sentence after the conversion can be displayed on the display device 420 so that the receptionist can confirm the sentence and make the necessary editing via the input device. Therefore, the contents of the customer's remark can be recorded in a more readable sentence, and the reference can be made easier.

When a translation program is installed as a program in the memory 120, the data converted into a sentence mixed with kana-kanji can be translated into a sentence in a specific language. Of course, when accepted in a foreign language such as English, it can be converted to a sentence mixed with kana-kanji or another foreign language sentence.

Next, the voice data converted into text data is received by the third identifier automatic generation section 360, and it is checked whether or not the text data includes a keyword registered in advance, and the keyword is included. In this case, the display attribute of the portion is set to a specific mode so that the display mode can be changed. For example, display color, background color, blinking,
Set to add shading, underline, etc.

Further, in the above example, an example is shown in which an identifier is added to text data itself, but the present invention is not limited to this. For example, when a character string corresponding to the keyword is included, a set of audio data corresponding to the keyword can be associated with the keyword, and the keyword can be used as an identifier for the set of audio data. In this case, it can be made to function as an identifier of both the voice data and the corresponding text data. At this time, at least one of the above-described time stamp and storage address can be further associated. In particular, by associating with a time stamp, it is convenient to arrange the speech order of the dialogue.

As described above, in the recording device 200, the dialogue with the customer received by telephone is stored in the same customer voice data file 210 as customer voice data, receptionist voice data, identifier data, and text data. Is done. Therefore, the processor can later hear the customer's voice by referring to the necessary customer voice data file, can also know the content by text data, and further, by using the identifier, The audio data can be reproduced only in the necessary places. Therefore, necessary information can be efficiently acquired. In addition, text data can be displayed or printed by a printer (not shown) for use. At this time, since the display mode of the specific term has been changed, the search for the specific keyword becomes easy.

In the above example, voice data is converted into text data as a batch process after recording. The present invention is not limited to this. For example, audio data may be converted to text data in real time during reception by telephone. In that case, all voice data of the customer can be replaced with text, but a simpler method can be adopted. For example, with respect to the content of the customer's remarks, the receptionist performs a reconfirmation voice into the microphone for important terms contained therein, and converts the voice of the receptionist into a character string by the voice recognition unit 330. You may. In this case, the recognition rate can be improved by learning the voice recognition unit 330 for a specific speaker. Moreover, the recognition rate can be further improved because the receptionist is conscious of speech recognition and pronounces clearly.
In addition, the third identifier automatic generation unit 360 can output a request for automatically generating an identifier for an important word generated by the acceptor.

As another mode of voice recognition,
For example, it is possible to automatically perform voice recognition on the voice data between the forward key and the reverse key. In this way, since important remark contents can be converted into text data, it becomes easy for the processor to search for important parts.

Next, another embodiment of the present invention will be described. FIG. 6 shows the hardware system configuration. In the present embodiment, two voice recognition units 330, third identifier automatic generation units 360, sound volume detection units 340, and first identifier automatic generation units 350 are provided for each of the customer and the receiver. The recording apparatus 200 has the same hardware system configuration as the embodiment shown in FIG. 3 described above, except that the text data is divided into a customer text 214 and a receptionist text 215. .

Various devices are connected to the information processing apparatus 100 via the interface 130. That is,
A recording device 200 and an analog / digital converter (A /
D) 310 and an analog / digital converter (A / D) 3
20, voice recognition units 330a and 330b, volume detection units 340a and 340b, first identifiers 350a and 350b, and third identifier automatic generation units 360a and 36.
0b, a voice synthesizing unit 370, and a digital-to-analog converter (D / A) 380 are connected so as to be built in the information processing apparatus 100. Of course, each of these devices may be an external device. The information processing device 10
The input device 410, the display device 420, the communication control device 430, and the audio output unit 440 are connected to 0. In the present embodiment, the voice recognition units 330a and 33a
0b, volume detectors 340a and 340b, first identifiers 350a and 350b, and third identifier automatic generator 36
Although 0a and 360b are provided for the customer and the receiver, respectively, the present invention is not limited to this. A single configuration may be used for both customers and receptionists. Further, in order to correspond to three or more speakers, a voice recognition unit and the like may be multiplexed to correspond to three persons, respectively.

The recording device 200 is constituted by a hard disk device, as in the first embodiment. The recording device 200 is provided with a customer voice file 210 for each customer who makes a call, and customer data 290 including the customer's address, telephone number, various attribute data, and the like. The customer voice file 210 includes customer voice data 212 that is a digitized voice signal of the customer, and receiver voice data 213 that is a digitized voice signal of the receiver.
Customer text data 214 obtained by voice recognition, receptionist text data 215, and identifier data 2 described later
11 is recorded.

In this embodiment, the identifier data 2
11, customer voice data 212, receptionist data 213, customer text data 214, and receptionist text data 2
15 are stored in the same file. However, the present invention is not limited to this. Some or all of the data may be stored in an independent form without being collected for each customer. For example, the identifier data 211 may be stored by itself like the customer data 290. The identifier data 211 and the customer voice data 21
2. The acceptor data 213, the customer text data 214, and the acceptor text data 215 may all be stored independently.

The voice recognition units 330a and 330b read the digitized voice signal, convert the voice into a corresponding character or character string, and generate text data. In the present embodiment, they are connected as independent devices. Of course, you may make it process by CPU110. In the present embodiment, the voice recognition unit 33
0a is used for an unspecified speaker in order to make customer voice recognition. In addition, the voice recognition unit 330b uses a voice corresponding to a specific speaker in order to perform recognition of the voice of the receiver. Of course, both of them may be adapted to an unspecified speaker.

The volume detectors 340a and 340b have a function of detecting a point at which the level sharply increases after the sound is interrupted. That is, the volume detector 340a
Detect the level of the voice signal of the customer, determine whether it is equal to or greater than a predetermined reference value, if it is less than the reference value, determine that the speech is interrupted, and then when the state exceeds the reference value, It determines that the speech has begun and outputs a cue signal. FIG. 4 shows an example. For example, before time t4, the volume level remains low for a while (silence state), and at time t4, the volume level sharply increases. The volume detector 340a detects this state,
Outputs the first signal. This head signal is input to the third identifier automatic generation unit 360a. Also, the volume detection unit 340
b detects the volume of the receptionist, detects the leading signal after the disconnection in the same manner as in the case of the customer described above, and sends the leading signal to the third identifier automatic generation unit 340b. In the present embodiment, the sound volume is detected for both the customer and the receptionist. However, only one of them, for example, the voice of the customer may be detected.

First identifier automatic generation units 350a and 35a
0b receives the output signals of the corresponding volume detectors 340a and 340b, respectively,
Generate an identifier. This identifier will be described later. The generated identifiers are stored in the corresponding customer voice data files 210, respectively.

Third identifier generating sections 360a and 360b
Determines whether or not the speech text data converted into the text data in the corresponding speech recognition units 330a and 330b includes a preset keyword, and generates a third identifier when the keyword is included.

The voice synthesizing section 370 reads the text data and converts it into a voice signal having a predetermined sound quality. This audio signal is analog data, and is output from the audio output unit 440.
Is output as voice.

The D / A 380 converts customer voice and receptionist voice recorded in the form of digital data into analog signals. The converted analog audio signal is output as audio by the audio output unit 440.

The voice output device 440 is provided with a voice synthesizer 370
The analog audio signal output from is converted to sound and output. Specifically, it is composed of an amplifier and a speaker.
In addition, it can also be comprised by headphones.

Next, the operation of the reception process in the present embodiment will be described with reference to the flowchart of FIG. 5 in addition to FIG. However, the basic processing procedure is the same as that of the embodiment of FIG. 3 described above, and therefore, the description will focus on the differences.

In the present embodiment, the processing of steps 1001 to 1009 is executed as in the above-described embodiment.
The description of these processes will not be repeated.

It is checked whether there is an identifier generation request (step 1010). This is done by checking the identifier input from the first identifier automatic generation unit 350a or 350b and the presence / absence of an identifier (second identifier) generation request from the input device 410. First identifier automatic generation unit 350a or 35
0b, the CPU 1
10 obtains the time (including year, month and day) at that time from a built-in calendar, further obtains a storage destination address of audio data to be recorded at that time, The identifier is stored in the customer voice data file 210 as identifier data 211 together with a preset attribute of the identifier, in this case, attribute information indicating the forward key (step 1011). Thereafter, step 101
Steps 2 and 1013 are executed.

Next, a case will be described in which an identifier is associated with voice data converted into text data. That is,
In this case, the corresponding third identifier automatic generation unit 360a
Alternatively, at 360b, it is checked whether or not the corresponding text data includes a keyword registered in advance, and if the keyword is included, the display attribute of the part is specified by a specific attribute so that the display mode can be changed. Set the mode.
For example, the display color, the background color, the blinking, the shading, the underline, and the like are set.

In the present embodiment, for example, the receptionist performs a reconfirmation voice to the microphone with respect to important terms contained in the contents of remarks of the customer,
The voice of the acceptor may be converted into a character string by the voice recognition unit 330b. In this case, the voice recognition unit 33
By learning 0b for a specific speaker, the recognition rate can be improved. Moreover, the recognition rate can be further improved because the receptionist is conscious of speech recognition and pronounces clearly. In addition, the third identifier automatic generation unit 360b can output a request for automatically generating an identifier for an important word generated by the acceptor. In this case, the important keyword is associated with a group of audio data including the keyword.

Further, as another mode of voice recognition,
As described above, for example, the voice data of the portion sandwiched between the forward key and the reverse key can be automatically recognized.

In this embodiment, since a plurality of voice recognition units and the like are provided for each speaker, each speaker can be processed independently. Therefore, even if the amount of information to be processed is large, it is possible to respond quickly.

In each of the above examples, speech recognition is performed in the same information processing apparatus 100. However, the present invention is not limited to this. For example, it may be performed in another information processing device.

In each of the above examples, A / D is used. However, when audio data is input to the information processing apparatus 100 in the form of digital data, A / D can be omitted. . In this case, D / A is used for reproduction by the speaker 531. A / D is used to digitize the input from the microphone 532.

Further, in each of the above-described examples, the content of the dialogue between the customer and the receptionist is recorded. However, it is also possible to record only the utterance contents of one speaker of the dialogue.
For example, only the voice of the customer may be recorded.

[0093] In addition, although the above-mentioned example shows the case of receiving by telephone, the present invention is also applicable to a case where no telephone is used, that is, a case where direct conversation is performed at a window or the like.
In this case, in order to clarify the distinction of the speakers, a microphone to be used may be separately prepared in advance, and the speakers may be distinguished based on differences in microphone inputs.

[0094]

According to the present invention, a complaint, an inquiry, a report of a failure, a report of an accident, etc. can be received by voice, the received information is recorded, and an identifier is provided at a specific place. Since it can be provided, it can be transmitted to the processor in a form that can be used for later processing.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an outline of an environment to which a voice reception processing system of the present invention is applied.

FIG. 2 is an explanatory diagram showing an outline of information exchange at a reception center to which the present invention is applied.

FIG. 3 is a block diagram showing an example of a configuration of a hardware system used in the embodiment of the present invention.

FIG. 4 is an explanatory diagram showing an example of identifier generation of an identifier by a volume detector used in the present invention.

FIG. 5 is a flowchart illustrating a procedure of a dialog recording process according to the embodiment;

FIG. 6 is a block diagram showing an example of a configuration of a hardware system used in another embodiment of the present invention.

[Description of Signs] 100: Information processing device, 110: Central processing unit (CP
U), 120 ... memory, 130 ... interface, 20
0: recording device, 210: customer voice data file, 21
1 ... Identifier data, 212 ... Customer voice data, 213 ...
Recipient voice data, 214 ... text data, 310,
320 ... analog-to-digital converter (A / D), 33
0, 330a, 330b ... voice recognition units, 340, 340
a, 340b: Volume detection unit, 350, 350a, 350
b: first identifier automatic generation unit, 360, 360a, 360
c: third identifier automatic generation unit, 410: input device, 420
... display device, 430 ... communication control device, 500 ... telephone device, 510 ... line control unit, 520 ... transmit / receive device, 531
... speaker, 532 ... microphone.

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Hideo Inoue 504-2 Shinanomachi, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside Hitachi Electronics Service Co., Ltd. (72) Inventor Reika Yamagishi 504-2 Shinanocho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Within Hitachi Electronics Service Co., Ltd. (72) Inventor Mutsuji Takesada 504-2 Shinanomachi, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture F-term (reference) 5B075 KK02 KK34 KK35 ND03 ND14 ND24 ND26 NK02 NK13 NK24 NK31 UU40

Claims (11)

[Claims] 1. A dialogue recording system for recording contents of at least one of dialogues, comprising: a recording device for recording voice data of a dialogue; and an identifier for identifying a specific part of the voiced data to be recorded. And an information processing device that performs a process of causing the recording device to perform recording. The information processing device requests the generation of an identifier for identifying a specific portion of the audio data recorded in the recording device. A dialogue recording system which receives an identifier, generates an identifier, and records the identifier in the recording device in association with audio data to be recorded. 2. A dialogue recording system according to claim 1, wherein a volume detection means for detecting a rise of the voice after the voice is interrupted, based on a volume of the input voice, and a voice input at that time. Means for outputting an identifier generation request signal for identifying data, wherein the information processing apparatus receives the identifier generation request signal and generates an identifier. 3. The dialogue recording system according to claim 1, further comprising an input device having a key for requesting generation of an identifier, wherein said information processing device is configured to output a signal for requesting generation of an identifier from said input device. Receiving the information, the identifier is generated. 4. The dialog recording system according to claim 3, wherein the identifier generation request signal from the input device identifies an area recorded in the future in time from the position of the identifier, and A dialogue recording system characterized in that an area recorded in the past is identified from the position of the identifier, and the input device has a function of inputting each of the areas. 5. The dialogue recording system according to claim 1, wherein the information processing device includes, as the identifier, time information at a time point when the identifier generation request signal is received. A dialogue recording system characterized by using. 6. The dialogue recording system according to claim 1, wherein the information processing device receives the identifier generation request signal as the identifier and outputs a corresponding voice. A dialog recording system using an address for recording data in the recording device. 7. The dialogue recording system according to claim 1, further comprising: voice recognition means for performing voice recognition on digitized voice data and converting the digitized voice data into a character string. The information recording apparatus according to claim 1, wherein the information processing apparatus causes the recording device to record the converted character string as text data. 8. The dialogue recording system according to claim 7, wherein the character string converted from the voice by the voice recognition unit is checked to determine whether or not the character string includes a predetermined keyword. A dialog recording system, further comprising means for changing a display mode of a corresponding part. 9. The dialogue recording system according to claim 1, wherein the voice recognition unit performs voice recognition on the digitized voice data and converts the voice data into a character string. The character string converted from the voice by the voice recognition means is examined to determine whether it includes a predetermined keyword. If the keyword includes a keyword, the keyword is associated with a set of voice data corresponding to the keyword, and the keyword is identified. A dialogue recording system, further comprising: means for setting an identifier for a group of voice data. 10. The method of claim 1, 2, 3, 4, 5, 6, 7,
The dialogue recording system according to any one of claims 8 and 9, wherein the information processing device records each voice of a speaker of the dialogue in a different area in the recording device. Recording system. 11. A telephone device for receiving a call from the outside and making a telephone call, and the telephone device according to claim 1, 2, 3, 4, 5, 6, 7,
11. A reception processing system comprising the dialogue recording system according to any one of 8, 9, and 10, wherein the telephone device includes a microphone and a speaker for talking, and receives a voice signal and the microphone. And an audio signal to the information processing apparatus, and the information processing apparatus causes the recording apparatus to record the input audio signal.