JP2010182191A - Business form input device, business form input system, business form input method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve accuracy and efficiency of business form input operation using voice recognition. <P>SOLUTION: When a keyword stored in a recognition keyword database 204 is recognized in input voice data, a recognition event determining section 103 selects an event (action) corresponding to the keyword. A business form screen changing section 104 specifies change of a business form screen or change of entry field selection according to the event, and a business form display section 105 updates and displays a screen of an operator terminal. A language model selecting section 303 selects a language model 301 required for input to the business form screen changed with reference to a business form meta-information language model-corresponding database 302. Data input to the business form using the selected language model 301 is stored in a business form database 202. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a form input technology for inputting information such as conversation contents into a form by using voice recognition in a telephone reception work of a call center.

In the telephone reception work in the call center, the operator needs to input the contents of the call such as the customer's inquiry and request into the form while answering the call with the customer, and appropriately respond to the customer. In addition, it is possible to input the history of the call contents after the call with the customer is finished, but there are problems such as input errors and work burden when there is a large amount of information to be input. there were.
Therefore, various form input techniques using voice recognition have been developed.

For example, in the call center system described in Patent Document 1, the user's voice is received, the recognition result obtained by voice recognition, the recognition result is displayed on the operator's display device, and the recognition result displayed by the operator is referred to. Then, the recognition result with the higher recognition rate is selected as the final recognition result among the recognition results obtained by recognizing the repeated voice.
In addition, the call center system described in Patent Document 2 acquires voice recognition information related to an operator's call content, refers to determination information related to vocabulary and phrases stored in the database, and whether or not the operator's response is appropriate. Such a determination is made.
In the conventional call center system described in Patent Document 3, the server recognizes the telephone response as voice and outputs it as character data to the display device, and outputs past response information stored in the server as character data. The telephone responder is a system that can appropriately respond with reference to the character data.
In addition, a technique has been developed in which a plurality of language models (dictionaries) used for speech recognition are prepared or a specialized language model is prepared.

Japanese Patent Laid-Open No. 10-322450 JP 2008-2111271 A JP 2005-110034 A

In the above call center system, it is possible to display and store the recognition result of voice recognition of telephone reception as character information, but input corresponding to complicated forms with many forms and many input fields However, it was difficult to perform the process, and there were problems such as an input error in the form. In addition, it is difficult for the operator to manually change the language model of the form and voice recognition while answering the telephone, and there is a problem that a switching error occurs.
The present invention has been made in view of the above points, recognizes the contents of utterances of customers and operators using voice recognition, dynamically changes a form column according to the utterance contents, and selects an appropriate language model. The purpose of this is to improve the recognition accuracy and to improve the efficiency and accuracy of the input operation to the form.

  In order to achieve the above-described object, the present invention is a form input device that supports data input work to a form, and associates the meta information of the form with a language model and stores it as form meta information language model corresponding data. A form meta information language model corresponding data storage means, a keyword data storage means for associating the meta information, keywords and event information of the form and storing them as keyword data, a voice recognition means for recognizing voice data, When the keyword is recognized in the voice data, event execution means for executing an event based on the keyword data, and when the form to be input is changed by the event, based on the data corresponding to the form meta information language model Language model for selecting the language model corresponding to the meta information of the changed form Selection means, characterized by having a a form input means for determining an input information of said document by using said selected language model.

Here, in the form input device of the present invention, a language model corresponding to the meta information of the form (form template), a keyword corresponding to the meta information of the form, and event information are stored in advance. When the input voice data is recognized and a keyword is recognized in the voice data, the form input device performs an event based on the event information corresponding to the keyword, changes the form screen, etc. The language model based on the corresponding form input field (meta information) is used to accept input or to perform automatic input.
Moreover, it is also possible to store the limitation of the input information of the form as event information.
In the present invention, language model selection and form input can be automatically performed, and manual form input, manual language model switching, form switching, and the like are not performed. Therefore, reduction of form input work and improvement of accuracy can be realized.

  Further, the second invention associates the meta information of the form with the language model and stores it as form meta information language model correspondence data, associates the form meta information with the keyword and the event information, A step of storing as data, a step of recognizing audio data, a step of executing an event based on the keyword data when the keyword is recognized in the audio data, and a form to be input changed by the event If so, a step of selecting a language model corresponding to the meta information of the changed form based on the form meta information language model correspondence data, and determining input information of the form using the selected language model A form input method characterized by comprising the steps of:

According to a third aspect of the present invention, there is provided a form input system having an operator terminal capable of voice communication with a customer terminal, and a server capable of transmitting and receiving data via the operator terminal and a network. A form meta information language model correspondence data storage means for associating information with a language model and storing it as form meta information language model correspondence data, and associating and storing the form meta information, keyword and event information as keyword data Keyword data storage means, voice recognition means for recognizing voice data received from the customer terminal and the operator terminal, and execution of an event based on the keyword data when the keyword is recognized in the voice data Event execution means to be input by the event Means for changing a form and displaying it on the operator terminal; language model selecting means for selecting a language model corresponding to the meta information of the changed form based on the data corresponding to the form meta information language model; and the selection And a form input unit that determines input information of the form using the language model thus formed.
The fourth invention is a program for operating a computer as a form input device according to the first invention.

  According to the present invention, a form input operation using voice recognition can be easily and quickly performed, and the accuracy of the input operation to the form is improved.

2 is a hardware configuration diagram of a form input system 1. FIG. It is a figure which shows the functional block structure of the form input system 1, and the flow of data. It is a figure which shows an example of the meta information of a form. It is a figure which shows an example of a database corresponding to a form meta information language model. It is a figure which shows an example of a recognition keyword database. It is a sequence diagram which shows the flow of a data registration process. It is a sequence diagram which shows the flow of a form input process. It is a figure which shows an example of a form screen. 2 is a hardware configuration diagram of a form input system 11. FIG. It is a figure which shows the functional block structure of the form input system 11, and the flow of data. It is a sequence diagram which shows the flow of a data registration process. It is a figure which shows an example of the meta information of a form. It is a figure which shows an example of the speech determination database. It is a sequence diagram which shows the flow of a form input process. It is a figure which shows an example of a form screen.

  Hereinafter, preferred embodiments of a form input apparatus according to the present invention will be described in detail with reference to the accompanying drawings. In the following description and the accompanying drawings, the same reference numerals are given to components having substantially the same functional configuration, and redundant description is omitted.

FIG. 1 is a hardware configuration diagram of a form input system. The form input system 1 includes a server 3, an operator terminal 5, a form manager terminal 7, and a customer terminal 9.
The customer terminal 9 is a terminal device capable of performing voice communication via the public line network 12 installed in the customer 8, and is a telephone, a mobile phone, a computer, or the like. As shown in FIG. 1, an example will be described in which a customer 8 connects to a public line network 12 from a customer terminal 9 and makes an inquiry to the operator terminal 5 through a PBX (Private Branch eXchange) 13. The server 3, the operator terminal 5, and the form manager terminal 7 can transmit and receive data via the network 14. The PBX 13 is, for example, an IP-PBX or the like, and can construct a telephone network using the network 3 connected to the server 3, one or more operator terminals 5, and the form manager terminal 7. It is possible to realize a telephone call with the operator 4.

  The operator terminal 5 is a computer or the like, and includes a control unit 51, a storage unit 52, an input unit 53, an output unit 54, and a communication unit 55. The control unit 51 is a central processing unit (CPU), a microprocessor, or the like, and controls each unit of the operator terminal 5. The storage unit 52 is a storage device such as a nonvolatile memory, a volatile memory, or a hard disk. The input unit 53 is an input device such as a keyboard or a mouse, or a voice input device such as a headset or a microphone. The output unit 54 is an audio output device such as a display device or a speaker. The communication unit 55 is a device that performs data transmission / reception with the server 3 and voice communication with the customer terminal 9 via the network 14.

  The form manager terminal 7 is a computer or the like, and is installed in the form manager 6. The form manager terminal 7 includes a control unit 71, a storage unit 72, an input unit 73, an output unit 74, and a communication unit 75. These functions and operations are the control unit 51 and the storage unit 52 of the operator terminal 5 described above. , The same as the input unit 53, the output unit 54, and the communication unit 55.

  The server 3 is a server computer and corresponds to a form input device. The server 3 includes a control unit 31, a storage unit 32, an input unit 33, an output unit 34, and a communication unit 35. The control unit 31 is a central processing unit (CPU), a microprocessor, or the like, and controls each unit in the server 3 and executes processing to be described later. The storage unit 32 is a storage device such as a nonvolatile memory, a volatile memory, or a hard disk. The input unit 33 is an input device such as a keyboard or a mouse. The output unit 34 is a display device or the like. The communication unit 35 is a device that transmits and receives data to and from the operator terminal 5 and the form manager terminal 7 via the network 14.

  FIG. 2 is a diagram illustrating a functional block configuration of the server and a data flow. The functions of the server 3 include a form creation unit 81, a form management unit 82, and a voice recognition dictionary management unit 83. The form creation unit 81 includes a voice input unit 101, a voice recognition unit 102, a recognition event determination unit 103, a form screen change unit 104, a form display unit 105, and a form input unit 106. The form management unit 82 includes a form meta information database 201, a form database 202, a form management data input unit 203, and a recognition keyword database 204. The speech recognition dictionary management unit 83 includes a language model 301, a form meta information language model correspondence database 302, and a language model selection unit 303.

  The voice input unit 101 accepts input of voice (utterance) data input from the input units of the operator terminal 5 and the customer terminal 9. The voice recognition unit 102 recognizes voice data input using a language model (dictionary) 301. The language model 301 used at this time is selected by the language model selection unit 303 using the form meta information language model correspondence data stored in the form meta information language model correspondence database 302.

  The recognition event determination unit 103 determines the operation (event) of the information in the form meta information database 201 and the recognition keyword database 204, and the form screen change unit 104 dynamically changes the form display data according to an instruction from the recognition event determination unit 103 and the like. change. The form display unit 105 displays the form on the operator terminal 5 or the like. The form input means 106 determines input information by receiving input to the form by the operator terminal 5 or the like, and stores it in the form database 202.

  The form management data input means 203 receives the form template and its meta information, the voice recognition keyword and the corresponding event information input from the form manager terminal 7 by the form manager 6. The form meta information database 201 stores a form template and its meta information. The form database 202 stores data input to the form. The recognition keyword database 204 stores keywords corresponding to the meta information and event (operation) information corresponding to the keywords.

  The language model 301 is a plurality of dictionaries used when the form input unit 106 determines form input information, and is also used when the speech recognition unit 102 performs speech recognition. The form meta information language model correspondence database 302 is information indicating correspondence between form meta information and a language model used for inputting the form. The language model selection unit 303 selects a language model 301 suitable for speech recognition and form input.

Below, the information memorize | stored in each database of the server 3 is demonstrated.
FIG. 3 is a diagram showing a part of the form meta information stored in the form meta information database 201. The form meta information is form data and indicates information about each input field on the form screen. In the form, a single selection combo box, a multiple selection check box, a date / time edit field, a number edit field, a free input field, a push button, and other input fields are provided on the screen. An example of the meta information shown in FIG. 3 is described in XML. For example, the 601 part shown in FIG. 3 has an attribute “Combobox” and an attribute value “23/10/2008 13:00” for the form column ID “accidentdate”. "Is defined.

  FIG. 4 is a diagram illustrating an example of form meta information language model correspondence data stored in the form meta information language model correspondence database 302. The form meta information language model correspondence data shown in FIG. 4 includes a form column ID of meta information and a corresponding language model (dictionary name), and is expressed as a table item in a relational database. For example, 641 parts shown in FIG. 4 indicate that the language model 301 representing “date and time” corresponds to the form field ID “accidentdate” of 601 parts shown in FIG. The language model selection unit 303 uses the form meta information language model correspondence data to select an optimal language model 301 to be used when inputting data into the form field for each form field (form input field) of the form.

FIG. 5 is a diagram illustrating an example of recognized keyword data stored in the recognized keyword database 204. The recognized keyword data shown in FIG. 5 includes a form column ID, a recognized keyword, an attribute, and event (operation) information corresponding to the keyword, and is expressed as a table item in a relational database. For example, in the case of 651 shown in FIG. 5, when the keyword “occurrence date” is recognized by voice for the form column ID “accidentdate”, the form attribute is “Combobox” and the operation is defined as “calendar selection”. Is done. The recognition event determination unit 103 determines the operation stored in the recognition keyword database 204 when the keyword stored in the recognition keyword database 204 is recognized in the recognition data recognized by the voice recognition unit 102, and the form screen change unit 104 performs the operation and changes the form screen.
In this way, in correspondence with the form column “accidentdate”, 641 parts shown in FIG. 4 are defined as form meta information language model correspondence data, and 651 parts shown in FIG. 5 are defined and stored as keyword data.

Similarly, the meta information 602 in FIG. 3 corresponds to 642 in the form meta information language model correspondence database 302 shown in FIG. 4 and 652 in the recognition keyword database 204 shown in FIG. 3 corresponds to 643 parts of the form meta information language model correspondence database 302 shown in FIG. 4 and 653 parts of the recognition keyword database 204 shown in FIG. 3 corresponds to 644 parts of the form meta information language model correspondence database 302 shown in FIG. 4 and 654 parts of the recognition keyword database 204 shown in FIG. Also, the meta information of 605 in FIG. 3 corresponds to 645 in the form meta information language model correspondence database 302 shown in FIG. 4 and 655 in the recognition keyword database 204 shown in FIG.
In this way, the form input field and operation corresponding to the keyword are defined, and the language model 301 for the input in each input field of the form is defined.

Next, the operation of the form input system 1 will be described.
Here, the case where the operator 4 is a call center operator and the customer 8 makes an inquiry to the call center from the customer terminal 9 will be described as an example.
First, the form manager 6 adds, updates, and deletes the form meta information in advance before starting the call center business. FIG. 6 is a sequence diagram showing a flow of data registration work performed by the form manager 6 from the form manager terminal 7.
The form manager terminal 7 inputs the form meta information (step S1001) and transmits it to the server 3. The server 3 registers the form meta information in the form meta information database 201 (step S1002), refers to the form column ID of the input form meta information, searches for and selects the language model 301 used in the form. prepare.

Next, the form manager terminal 7 inputs the recognition keyword data shown in FIG. 5 (step S1004) and transmits it to the server 3. The server 3 stores the keyword information in the recognition keyword database 204 (step S1005).
Next, a case where a call center operation is started and a customer terminal 9 makes an inquiry by telephone to the operator terminal 7 will be described as an example.

FIG. 7 is a sequence diagram showing the flow of the form input operation by this system.
When an utterance such as a telephone inquiry and reply between the customer 8 and the operator 4 is started, the voice data (utterance information) of the utterance input from the input unit 53 such as the voice input device of the operator terminal 5 is the voice of the server 3. It is transmitted to the input means 101 (step S1201). The voice recognition unit 102 of the server 3 recognizes the voice data (step S1202), and determines whether there is an utterance of the keyword stored in the recognition keyword database 204 (step S1203).

  If there is an utterance of a keyword stored in advance in the recognized keyword database 204 (“Yes” in step S1203), the recognized event determining unit 103 selects an event (operation) corresponding to the keyword according to the data in the recognized keyword database 204 ( Step S1204). Next, the operator moves the focus (such as moving the mouse to the input field) to determine whether or not there is an operation such as changing the input field (step S1205). If there is an action (“Yes” in step S1205), the form screen changing unit 104 issues an instruction to change the form screen or change the input field selection according to the event (step S1206). The screen is updated and displayed on the output unit 54. Further, the language model selection unit 303 selects and prepares the language model 301 necessary for inputting the changed form screen or the input field (step S1207).

In response to the form screen change instruction from the server 3, the operator terminal 5 changes the form screen displayed on the output unit 54 (step S1208), and the operator terminal 5 uses the language model 301 prepared for the form. Accepts form entry. It is determined whether or not the utterance between the customer 8 and the operator 4 has ended (step S1209). When the utterance ends, the form data (information input as the form template) is transmitted to the form input means 106 of the server 3 ( In step S1210), the form input unit 106 registers the form data in the form database 202 (step S1211).

If there is no keyword registered in the utterance in step S1203, or if there is no event corresponding to the keyword in step S1205, the speech-recognized recognition data is directly registered in the form database 202 (step S1211).
In step S1209, if the utterance does not end, speech recognition of the utterance is continued, and the operations from step S1201 to step S1211 are repeatedly executed until the utterance is ended.

For example, when the customer 8 requests the call center to repair a broken machine, the call center operator 4 utters “The failure date and time is October 28”. Among the utterances, the keyword “occurrence date” is stored in the recognition keyword database 204 shown in FIG. 5, the form column ID is “accidentdate”, the attribute is “Combobox”, and the event (operation) is “ Corresponding to “Calendar selection”.
There is a keyword utterance in step S1203, and an event corresponding to the keyword exists in step S1205. The server 3 prepares a form screen including the form column ID “accidentdate” (step S1206), and is shown in FIG. Based on the form meta information language model correspondence database 302, the “date and time” language model 301 corresponding to the form column ID “accidentdate” is selected and prepared (step S1207).

The operator terminal 5 displays a form screen including the form column ID “accidentdate” (step S1208).
FIG. 8 shows an example of a form displayed on the operator terminal 5. A form column 701 shown in FIG. 8 is an input column for “occurrence date”. The server 3 can perform speech recognition using the recognized keyword and the language model 301 based on the recognized keyword (in this case, the “date and time” language model), and can automatically set the occurrence date in the input field. . For example, in this case, as shown in the form column 701 in FIG. 8, “October 28, 2008” is automatically entered as the “occurrence date” in the form entry field 701. Further, the event information corresponding to the form column ID “accidentdate” becomes “calendar selection”, and the operator 4 can input “date” selected from the calendar displayed on the operator terminal 5 in the input column 701.

Further, when the call center operator 4 focuses on the symptom column 702 shown in FIG. 8 using the input unit 53 such as a mouse, the server 3 selects the language model 301 related to “symptom” and the symptom that the customer 8 speaks. It is possible to recognize the voice data related to the data and register the recognized data in the form database 202 as form data.
Further, when the customer 8 speaks to the operator 4 as “No need to work” or “No problem” in the input in the work column 703, the server 3 displays the keyword data 655 shown in FIG. “No problem” is recognized as the registered keyword, and the combo box in the work column 703 is automatically operated as “input impossible”.

  As described above, the form system 1 recognizes speech of the customer 8 and the operator 4 in real time, dynamically changes the form screen based on the keyword spoken therein, and stores it in association with the keyword. Using the selected language model 301, the language model 301 necessary for the form input is selected, and the form column associated with the keyword is automatically input using the selected language model 301, or other decisions are made. Perform the specified action. In addition, regarding the input of a form, it is possible to recognize spoken voice and use it as input information. Further, it is possible to restrict the input information and input items of the form. Thereby, the operator's input work is reduced, the form input work time is shortened, and the accuracy of the form input work is improved.

  In the present embodiment, the server 3 is a single server computer, but it is also possible to perform processing in a distributed manner by a plurality of server computers connected via a network. For example, among the functions of the server 3, the speech recognition server performs the processing related to the speech recognition function, the speech input unit 101, the speech recognition unit 102, and the speech recognition dictionary management 83, and the form management server performs the other functions. It is also possible to execute the above processing by transmitting and receiving data between the server and the form management server.

Next, a second embodiment of the present invention will be described.
FIG. 9 is a hardware configuration diagram of the form input system 11. The form input system 11 has almost the same structure as the form input system 1 shown in FIG. 1, and is different from the form input system 1 in that an operator manager terminal 21 is added to the network.
The operator manager terminal 21 is a computer used by the operator manager 22, and includes a control unit 91, a storage unit 92, an input unit 93, an output unit 94, and a communication unit 95. These control the operator terminal 5 described above. Unit 51, storage unit 52, input unit 53, output unit 54, and communication unit 55 have the same functions and perform operations.

FIG. 10 is a diagram showing a functional block configuration and data flow of the form input system 11. The functional block configuration of the form input system 11 is different from that of the form input system 1 shown in FIG. 2 in that an operator management unit 84 is added. The operator management unit 84 includes an operator management data input unit 401, an utterance determination database 402, and an utterance determination unit 403.
The operator management data input unit 401 accepts keyword information input from the operator manager terminal 21 and stores it in the utterance determination database 402. Based on the data stored in the utterance determination database 402, the utterance determination unit 403 performs an event (operation) such as a form input operation based on an operator's utterance or the like, or a change of the form screen.

Next, the operation of the form input system 11 will be described.
The form manager 6 and the operator manager 22 add, update, and delete data such as form meta information and keywords in advance before starting the call center operation. FIG. 11 is a sequence diagram showing the flow of data registration work. The data registration work shown in FIG. 11 is the same as the data registration work shown in FIG. 6, but the data registration work by the operator manager terminal 21 is different.
The form meta information input in step S1001 is transmitted to the operator manager terminal 21, and the operator manager terminal 21 stores data in which the meta information, the keyword, and the action are associated as the utterance determination database 402.

FIG. 12 is a diagram illustrating an example of meta information of a form, and FIG. 13 is a diagram illustrating an example of data in an utterance determination database.
The utterance determination data shown in FIG. 13 includes a form column ID of meta information, and a corresponding speaker (speaker of a keyword), a recognized keyword, and an event (action). An example of the meta information shown in FIG. 12 is described in XML. For example, in the case where the utterance determination data 835 shown in FIG. 13 is inputted with the form column ID “comment”, the utterance “Kramer” is recognized by the operator, It indicates that “add attention to the comment input field after notifying the operator administrator” is defined as the action.

  12 and 13, the meta information 801 shown in FIG. 12 corresponds to the utterance determination data 831 shown in FIG. 13, and the meta information 802 shown in FIG. 12 corresponds to the utterance determination data 832 shown in FIG. 12 corresponds to the utterance determination data 833 shown in FIG. 13, the meta information 804 shown in FIG. 12 corresponds to the utterance determination data 834 shown in FIG. 13, and the meta information 805 shown in FIG. Corresponding to the utterance determination data 835 shown in FIG. 13, the meta information 806 shown in FIG. 12 corresponds to the utterance determination data 836 shown in FIG.

FIG. 14 is a sequence diagram showing a flow of a form input operation by the form input system 11. The form input work by the form input system 11 is almost the same as the form input work by the form input system 1 shown in FIG. 7, but the work by the operator manager terminal 21 is added.
The voice recognition unit 102 of the server 3 recognizes the voice data (step S1202), and determines whether there is an utterance of the keyword stored in the recognition keyword database 204 and the utterance determination database 402 (step S3001).

  When there is an utterance of a keyword stored in advance in the recognition keyword database 204 and utterance determination database 402 (“Yes” in step S3001), the recognition event determination unit 103 corresponds to the keyword stored in the recognition keyword database 204 and utterance determination database 402. An event (operation) to be performed is selected (step S3002). Of these events, when the operator manager 22 is notified and the operator manager 22 needs to operate, the operator manager terminal 21 is notified of the event and the operator manager terminal 21 is instructed to input an operation or instruction from the operator manager terminal 21. Accept (step S3003). For example, here, when the operator manager 22 performs an operation on the form input, the form screen changing unit 104 instructs the change of the form screen or the change of the input field selection according to the operation (step S1206). Subsequent operations are the same as those in FIG.

FIG. 15 is a diagram illustrating an example of a form input screen. For example, when the operator 4 utters “You are a Kramer” during input work on the form shown in FIG. 15 and is recognized by the voice recognition unit 102, the recognition event determination unit 103 sends it to the operator manager terminal 21. It is possible to receive the notification and receive the comment input by the operator manager 22 and display it in the comment input field 911 shown in FIG.
For example, in this case, using the language model 301 associated with the comment input field 911 in the meta information language model correspondence database 302, the utterance of the operator manager 22 is recognized as voice data, and the recognition data is input to the form. It can be displayed.
Further, when the customer 8 utters “sue your company” to the operator 4 and is recognized by the voice recognition means 102, the combo in the form column “urgency” according to the utterance determination data 836 shown in FIG. 13. It is also possible to automatically enter “best” in the box.

  In the second embodiment, the operator manager 22 can input data for associating keywords and operations related to form input in the utterance between the customer 8 and the operator 4. Therefore, even if the operator 4 has little experience, the judgment and processing of the experienced operator manager 22 can be reflected in the form input work, the accuracy of the form input work is improved, and the work load on the operator 4 is also increased. Reduce.

It is also possible to create a program that realizes the form input process by the server 3 described above, and a general-purpose computer reads the program to realize the form input apparatus. This program may be recorded on a recording medium such as a CD-ROM, or distributed via a network.
The preferred embodiments of the form input system according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It will be apparent to those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea disclosed in the present application, and these naturally belong to the technical scope of the present invention. Understood.

  As described above, according to the present invention, the administrator prepares the input data template, the speech recognition keyword, and the language model in advance in the operation of creating the history data accompanied by the voice utterance, and creates the respective relationships. Thus, it can be applied to a purpose of simplifying input by using voice recognition without imposing a burden on general users. For example, when the meeting minutes need to be created, the meeting organizer registers the minutes template, language model and utterance keywords in advance, and creates the minutes by speech recognition from the utterances of the meeting participants It is also applicable to.

DESCRIPTION OF SYMBOLS 1 ......... Form input system 3 ......... Server 5 ......... Operator terminal 7 ......... Form manager terminal 9 ......... Customer terminal 101 ......... Voice input means 102 ......... Voice recognition means 103 ......... Recognition event determination means 104... Form screen change means 105... Form display means 106... Form input means 201... Form meta information database 202... Form database 203. ... recognition keyword database 301 ......... language model 302 ......... form meta information language model correspondence database 303 ......... language model selection means

Claims (6)

A form input device for supporting data input work on a form,
A form meta information language model compatible data storage means for associating the meta information of the form with a language model and storing it as form meta information language model compatible data;
A keyword data storage unit that associates the meta information of the form, the keyword, and the event information, and stores it as keyword data;
Voice recognition means for recognizing voice data;
Event execution means for executing an event based on the keyword data when the keyword is recognized in the audio data;
When a form to be input is changed by the event, based on the form meta information language model compatible data, a language model selection unit that selects a language model corresponding to the meta information of the changed form;
A form input means for determining input information of the form using the selected language model;
A form input device characterized by comprising: The event execution means, when a predetermined keyword is recognized, notifies the form input means of the limitation of input information,
The form input device according to claim 1, wherein the form input unit restricts input of input information of the form. The event execution means notifies the administrator when a predetermined keyword is recognized,
2. The form input device according to claim 1, wherein the form input unit recognizes the administrator's voice data using the selected language model and accepts it as input information of the form. An operator terminal capable of voice communication with the customer terminal;
A server capable of transmitting and receiving data via the operator terminal and a network;
A form input system having
The server
A form meta information language model compatible data storage means for associating the meta information of the form with the language model and storing the form meta information language model compatible data;
A keyword data storage unit that associates the meta information of the form, the keyword, and the event information, and stores it as keyword data;
Voice recognition means for recognizing voice data received from the customer terminal and the operator terminal;
Event execution means for executing an event based on the keyword data when the keyword is recognized in the audio data;
Means for changing a form to be input by the event and displaying it on the operator terminal;
Language model selection means for selecting a language model corresponding to the meta information of the changed form based on the form meta information language model correspondence data;
A form input means for determining input information of the form using the selected language model;
A form input system characterized by comprising: Associating the meta information of the form with the language model and storing the form as meta data corresponding to the form meta information language model;
Storing the form meta information, the keyword, and the event information in association with each other as keyword data;
Recognizing audio data;
If the keyword is recognized in the audio data, executing an event based on the keyword data;
When a form to be input is changed by the event, a step of selecting a language model corresponding to the meta information of the changed form based on the form meta information language model correspondence data;
Determining input information of the form using the selected language model;
A form input method characterized by comprising: Computer
A form meta information language model compatible data storage means for associating the meta information of the form with the language model and storing the form meta information language model compatible data;
A keyword data storage unit that associates the meta information of the form, the keyword, and the event information, and stores it as keyword data;
Voice recognition means for recognizing voice data;
Event execution means for executing an event based on the keyword data when the keyword is recognized in the audio data;
When a form to be input is changed by the event, based on the form meta information language model compatible data, a language model selection unit that selects a language model corresponding to the meta information of the changed form;
A form input means for determining input information of the form using the selected language model;
A program characterized by operating as

Images