JP2002259908A - Writing data processing system, writing data processing server, and writing data processing device - Google Patents

Abstract

(57) [Problem] To provide a handwritten data processing system, a handwritten data processing server, and a handwritten data processing device capable of correctly converting stroke data of a handwritten character input to a handwritten character input device into text data. . SOLUTION: Handwritten data input by handwriting with a pen 7 on a pressure-sensitive pad 6a of a handwriting character input tablet 6 and voice data input by a microphone 8 are transmitted via the Internet 3 via a client terminal device 5, Transmitted to the handwriting data processing server 2,
Then, text data is created based on the writing data and the voice data, and is returned to the client terminal device 5.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a writing data processing system, a writing data processing server, and a writing data processing device.

[0002]

2. Description of the Related Art Conventionally, there has been known a handwritten character input device for inputting stroke data of a handwritten input character on a pressure-sensitive pad using a pressure-sensitive pad for detecting a pressed position of XY coordinates. I have. In this handwritten character input device,
The stroke data of the character handwritten on the pressure-sensitive pad is stored in the storage device as it is, or the handwritten character is printed by the printing device. Also, it has been desired to convert stroke data of input characters into text data by character recognition.

[0003]

However, the current general handwritten character recognition technology has a problem that the accurate character recognition rate is about 80%, which is not suitable for practical use.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has a handwriting data processing system capable of correctly converting stroke data of handwritten characters input to a handwritten character input device into text data. It is an object to provide a data processing server and a writing data processing device.

[0005]

According to the first aspect of the present invention, there is provided a writing data processing system for achieving the above object.
In a handwriting data processing system in which a terminal device and a server are connected via a network, and the writing data transmitted from the terminal device is converted into text data by the server and returned to the terminal device, the terminal device includes a writing device. Writing data input means for inputting data, terminal-side transmitting means for transmitting the writing data input to the writing data input means to the server, and terminal-side receiving means for receiving text data from the server, The server includes a server-side receiving unit that receives the writing data sent from the terminal device, a writing data storage unit that stores the writing data received by the server-side receiving unit, and a server that is stored in the writing data storage unit. Display means for displaying written data;
Text data input means for inputting text data by an operator with reference to the handwritten data displayed on the display means, and text data transmission means for transmitting the text data input from the text data input means to the terminal device And

In the invention of this configuration, when handwritten character handwritten data is input to the handwritten data input means connected to the terminal device, the terminal-side transmitting means transmits the handwritten data to the server. In the server, the server-side receiving means receives the handwritten data, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the display means displays the handwritten data stored in the handwritten data storage means. The operator refers to the handwritten data displayed on the display means, inputs text data from the text data input means, the input text data is transmitted to the terminal device by the text data transmission means, and provided in the terminal device. The received terminal-side receiving means receives the text data from the server.

In the handwriting data processing system according to the second aspect of the present invention, the terminal device and the server are connected via a network, and the handwriting data transmitted from the terminal device is converted into text data by the server. In the handwriting data processing system for returning to the terminal device, the terminal device, handwriting data input means for inputting handwriting data,
A terminal-side transmitting unit that transmits the writing data input to the writing data input unit to the server, and a terminal-side receiving unit that receives text data from the server, wherein the server is transmitted from the terminal device. Server-side receiving means for receiving incoming handwritten data, and handwritten data storage means for storing the handwritten data received by the server-side receiving means,
Text data creation means for creating text data by character recognition of the handwritten data stored in the handwriting data storage means, text correction means for correcting the text data created by the text data creation means, and the text correction means Transmitting means for transmitting the corrected text data to the terminal device.

In this invention, when handwritten data of handwritten characters is inputted to the handwritten data input means connected to the terminal device, the terminal side transmitting means transmits the handwritten data to the server. The receiving means receives the writing data sent from the terminal device, the writing data storing means stores the writing data received by the server-side receiving means, and the text data creating means stores the writing data stored in the writing data storing means. The text correction unit corrects the text data generated by the text data generation unit, and the text data transmission unit transmits the text data corrected by the text correction unit to the terminal device. Then, the terminal side receiving means receives the text data from the server.

According to a third aspect of the present invention, in addition to the configuration of the second aspect, the server further includes a display unit, and the text data creation unit includes the display unit. The server-side receiving unit performs character recognition on the received handwritten data to create a character recognition candidate, and the display unit includes the handwritten data received by the server-side receiving unit and the character recognition candidate created by the text data creating unit. Is displayed.

[0010] In this invention, in addition to the operation of the invention according to claim 2, in the server, the text data creating means creates a character recognition candidate by character-recognizing the handwritten data received by the server-side receiving means. Then, the display means displays the handwritten data received by the server-side receiving means and the character recognition candidates created by the text data creating means.

In the writing data processing system according to a fourth aspect of the present invention, in addition to the configuration of the third aspect of the present invention, an operator refers to the writing data displayed on the display means, and The text correction unit can correct the character recognition candidate.

[0012] In the invention having this configuration, in addition to the effect of the invention described in claim 3, in the server, the operator refers to the handwritten data displayed on the display means, and the text correction means causes the operator to perform the operation. Character recognition candidates can be modified.

In the writing data processing system according to the present invention, the terminal device and the server are connected via a network, and the writing data transmitted from the terminal device is converted into text data by the server. In the handwriting data processing system for returning to the terminal device, the terminal device, handwriting data input means for inputting handwriting data,
Voice data input means for inputting voice at the time of writing data input, writing / voice data associating means for associating the writing data input from the writing data input means with the voice data input from the voice data input means, Terminal-side transmitting means for transmitting the writing data and voice data associated with the writing / voice data associating means to the server, and terminal-side receiving means for receiving text data from the server; and A server-side receiving means for receiving handwritten data and voice data sent from the terminal device, a handwritten data storage means for storing the handwritten data received by the server-side receiving means, Voice data storage means for storing the obtained voice data, and the handwritten data and the sound stored in the handwritten data storage means. Text data creating means for creating text data based on the voice data stored in the data storage means, and text data sending means for sending the text data created by the text data creating means to the terminal device. I have.

In the invention of this configuration, handwriting data of handwritten characters is input to the writing data input means connected to the terminal device, and voice at the time of writing data is input to the voice data input means. The associating means associates the writing data input from the writing data input means with the audio data input from the audio data input means, and the terminal-side transmitting means is associated with the writing / audio data associating means. Transmitting the writing data and the voice data to the server; In the server, the server-side receiving means receives the handwritten data and the voice data sent from the terminal device, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the voice data storage means is the server. The voice data received by the side receiving means is stored, and the text data creating means creates text data based on the handwritten data stored in the handwritten data storage means and the voice data stored in the voice data storage means, The text data transmitting means transmits the text data created by the text data creating means to the terminal device. On the terminal device side, the terminal side receiving means receives the text data from the server.

According to a sixth aspect of the present invention, in the handwriting data processing system, in addition to the configuration of the handwriting data processing system according to the fifth aspect, the text data creating means includes:
Character recognition candidate generation means for character-recognizing the writing data stored in the writing data storage means to generate character recognition candidates; and character recognition candidate accuracy for determining the certainty of the character recognition candidates generated by the character recognition candidate generation means. A calculation unit, a character recognition candidate generated by the character recognition candidate generation unit corresponding to the writing data, and the character recognition candidate generation unit corresponding to the writing data targeted when the character recognition candidate is generated. Character / speech data association means for associating a predetermined range of the voice data with the handwritten data, and speech in the predetermined range associated with the character recognition candidate by the character / speech data association means In the data, a degree calculating means for calculating a degree at which the character recognition candidate is uttered, and the character recognition candidate accuracy calculating means determine the degree. And certainty of serial character recognition candidate, the character recognition candidates said degree calculation means has calculated and a first finish the text data preparation means for preparing text data based on the degree to which the utterance.

In the invention having this configuration, in addition to the operation of the handwriting data processing system according to the fifth aspect, in the text data creating means, the character recognition candidate generating means converts the handwriting data stored in the handwriting data storage means into character data. Recognition and character recognition candidates are generated, the character recognition candidate accuracy calculation means obtains the certainty of the character recognition candidates generated by the character recognition candidate generation means, and the character / speech data associating means obtains a character corresponding to the writing data. The character recognition candidate generated by the recognition candidate generation means is associated with voice data in a predetermined range corresponding to writing data targeted when the character recognition candidate generation means generates the character recognition candidate. Calculating the degree to which the character recognition candidate is uttered in a predetermined range of voice data associated with the character recognition candidate by the voice data association means The first text-based text data creating means converts the text data based on the likelihood of the character recognition candidate calculated by the character recognition candidate accuracy calculation means and the degree of utterance of the character recognition candidate calculated by the degree calculation means. create.

Further, in the handwriting data processing system according to the present invention, in addition to the configuration of the handwriting data processing system according to the sixth aspect, the degree calculating means converts the voice data in the predetermined range into text data. A voice-to-text conversion means for converting; and a frequency calculation means for calculating a frequency at which the character recognition candidate is included in the text data converted by the voice-to-text conversion means, wherein the frequency calculated by the frequency calculation means is provided. Is the above-mentioned degree.

In the invention having this configuration, in addition to the operation of the handwriting data processing system according to claim 6, the voice / text conversion means of the degree calculation means converts voice data in a predetermined range into text data, The calculating means calculates the frequency at which the character recognition candidate is included in the text data converted by the voice / text converting means, and sets the frequency calculated by the frequency calculating means as the degree.

In the writing data processing system according to the present invention, in addition to the configuration of the writing data processing system according to the sixth aspect, the degree calculating means synthesizes a voice pattern of the character recognition candidate. Audio pattern synthesizing means, and a voice pattern synthesized by the audio pattern synthesizing means, provided with spotting means for determining the likelihood included in the audio data in the predetermined range, wherein the likelihood determined by the spotting means is The configuration is characterized by the degree.

In the invention having this configuration, in addition to the operation of the handwriting data processing system according to the sixth aspect, in the degree calculating means, the voice pattern synthesizing means synthesizes the voice pattern of the recognition candidate, and the spotting means generates the voice pattern. The likelihood that the voice pattern synthesized by the pattern synthesizing unit is included in the audio data in the predetermined range is determined, and the certainty determined by the spotting unit is defined as the degree.

According to a ninth aspect of the present invention, in addition to the configuration of the handwriting data processing system according to any one of the sixth to eighth aspects, the first clean-text data creating means may further comprise: Display means for displaying the character recognition candidates; character recognition candidate designating means for designating the character recognition candidates to be displayed by the display means; and determination of voice data corresponding to the character recognition candidates designated by the character recognition candidate designating means. There is provided a first reproducing means for reproducing the range, and a correcting means for allowing an operator to confirm the voice reproduced by the first reproducing means and correct the character recognition candidate.

According to the invention of this configuration, in addition to the operation of the writing data processing system according to any one of claims 6 to 8,
The display means of the first fair copy text data creating means includes:
The character recognition candidates are displayed, and the first reproducing means reproduces a predetermined range of the voice data corresponding to the character recognition candidates specified by the character recognition candidate designating means. After confirming the reproduced voice, the character recognition candidate is corrected.

According to a tenth aspect of the present invention, in addition to the configuration of the handwriting data processing system according to any one of the fifth to ninth aspects, the text data creating means includes a memory for storing the handwriting data. Writing data display means for displaying the writing data stored in the means, writing data indicating means for indicating the writing data displayed by the writing data display means, and voice corresponding to the writing data specified by the writing data indicating means. A second reproducing means for reproducing a predetermined range of data, and a second fine-text data creating means for creating text data while an operator checks the handwritten data displayed by the display means.

According to the invention of this configuration, in addition to the operation of the writing data processing system according to any one of claims 5 to 9,
The writing data display means of the text data creating means displays the writing data stored in the writing data storage means, the writing data indicating means indicates the writing data to be displayed by the writing data display means, and the second reproducing means A predetermined range of voice data corresponding to the handwritten data instructed by the handwritten data instructing means is reproduced, and text data is created by an operator while confirming the handwritten data displayed by the display means by the second copy-written text data creating means. can do.

In the writing data processing server according to the present invention, the terminal device is connected via a network, and the writing data transmitted from the terminal device is converted into text data and returned to the terminal device. A server for receiving handwritten data sent from the terminal device, a handwritten data storage for storing handwritten data received by the server-side receiving means, and a handwritten data storage Display means for displaying the writing data stored in the display means, text data input means for inputting text data by the operator with reference to the writing data displayed on the display means, and input from the text data input means. Text data transmitting means for transmitting text data to the terminal device.

In the invention of this configuration, the server-side receiving means receives the handwritten data, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the display means is stored in the handwritten data storage means. Displaying the writing data, the operator refers to the writing data displayed on the display means,
Input text data from text data input means,
The input text data is transmitted to the terminal device by the text data transmitting means.

According to a twelfth aspect of the present invention, the writing data processing server connects the terminal device via a network, converts the writing data transmitted from the terminal device into text data, and returns the text data to the terminal device. A server for receiving handwritten data sent from the terminal device, a handwritten data storage for storing handwritten data received by the server-side receiving means, and a handwritten data storage Text data creating means for creating text data by character recognition of the handwritten data stored in the text data, text correcting means for correcting the text data created by the text data creating means,
Clear text data transmitting means for transmitting the text data corrected by the text correcting means to the terminal device.

In the invention of this configuration, the server-side receiving means receives the handwritten data sent from the terminal device, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the text data creating means The character data of the handwritten data stored in the handwritten data storage means is recognized to create text data, the text correction means corrects the text data created by the text data creation means, and the clear text data transmission means is a text correction means. The corrected text data is transmitted to the terminal device.

According to a thirteenth aspect of the present invention, in addition to the configuration of the twelfth aspect, the writing data processing server further includes a display unit, and the text data creating unit includes the server-side receiving unit. Character recognition of the handwritten data received, to create a character recognition candidate, the display means,
The writing data received by the server-side receiving means and the character recognition candidates created by the text data creating means are displayed.

[0030] In the invention having this configuration, in addition to the effect of the twelfth aspect, the text data creating means performs character recognition on the handwritten data received by the server-side receiving means to create a character recognition candidate. The display means displays the handwritten data received by the server-side receiving means and the character recognition candidates created by the text data creating means.

Further, in the handwriting data processing server according to the fourteenth aspect, in addition to the configuration according to the thirteenth aspect, the operator refers to the handwriting data displayed on the display means, and The text correction unit can correct the character recognition candidate.

[0032] In the invention having this configuration, in addition to the operation of the invention of the thirteenth aspect, the operator refers to the handwritten data displayed on the display means and corrects the character recognition candidate by the text correction means. can do.

In the writing data processing server according to the present invention, the terminal device is connected via a network, and the writing data transmitted from the terminal device is converted into text data and returned to the terminal device. A writing data processing server for receiving writing data sent from the terminal device and audio data associated with the writing data, and storing the writing data received by the receiving device. Writing data storage means, voice data storage means for storing voice data received by the server-side receiving means, writing data stored in the writing data storage means and voice data stored in the voice data storage means Text data creating means for creating text data based on And a text data transmitting means for transmitting the text data to the terminal device.

In the invention of this configuration, the server-side receiving means receives the writing data and the voice data sent from the terminal device, and the writing data storage means stores the writing data received by the server-side receiving means, The voice data storage means stores the voice data received by the server-side receiving means, the text data creation means, based on the handwritten data stored in the handwritten data storage means and the voice data stored in the voice data storage means, The text data is created, and the text data sending unit sends the text data created by the text data creating unit to the terminal device.

In the writing data processing server of the invention according to claim 16, in addition to the configuration of the invention described in claim 15, the text data creating means includes a step of converting the writing data received by the server-side receiving means into text data. Character recognition candidate generating means for recognizing and generating a character recognition candidate, character recognition candidate accuracy calculating means for determining the certainty of the character recognition candidate generated by the character recognition candidate generating means, and the character corresponding to the writing data The character recognition candidate generated by the recognition candidate generation means and the voice data in a predetermined range corresponding to the writing data targeted when the character recognition candidate generation means generates the character recognition candidate, the writing data Character / speech data associating means for associating with the character recognition candidate by the character / speech data associating means. In the voice data of the range, a degree calculating means for calculating the degree at which the character recognition candidate is uttered, the likelihood of the character recognition candidate obtained by the character recognition candidate accuracy calculating means, and the degree calculating means calculating First text data creating means for creating text data based on the degree of utterance of the character recognition candidate.

In the invention having this configuration, in addition to the effect of the invention according to claim 15, in the text data creating means, the character recognition candidate creating means performs character recognition on the writing data stored in the writing data storage means. A character recognition candidate is generated, a character recognition candidate accuracy calculation unit obtains a certainty of the character recognition candidate generated by the character recognition candidate generation unit, and a character / speech data association unit generates a character recognition candidate corresponding to the writing data. The character recognition candidate generated by the means and the character recognition candidate generating means are associated with voice data in a predetermined range corresponding to the writing data targeted when the character recognition candidate is generated, and the degree calculating means comprises character / voice data. The degree to which the character recognition candidate is uttered is calculated in the voice data in a predetermined range associated with the character recognition candidate by the association means, and the first fair copy text is calculated. Strike data creating means creates the text data based on the degree to which the likelihood of the character recognition candidates character recognition candidates accuracy computing unit determined, the character recognition candidates degree calculation means has calculated is uttered.

[0037] In the handwriting data processing server according to the seventeenth aspect, in addition to the configuration according to the sixteenth aspect, the degree calculating means converts the voice data in the predetermined range into text data. A text conversion unit, and a frequency calculation unit for calculating a frequency at which the character recognition candidate is included in the text data converted by the voice-to-text conversion unit, wherein the frequency calculated by the frequency calculation unit is the degree The configuration is characterized by the following.

[0038] In the invention having this configuration, in addition to the effect of the invention according to claim 16, the voice / text conversion means of the degree calculation means converts voice data in a predetermined range into text data, and the frequency calculation means The frequency at which the character recognition candidate is included in the text data converted by the voice-to-text conversion means is calculated, and the frequency calculated by the frequency calculation means is defined as the degree.

In the writing data processing server according to the present invention, in addition to the configuration according to the sixteenth aspect, the degree calculation means may synthesize a voice pattern of the character recognition candidate. Means, a sound pattern synthesized by the sound pattern synthesizing means, spotting means for determining the likelihood included in the audio data in the predetermined range, and setting the likelihood determined by the spotting means to the degree. It has a characteristic configuration.

In the invention having this structure, in addition to the effect of the invention according to the sixteenth aspect, in the degree calculating means, the voice pattern synthesizing means synthesizes the voice pattern of the recognition candidate with the character, and the spotting means changes the voice pattern synthesizing means. The likelihood that the synthesized voice pattern is included in the voice data in the predetermined range is determined, and the likelihood determined by the spotting means is defined as the degree.

According to a nineteenth aspect of the present invention, in the writing data processing server, in addition to the configuration of the invention described in any one of the sixteenth to eighteenth aspects, the first clear-cut text data creating unit includes the character recognition unit. Display means for displaying candidates; character recognition candidate designating means for designating character recognition candidates displayed by the display means; and reproduction of a predetermined range of voice data corresponding to the character recognition candidates designated by the character recognition candidate designating means. And a correction means for correcting the character recognition candidate by the operator confirming the sound reproduced by the first reproduction means.

According to the invention having this structure, claims 16 to 18 are provided.
In addition to the operation described in any one of the above, the display means of the first copy-written text data creating means displays a character recognition candidate, and the first reproducing means displays the character recognition candidate designated by the character recognition candidate designating means. Then, a predetermined range of the voice data corresponding to the character data is reproduced, and the operator corrects the character recognition candidate by confirming the voice reproduced by the first reproducing means by the correcting means.

According to a twentieth aspect of the handwriting data processing server of the present invention, in addition to the configuration of the invention of any one of the fifteenth to nineteenth aspects, the text data creating means further comprises:
Writing data display means for displaying the writing data stored in the writing data storage means, writing data indicating means for indicating writing data displayed by the writing data display means, and writing instructed by the writing data indicating means. A second reproducing unit for reproducing a predetermined range of audio data corresponding to the data; and a second fine-text data generating unit for generating text data while an operator checks the handwritten data displayed by the display unit. ing.

According to the invention having this structure, claims 15 to 19 are provided.
In addition to the above-described operation, the writing data display means of the text data creating means displays the writing data stored in the writing data storage means, and the writing data indicating means displays the writing data displayed by the writing data display means. And the second reproducing means reproduces a predetermined range of audio data corresponding to the writing data specified by the writing data specifying means,
The text data can be created while the operator confirms the handwritten data displayed by the display means by the clear copy text data creating means.

According to a twenty-first aspect of the present invention, there is provided the handwriting data processing device, wherein the handwriting data input means for inputting handwriting data, the voice data input means for inputting voice data at the time of inputting the handwriting data, Writing / speech data associating means for associating the writing data input from the input means with the speech data input from the audio data inputting means; and the writing data and the sound data associated by the writing / speech data associating means. And text data creating means for creating text data based on the above.

According to the invention of this configuration, the writing data is input from the writing data input means, the voice data at the time of inputting the writing data is input from the voice data input means, and the writing data input means is input by the writing / voice data association means. The writing data input from the means and the voice data input from the voice data input means are associated with each other, and the text data creating means associates the writing data and the voice data associated with the writing / voice data associating means. Text data is created based on the text data.

In the handwriting data processing apparatus according to the twenty-second aspect, in addition to the configuration of the twenty-first aspect,
The text data creating means includes a character recognition candidate generating means for performing character recognition on the writing data input from the writing data input means to generate a character recognition candidate, and a character recognition candidate generated by the character recognition candidate generating means. Character recognition candidate accuracy calculation means for determining the likelihood, and calculating the degree to which the character recognition candidate is uttered in voice data in a predetermined range with respect to the time when the writing data corresponding to the character recognition candidate is input. A degree calculating unit; and a fair copy text data creating unit for creating text data based on the probability of the character recognition candidate and the degree calculated by the degree calculating unit.

In the invention having this structure, in addition to the effect of the invention described in claim 21, the text data creating means may include:
The character recognition candidate generating means generates a character recognition candidate by performing character recognition on the writing data input from the writing data input means, and the character recognition candidate accuracy calculating means determines the likelihood of the character recognition candidate generated by the character recognition candidate generating means. The degree calculating means calculates the degree of utterance of the character recognition candidate in the voice data in a predetermined range with respect to the time when the writing data corresponding to the character recognition candidate is input, Thus, text data is created based on the likelihood of the character recognition candidate and the degree calculated by the degree calculating means.

According to a twenty-third aspect of the present invention, in addition to the configuration of the twenty-second aspect, the handwriting data processing apparatus further comprises:
The degree calculating means, a voice recognition means for converting the voice data of the predetermined range to text data, a frequency calculation means for calculating the frequency of the character recognition candidate is included in the text data converted by the voice recognition means, And the frequency calculated by the frequency calculation means is set as the degree.

In the invention having this structure, in addition to the effect of the invention according to claim 22, in the degree calculating means, the voice recognizing means converts voice data in a predetermined range into text data, and the frequency calculating means comprises the voice recognizing means. Is calculated, and the frequency calculated by the frequency calculation means is defined as the degree.

According to a twenty-fourth aspect of the present invention, in addition to the configuration of the twenty-second aspect, the handwritten data processing apparatus further comprises
The degree calculating means includes a voice pattern synthesizing means for synthesizing the voice pattern of the character recognition candidate, and a spotting means for determining the likelihood that the synthesized voice pattern is included in the predetermined range of voice data, The certainty determined by the spotting means is set to the degree.

In the invention having this configuration, in addition to the effect of the invention according to claim 22, in the degree calculating means, the voice pattern synthesizing means synthesizes a voice pattern of a character recognition candidate,
The spotting means is configured to determine the likelihood that the synthesized voice pattern is included in the voice data in the predetermined range, and use the likelihood determined by the spotting means as the degree. .

According to a twenty-fifth aspect of the present invention, in addition to the constitutions of the twenty-first to twenty-fourth aspects, the text data creating means may include a character recognition program generated by the character recognition candidate generating means. Character / speech data associating means for associating candidate data with the voice data via handwritten data, display means for displaying the character recognition candidates, and instructing means for instructing the character recognition candidates to be displayed by the display means; A first reproducing unit for reproducing a predetermined range of voice data corresponding to the character recognition candidate specified by the specifying unit; and an operator confirming the voice reproduced by the first reproducing unit to check the character recognition candidate. And correcting means for correcting

In the invention having this configuration, in addition to the effects of the twenty-first to twenty-fourth aspects, in the text data creating means, the character / speech data associating means includes a character recognition candidate The data and the voice data are associated with each other via handwritten data, the display means displays a character recognition candidate, the instructing means instructs the character recognition candidate to be displayed by the display means, and the first reproducing means operates by the instructing means. A predetermined range of the voice data corresponding to the designated character recognition candidate is reproduced, and the operator corrects the character recognition candidate by checking the voice reproduced by the first reproducing means by the correction means.

In the writing data processing apparatus according to the twenty-sixth aspect, in addition to the configuration of the writing data processing apparatus according to any one of the twenty-first to twenty-fourth aspects, the text data creating means may include the writing data. Writing data display means for displaying the writing data input by the data input means, writing data indicating means for indicating the writing data to be displayed by the writing data display means, and writing data indicated by the writing data indicating means. A second reproduction unit for reproducing a predetermined range of the corresponding voice data; and a second fine-text data generation unit for generating text data while an operator checks the handwritten data displayed by the display unit. I have.

According to the invention having this configuration, in addition to the functions of the twenty-first to twenty-fourth aspects, the text data creating means includes:
When the writing data display means displays the writing data input by the writing data input means, and the writing data display means indicates the writing data to be displayed by the writing data display means,
The second reproducing means reproduces a predetermined range of voice data corresponding to the writing data instructed by the writing data instructing means, and creates the second clear text data while the operator confirms the writing data displayed by the display means. Create text data by means.

[0057]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a conceptual diagram of a writing data processing system 1 according to a first embodiment of the present invention. In the handwriting data processing system 1 of this embodiment, a client terminal device 5 is connected via the Internet 3 to a handwriting data processing server 2 installed in a predetermined company. Further, another client terminal device 9 and the like are also connected to the Internet 3. Client terminal device 5
Has a substantially rectangular pressure-sensitive pad 6 into which handwritten characters are input.
The tablet 6 for handwritten character input provided with a is connected.

Inside the pressure-sensitive pad 6a of the handwritten character input tablet 6, two conductive films having a resistance of several hundred ohms are stretched in parallel with the surface of the handwritten character input tablet 6 with a slight gap. Yes, pressure sensitive pad 6a
When the surface of the pen is pressed with the tip of the pen 7, the two conductive films come into contact with each other to read the X coordinate and the Y coordinate. The pressure-sensitive pad 6a is made of a transparent conductive film, and a liquid crystal display (not shown) having substantially the same size as the pressure-sensitive pad 6a is provided on the back side of the transparent conductive film. Characters input to the pad 6a are displayed and can be visually recognized from the pressure-sensitive pad 6a.

The handwriting character input tablet 6 is also provided with a microphone 8. The stroke data of the handwritten characters input to the handwritten character input tablet 6 is transmitted from the client terminal device 5 to the Internet 3.
Is transmitted to the handwriting data processing server 2 via.

Next, the structure of the writing data processing server 2 will be described with reference to FIG. FIG. 2 is a block diagram showing a device configuration of the writing data processing server 2.

As shown in FIG. 2, the writing data processing server 2 includes a CPU 2 for controlling the writing data processing server 2.
The CPU 20 includes a RAM 21 for temporarily storing various data, and a ROM for storing BIOS and the like.
22 and an I / O bus 23 for mediating data transfer
And are connected. A hard disk device 28 is connected to the I / O bus 23, and the hard disk device 28
, Stroke data of a handwritten character received from the client terminal device 5 (hereinafter referred to as “writing data”).
And a handwritten data storage area 2a for storing text data (hereinafter referred to as "written data") to be returned to the client terminal device 5.
8b, a character recognition dictionary storage area 28c for storing a character recognition dictionary for character recognition of handwritten data to create quality copy data, and a voice data storage area 28d for storing voice data received from the client terminal device 5. When,
A word dictionary storage area 28e storing a word dictionary, and another information storage area 28f storing other information;
A program storage area 28g for storing various programs executed by the handwriting data processing server 2 is provided.

The I / O bus 23 has a video controller 24, a key controller 25, and a CD-ROM
A drive 26, a LAN card 29, a sound card 33, and a mouse controller 35 are connected, a CRT 30 is connected to the video controller 24, a keyboard 31 is connected to the key controller 25, and a router 32 is connected to the LAN card 29. Is connected, and the router 32 is connected to the Internet 3. Further, a speaker 34 is connected to the sound card 33, and a mouse 36 is connected to the mouse controller 35. In addition, CD-R
The CD-ROM 27 inserted into the OM drive 26 includes:
Application programs, which are various programs executed by the writing data processing server 2 to realize the writing data processing system 1 of the present embodiment, are stored.
From the D-ROM 27, it is set up in the hard disk device 28 and stored in the program storage area 28g.

Next, the structure of the client terminal device 5 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a device configuration of the client terminal device 5.

As shown in FIG. 3, the client terminal device 5 is a known notebook personal computer.
The client terminal device 5 includes a client terminal device 5
Is provided, and the CPU 50 includes:
A RAM 51 for temporarily storing various data, and a BIO
A ROM 52 storing S and the like and an I / O bus 53 for mediating data transfer are connected. A hard disk device 58 is connected to the I / O bus 53. The hard disk device 58 has a handwritten data storage area 58a for storing handwritten data received from the handwritten character input tablet 6, and a handwritten data storage area 58a received from the handwritten data processing server 2. Data storage area 58b for storing the prepared data
And a voice data storage area 58c for storing voice data received from the handwriting character input tablet 6, and a program storage area 58d for storing various programs executed by the client terminal device 5.

The I / O bus 53 has a video controller 54, a key controller 55, and a CD-ROM
A drive 56, a LAN card 59, a sound card 67, and a mouse controller 65 are connected, an LCD 60 is connected to the video controller 54, a keyboard 61 is connected to the key controller 55, and a router 62 is connected to the LAN card 59. Is connected, and the router 62 is connected to the Internet 3. The microphone 8 provided in the handwritten character input tablet 6 is connected to the sound card 67. Further, a handwriting character input tablet 6 is connected to the I / O bus 53 so that handwriting data is input. Further, a sound card 67 is connected to the I / O bus 53, and the microphone 8 attached to the handwriting character input tablet 6 is connected to the sound card 67.

The CD-ROM 57 inserted into the CD-ROM drive 56 has various programs executed by the client terminal device 5 in order to implement the writing data processing system 1 of the present embodiment. The program is stored, and at the time of introduction, the application program is set up in the hard disk device 58 from the CD-ROM 57 and stored in the program storage area 58d.

Next, the operation of the writing data processing system 1 according to the first embodiment will be described with reference to FIGS. FIG. 4 is a flowchart of a process executed by the handwriting data processing server 2. FIG. 5 is a diagram showing a state in which handwriting characters are handwritten on the handwriting character input tablet 6, and FIG. 2 is a diagram showing a display screen of the CRT 30, and FIG. 7 is a diagram showing a display screen of the LCD 60 of the client terminal device 5.

First, as shown in FIG. 5, when a handwritten character is written on the pressure-sensitive pad 6a of the character input tablet 6 using the pen 7, the written data is converted into time-series stroke data on the client terminal device. 5, and is stored in a writing data storage area 58a of a hard disk device 58 via an I / O bus 53. In the client terminal device 5, the handwriting data storage area 58a
Is transmitted from the LAN card 59 to the handwritten data processing server 2 via the router 62 via the Internet 3 in chronological order.

The writing data processing server 2 performs processing according to the flowchart shown in FIG. First, when handwritten data, which is time-series stroke data, is sent from the client terminal device 5 via the router 32 via the Internet 3 (S1), the handwritten data is input to the LAN 29 and passed through the I / O bus 23. It is stored in the writing data storage area 28a of the hard disk device 28 (S3).

Next, the handwriting data stored in the handwriting data storage area 28a is displayed on the upper half of the divided screen of the CRT 30 (S5). Writing data processing server 2
The operator refers to the displayed handwritten data and inputs correct characters as the copy-written data from the keyboard 31, and the characters of the input copy-written data are displayed in the lower half of the divided screen of the CRT 30 (S7). ). The text data of the copy data is stored in the copy data storage area 28b of the hard disk device 28, and the operator of the writing data processing server 2 uses the mouse 36 to
When a transmission button (not shown) displayed at T30 is clicked, the handwritten data is transmitted from the writing data processing server 2 to the client terminal device 5 via the Internet 3 (S11).

In the client terminal device 5, the handwritten data received from the handwritten data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the handwritten data processing system 1 according to the first embodiment, handwritten data of handwritten characters input to the handwritten character input tablet 6 is processed via the Internet 3 by handwritten data processing. Transmitted to the server 2, the writing data processing server 2 allows the operator to input correct characters from the keyboard 31 while looking at the writing data of the handwritten characters displayed on the CRT 30.
Since the clear copy data composed of the text data is created and returned to the client terminal device 5 via the Internet 3, the operator of the tablet 6 for inputting handwritten characters can obtain the clean copy data composed of the correct text data.

Next, a writing data processing system 1 according to a second embodiment of the present invention will be described with reference to FIGS. 1 to 3 and FIGS. 5 to 8. FIG. 8 is a flowchart of a process performed by the handwriting data processing server 2 of the handwriting data processing system 1 according to the second embodiment. In the writing data processing system 1 of the second embodiment, the mechanical configurations of the writing data processing server 2 and the client terminal device 5 are the same as those of the first embodiment. The description of the mechanical configuration will be omitted with the help of the description.

First, as shown in FIG. 5, when a handwritten character is written on the pressure-sensitive pad 6a of the character input tablet 6 using the pen 7, the written data is converted into time-series stroke data on the client terminal device. 5, and is stored in a writing data storage area 58a of a hard disk device 58 via an I / O bus 53. In the client terminal device 5, the handwriting data storage area 58a
Is transmitted from the LAN card 59 to the handwritten data processing server 2 via the router 62 via the Internet 3 in chronological order.

The writing data processing server 2 performs processing according to the flowchart shown in FIG. First, when handwritten data, which is time-series stroke data, is transmitted from the client terminal device 5 via the router 32 via the Internet 3 (S21), the handwritten data is input to the LAN 29, and is transmitted through the I / O bus 23. It is stored in the writing data storage area 28a of the hard disk device 28 (S23).

Next, character recognition of the handwritten data stored in the handwritten data storage area 28a is performed (S25).
For this character recognition, an online recognition method is used. Specifically, based on the stroke data stored in the writing data storage area 28a, online recognition is performed with reference to the online recognition dictionary stored in the character recognition dictionary storage area 28c of the hard disk device 28 (S2).
5). More specifically, the CPU 20 compares the time-series data of strokes stored in the online recognition dictionary stored in the character recognition dictionary storage area 28c with the time-series data of strokes stored in the handwriting data storage area 28a. Characters are identified by comparison.

For example, as a result of performing online recognition in S25 in response to the input of “sales” of handwritten characters, a candidate called “sales” is generated as a first candidate, and “read” as a second candidate. When the candidate "up" is generated, the character code (text data) of the recognition candidate is used as the copy-written data as the copy-written data storage area 28
b (S25). Next, as shown in FIG. 6, the handwriting data stored in the handwriting data storage area 28a is displayed on the upper half of the divided screen of the CRT 30 (S26), and the recognition candidates stored in the clean copy data storage area 28b are displayed. The first candidate character is displayed in the lower half of the divided screen of the CRT 30 (S27).

At this time, if the operator of the handwriting data processing server 2 needs to correct the recognition candidate characters displayed in the lower half of the divided screen of the CRT 30, the mouse 3
If the user clicks a correction button (not shown) displayed on the CRT 30 using S6 (S29: YES), the character recognition candidate can be corrected (S31). In the correction of the character recognition candidate (S31), the next recognition candidate is displayed by designating the recognition candidate to be corrected with the pointer of the mouse 36 and pressing the return key (not shown) of the keyboard 31. By inputting a correct character from the keyboard 31, it can be corrected.

Next, using the mouse 36, the CRT 30
When the storage button (not shown) is clicked, the corrected character recognition data (correction data) is stored in the correction data storage area 28b of the hard disk drive 28 (S33). Note that even when there is no need for correction (S29:
NO), when a storage button (not shown) displayed on the CRT 30 is clicked using the mouse 36, character recognition data (clear copy data) is stored in the clear copy data storage area 28b of the hard disk drive 28.

Next, if the processing of all the words stored in the handwriting data storage area 28a has not been completed (S35: NO), the focus is moved to the next word (S39), and the aforementioned S25 is performed. To S33 are repeated.
When the processing of all the words stored in the handwriting data storage area 28a is completed (S35: YES), the handwritten data is transmitted from the handwriting data processing server 2 to the client terminal device 5 via the Internet 3. (S3
7).

In the client terminal device 5, the handwritten data received from the handwritten data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the handwritten data processing system 1 of the second embodiment, the handwritten data of the handwritten characters input to the handwritten character input tablet 6 is processed through the Internet 3 by the handwritten data processing system. The data is transmitted to the server 2, the character is recognized, further corrected by the operator, and the copy-written data is returned to the client terminal device 5 via the Internet 3. It is possible to obtain fair copy data composed of text data.

Next, FIGS. 1 to 3, FIGS. 5 to 7, and FIGS.
A description will be given of a writing data processing system 1 according to a third embodiment of the present invention with reference to FIG. FIG.
It is a flow chart of processing performed by writing data processing server 2 of processing system 1 of writing data of a 3rd embodiment.

In the writing data processing system 1 of the third embodiment, the mechanical configurations of the writing data processing server 2 and the client terminal device 5 are the same as those of the first embodiment. The description of the mechanical configuration is omitted with reference to the description of the first embodiment.

First, as shown in FIG. 5, using a pen 7 on the pressure-sensitive pad 6a of the tablet 6 for character input, the "business results", "sales", and "domestic 15,329 million yen" are used. , And "export 90.6 billion yen", the handwritten data is input to the client terminal device 5 as time-series stroke data, and is written via the I / O bus 53 to the handwritten data storage area 5 of the hard disk device 58.
8a. At this time, the writing data is stored in the writing data storage area 58a of the hard disk device 58 in association with the time data that the CPU 50 ticks in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You.

At the same time, the voice of the operator of the character input tablet 6 is input from the microphone 8 provided on the handwritten character input tablet 6 and is input to the sound card 67 of the client terminal device 5. / O bus 53
Are stored in the audio data storage area 58c of the hard disk device 58. At this time, the audio data is stored in the audio data storage area 58c of the hard disk device 58 in association with the time data ticked by the CPU 50 in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You. Therefore, the writing data and the audio data are respectively associated with the time data, so that the writing data and the audio data can be associated with each other.

For example, when the operator uses the pen 7 on the pressure-sensitive pad 6a of the character input tablet 6, the "business results", "sales", "domestic 15,329 million yen",
When writing handwritten characters as "export 90.6 billion yen"
"In terms of operating results, sales in the first half were 105,930 million yen, up 3.5% from the same period of the previous year, a record high in the first half. "In the beginning, it was a business result," said the sales increased by 11.4% to 15,329 million yen and export sales increased by 2.2% to 90.6 billion yen. However, sales in the first half increased 3.5% year-on-year to 1,05
930 million yen, a record high in the interim period. Domestic sales increased 11.4% year-on-year to
Export sales were ¥ 329 million, up 2.2% year on year to ¥ 90.6 billion. ] As audio data,
The audio data storage area 5 is associated with the time data.
8c.

In the client terminal device 5, the handwritten data (including time data) stored in the handwritten data storage area 58a and the voice data (including time data) stored in the voice data storage area 58c are stored in the LAN card 59. Via the Internet 3 via the router 62,
It is transmitted to the handwriting data processing server 2.

Next, in the handwriting data processing server 2, FIG.
Are performed according to the flowchart shown in FIG. First,
When writing data and voice data as time-series stroke data are transmitted from the client terminal device 5 via the router 32 via the Internet 3 (S4).
1), input to the LAN 29, and passed through the I / O bus 23,
Writing data storage area 28a of hard disk device 28
And stored in the audio data storage area 28d (S4
3).

Next, character recognition of the handwritten data stored in the handwritten data storage area 28a is performed (S45).
For this character recognition, an online recognition method is used. Specifically, based on the stroke data stored in the writing data storage area 28a, online recognition is performed with reference to the online recognition dictionary stored in the character recognition dictionary storage area 28c of the hard disk drive 28 (S4).
5). More specifically, the CPU 20 compares the time-series data of strokes stored in the online recognition dictionary stored in the character recognition dictionary storage area 28c with the time-series data of strokes stored in the handwriting data storage area 28a. Characters are identified by comparison, and character recognition data composed of text data is created. The result is displayed on the CRT 30 as shown in FIG.

Next, the character recognition data is associated with the time data (S47). In this process, first, the time data associated with the writing data as the basis of the character recognition data is associated with the character recognition data recognized from the writing data, and a character recognition data storage area (not shown) of the RAM 21 is stored. To memorize. This makes it possible to associate the character recognition data with the voice data via the time data.

Next, the candidate sentence is separated according to the time of the character recognition data (S49). As a result of the processing in S45, the character recognition result is recognized as "15,329 million yen in domestic business, 90.6 billion yen in export," and the portion where the stroke interval of the written data is long, that is, the time is opened. Then, the segmented part is determined as a sentence break and a delimiter process for separating the candidate sentences is performed. For example, “business results”, “reading”, “domestic 15,329 million yen”, and “export 90.6 billion yen”.

Next, morphological analysis is performed on the candidate sentences separated by the candidate sentence separation process in S49, and the candidate sentences are separated on a word-by-word basis (S51). For example, "management""performance","reading","domestic","15.329million",
It is divided into words such as "yen", "export", "90.6 billion", and "yen". At this time, for example, the first candidate of the word candidate corresponding to the written data “sales” may be “reading” and the second candidate may be “sales”.

Next, in the processing of S51, voice data (voice data corresponding to the word) corresponding to the time data of each word divided into words is converted from voice data to text data for 30 seconds before and after the word. Is performed (S53). This conversion is performed using a well-known voice recognition technology. Then
A word candidate frequency calculation process is performed (S55). For example, the first word candidate corresponding to the written data “sales” is “reading”, and the second candidate is a text conversion of voice data for 30 seconds before and after “sales”. However, sales in the first half of the year increased 3.5% year-on-year to 105,930 million yen, a record high in the first half, and domestic sales were 11.4 year-on-year.
Export sales increased 2.2% year on year to 90.6 billion yen. In the converted text, the frequency of the word "reading" is "0", but the frequency of the word "sales" is "4". Become.

Next, a process of creating a fair copy text data is performed (S57). In this process, a word candidate is determined based on the probability of character recognition and the appearance frequency of the word candidate obtained in S55. In the above example, the first word candidate is “reading” and the second word candidate is “sales”.
The appearance frequency of “reading” is “0” and the appearance frequency of “sales” is “4” in the part of the speech data converted from text data of seconds and converted to text. Is determined as the clean copy data, and is stored in the clean copy data storage area 28b of the hard disk device 28.

In the process of S55, the following expression is used.
Which word candidate is adopted can be determined.
That is, the probability value of the certainty of kanji recognition is p (x),
A large value obtained by multiplying p (x) by the ratio of the appearance frequency is adopted as a word candidate. In the above example, it is as follows. "Reading": (p (reading) + p (up)) x 0/4 "Sales": (p (selling) + p (up)) x 4/4 In the above example, "reading" is "0". So that
“Sales” is adopted as a word candidate.

Next, if the processing of all the words stored in the handwriting data storage area 28a has not been completed (S59: NO), the focus is moved to the next word (S63) and the above-mentioned S53 is performed. To S57 are repeated.
When the processing of all the words stored in the handwriting data storage area 28a is completed (S59: YES), the handwritten data is transmitted from the handwriting data processing server 2 to the client terminal device 5 via the Internet 3. (S6
1).

In the client terminal device 5, the handwritten data received from the handwriting data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the handwritten data processing system 1 according to the third embodiment, handwritten data of handwritten characters input to the handwritten character input tablet 6 is processed via the Internet 3 by handwritten data processing. Since the data is transmitted to the server 2, the characters are recognized, and the fairly-written data is created with reference to the frequency of occurrence of the word candidate in the voice data, and the fairly-written data is returned to the client terminal device 5 via the Internet 3. The operator of the tablet 6 for inputting handwritten characters can obtain the fair copy data including the correct text data.

Next, FIG. 1 to FIG. 3, FIG. 5 to FIG.
A description will be given of a writing data processing system 1 according to a fourth embodiment of the present invention with reference to FIG. FIG.
9 is a flowchart of a process performed by the handwriting data processing server 2 of the handwriting data processing system 1 according to the fourth embodiment.

In the writing data processing system 1 of the fourth embodiment, the mechanical configurations of the writing data processing server 2 and the client terminal device 5 are the same as those of the first embodiment. The description of the mechanical configuration is omitted with reference to the description of the first embodiment.

First, as shown in FIG. 5, using a pen 7 on a pressure-sensitive pad 6a of a tablet 6 for character input, the "business results", "sales", and "domestic 152.9 billion yen" are used. , And "export 90.6 billion yen", the handwritten data is input to the client terminal device 5 as time-series stroke data, and is written via the I / O bus 53 to the handwritten data storage area 5 of the hard disk device 58.
8a. At this time, the writing data is stored in the writing data storage area 58a of the hard disk device 58 in association with the time data that the CPU 50 ticks in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You.

At the same time, the voice of the operator of the character input tablet 6 is input from the microphone 8 provided on the handwritten character input tablet 6 and is input to the sound card 67 of the client terminal device 5. / O bus 53
Are stored in the audio data storage area 58c of the hard disk device 58. At this time, the audio data is stored in the audio data storage area 58c of the hard disk device 58 in association with the time data ticked by the CPU 50 in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You. Therefore, the writing data and the audio data are respectively associated with the time data, so that the writing data and the audio data can be associated with each other.

For example, when the operator uses the pen 7 on the pressure-sensitive pad 6a of the character input tablet 6, the "business results", "sales", "domestic 15,329 million yen",
When writing handwritten characters as "export 90.6 billion yen"
"In terms of operating results, sales in the first half were 105,930 million yen, up 3.5% from the same period of the previous year, a record high in the first half. "In the beginning, it was a business result," said the sales increased by 11.4% to 15,329 million yen and export sales increased by 2.2% to 90.6 billion yen. However, sales in the first half increased 3.5% year-on-year to 1,05
930 million yen, a record high in the interim period. Domestic sales increased 11.4% year-on-year to
Export sales were ¥ 329 million, up 2.2% year on year to ¥ 90.6 billion. ] As audio data,
The audio data storage area 5 is associated with the time data.
8c.

Then, in the client terminal device 5, the handwritten data (including time data) stored in the handwritten data storage area 58a and the voice data (including time data) stored in the voice data storage area 58c are stored in the LAN card 59. Via the Internet 3 via the router 62,
It is transmitted to the handwriting data processing server 2.

In the handwriting data processing server 2, processing according to the flowchart shown in FIG. 11 is performed. First, from the client terminal device 5 via the Internet 3,
When writing data and voice data, which are time-series stroke data, are sent via the router 32 (S71), L
The data is input to the AN 29 and is stored in the writing data storage area 28a and the audio data storage area 28d of the hard disk drive 28 via the I / O bus 23 (S73).

Next, character recognition of the handwritten data stored in the handwritten data storage area 28a is performed (S75).
For this character recognition, an online recognition method is used. Specifically, based on the stroke data stored in the writing data storage area 28a, online recognition is performed with reference to the online recognition dictionary stored in the character recognition dictionary storage area 28c of the hard disk drive 28 (S7).
5). More specifically, the CPU 20 compares the time-series data of strokes stored in the online recognition dictionary stored in the character recognition dictionary storage area 28c with the time-series data of strokes stored in the handwriting data storage area 28a. Characters are identified by comparison, and character recognition data composed of text data is created. The result is displayed on the CRT 30 as shown in FIG.

Next, the character recognition data is associated with the time data (S77). In this process, first, the time data associated with the writing data as the basis of the character recognition data is associated with the character recognition data recognized from the writing data, and a character recognition data storage area (not shown) of the RAM 21 is stored. To memorize. This makes it possible to associate the character recognition data with the voice data via the time data.

Next, the candidate sentence is separated according to the time of the character recognition data (S79). By the processing of S75, the character recognition result is recognized as "15,329 million yen in domestic export 900.6 billion yen in business performance reading", and the portion where the stroke interval of the written data is long, that is, the time is opened. Then, the segmented part is determined as a sentence break and a delimiter process for separating the candidate sentences is performed. For example, “business results”, “reading”, “domestic 15,329 million yen”, and “export 90.6 billion yen”.

Next, morphological analysis is performed on the candidate sentences separated by the candidate sentence segmentation process in S79, and the candidate sentences are segmented into words (S81). For example, "management""performance","reading","domestic","15.329million",
It is divided into words such as "yen", "export", "90.6 billion", and "yen". At this time, for example, the first candidate of the word candidate corresponding to the written data “sales” may be “reading” and the second candidate may be “sales”.

Next, in the processing of S81, the speech parameters (speech data) of each word divided in word units are synthesized (characters are sounded) (S83). This speech parameter is a well-known HMM (Hidden Markov Mo
del) Synthesize Tri-phone. For example,
For "sales", "u + r u-r + i + a i-a + g a-g + e ge"
Is synthesized with

Next, a composite parameter word is extracted (S85). Specifically, in the process of S81, word spotting for comparing sounds is performed to determine whether or not there is the same voice parameter during 30 seconds before and after voice data corresponding to the candidate word obtained by synthesizing the voice parameter (voice data). Do. For example, the first candidate of the word candidate corresponding to the written data "sales" is "reading", the second candidate is "sales", and the voice data for 30 seconds before and after that is "firstly the business results, In the first half, sales increased 3.5% year-on-year to 105,930 million yen, a record high in the first half, and domestic sales were 11.4% year-on-year.
Sales increased by 15,329 million yen, and export sales increased 2.2% to 90.6 billion yen. Means that the sound "sales" exists four times in this audio data. Here, when word spotting for comparing sounds with sounds is performed, the word spotting probability pw (sales) for “sales” is higher than the word spotting probability pw (sales) for “speaking”.

Next, a process of creating a fair copy text data is performed (S87). In this process, word candidates are determined based on the certainty of character recognition and the probability of word spotting obtained in S85. In the above example, the first word candidate is “reading” and the second candidate is “sales”, but the word spotting probability pw (sales) in the audio data for 30 seconds before and after is higher. Is higher than the word spotting probability pw (reading), the “sales” is determined as fair copy data, and stored in the hard copy data storage area 28 b of the hard disk drive 28.

In the process of S87, the following expression is used.
Which word candidate is adopted can be determined.
That is, the probability of character recognition is p (x), the probability of word spotting is pw (x), and pw (x) is pw (x).
The one with the larger value of (x) is adopted as a word candidate. In the above example, it is as follows. “Reading”: (p (reading) + p (upper)) × pw (reading) “sales”: (p (selling) + p (upper)) × pw (sales) In the above example, for example, (p ( The value of (reading) + p (up)) is “1.0”, the value of (p (sell) + p (up)) is “0.8”, and the value of pw (reading) is “0. 1 ", and if the value of pw (sales) is" 0.5 ", (p (reading) + p (upper)) x pw (reading) = 0.1 (p (sell) + p (up) ) × pw (sales) = 0.4, and “sales” is adopted as a word candidate.

Next, if the processing of all the words stored in the writing data storage area 28a has not been completed (S89: NO), the focus is moved to the next word (S93), and the above-mentioned S83 is performed. To S87 are repeated.
When the processing of all the words stored in the handwriting data storage area 28a is completed (S89: YES), the handwritten data is transmitted from the handwriting data processing server 2 to the client terminal device 5 via the Internet 3. (S9
1).

In the client terminal device 5, the handwritten data received from the handwriting data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the handwritten data processing system 1 according to the fourth embodiment, handwritten data of handwritten characters input to the handwritten character input tablet 6 is processed via the Internet 3 by handwritten data processing. The data is transmitted to the server 2, the characters are recognized, the word candidates are converted into voice data, word spotting with the voice data is performed, and the fair copy data is created. Since it is returned to the client terminal device 5, the operator of the tablet 6 for inputting handwritten characters can obtain fair copy data composed of correct text data.

Next, FIGS. 1 to 3, FIGS. 5 to 7, and FIGS.
With reference to FIG. 0 and FIG. 12, a writing data processing system 1 according to a fifth embodiment of the present invention will be described. FIG.
11 is a flowchart of a process performed by a handwriting data processing server 2 of a handwriting data processing system 1 according to a fifth embodiment.

In the writing data processing system 1 of the fifth embodiment, the mechanical configurations of the writing data processing server 2 and the client terminal device 5 are the same as those of the first embodiment. The description of the mechanical configuration is omitted with reference to the description of the first embodiment.

First, as shown in FIG. 5, using a pen 7 on the pressure-sensitive pad 6a of the tablet 6 for character input, the "business results", "sales", and "domestic 15,329 million yen" are used. , And "export 90.6 billion yen", the handwritten data is input to the client terminal device 5 as time-series stroke data, and is written via the I / O bus 53 to the handwritten data storage area 5 of the hard disk device 58.
8a. At this time, the writing data is stored in the writing data storage area 58a of the hard disk device 58 in association with the time data that the CPU 50 ticks in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You.

At the same time, the voice of the operator of the character input tablet 6 is input from the microphone 8 provided on the handwritten character input tablet 6, and is input to the sound card 67 of the client terminal device 5, and / O bus 53
Are stored in the audio data storage area 58c of the hard disk device 58. At this time, the audio data is stored in the audio data storage area 58c of the hard disk device 58 in association with the time data ticked by the CPU 50 in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You. Therefore, the writing data and the audio data are respectively associated with the time data, so that the writing data and the audio data can be associated with each other.

For example, when the operator uses the pen 7 on the pressure-sensitive pad 6a of the character input tablet 6, the "business results", "sales", "domestic 15,329 million yen",
When writing handwritten characters as "export 90.6 billion yen"
"In terms of operating results, sales in the first half were 105,930 million yen, up 3.5% from the same period of the previous year, a record high in the first half. "In the beginning, it was a business result," said the sales increased by 11.4% to 15,329 million yen and export sales increased by 2.2% to 90.6 billion yen. However, sales in the first half increased 3.5% year-on-year to 1,05
930 million yen, a record high in the interim period. Domestic sales increased 11.4% year-on-year to
Export sales were ¥ 329 million, up 2.2% year on year to ¥ 90.6 billion. ] As audio data,
The audio data storage area 5 is associated with the time data.
8c.

In the client terminal device 5, the handwritten data (including time data) stored in the handwritten data storage area 58a and the voice data (including time data) stored in the voice data storage area 58c are stored in the LAN card 59. Via the Internet 3 via the router 62,
It is transmitted to the handwriting data processing server 2.

Next, in the handwriting data processing server 2, FIG.
Processing according to the flowchart shown in FIG. 2 is performed. First, when writing data and voice data, which are time-series stroke data, are transmitted from the client terminal device 5 via the Internet 3 via the router 32 (S1).
01), is input to the LAN 29, passes through the I / O bus 23, and passes through the handwriting data storage area 2 of the hard disk device 28.
8a and the voice data storage area 28d (S
103).

Next, the display of the handwritten data (handwritten characters) stored in the handwritten data storage area 28a is performed on the upper half of the screen of the CRT 30 as shown in FIG. 6 (S1).
05). When an unknown character in the display of the handwritten data is indicated by the pointer of the mouse 36 (S10).
7) The voice data corresponding to the writing data is
0, reproduced by the sound card 33,
A sound is generated from the speaker 34 (S109). When the operator of the handwritten data processing server 2 listens to the reproduced voice and inputs the correct characters as the copy data from the keyboard 31, the characters of the input copy data are displayed in the lower half of the divided screen of the CRT 30. (S11
1). The text data of this copy data is stored in the copy data storage area 28b of the hard disk device 28,
When the operator of the writing data processing server 2 clicks a transmission button (not shown) displayed on the CRT 30 using the mouse 36, the handwritten data is transmitted from the writing data processing server 2 to the client terminal device 5 via the Internet 3. It is transmitted (S115).

In the client terminal device 5, the handwritten data received from the handwriting data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the writing data processing system 1 according to the fifth embodiment, the writing data of the handwritten characters input to the handwritten character inputting tablet 6 is processed via the Internet 3 by the handwriting data processing system. Transmitted to the server 2, the writing data processing server 2 allows the operator to input correct characters while listening to the voice data on the keyboard 3.
1 to create the clear copy data composed of the text data and return it to the client terminal device 5 via the Internet 3. Therefore, the operator of the tablet 6 for inputting handwritten characters can make the clean copy data composed of the correct text data. Can be obtained.

Next, FIGS. 1 to 3, FIGS. 5 to 7, and FIGS.
A description will be given of a writing data processing system 1 according to a sixth embodiment of the present invention with reference to FIG. FIG.
11 is a flowchart of a process performed by a writing data processing server 2 of a writing data processing system 1 according to a sixth embodiment.

In the writing data processing system 1 of the sixth embodiment, the mechanical configurations of the writing data processing server 2 and the client terminal device 5 are the same as those of the first embodiment. The description of the mechanical configuration is omitted with reference to the description of the first embodiment.

First, as shown in FIG. 5, using a pen 7 on the pressure-sensitive pad 6a of the tablet 6 for character input, the "business results", "sales", and "domestic 15,329 million yen" are used. , And "export 90.6 billion yen", the handwritten data is input to the client terminal device 5 as time-series stroke data, and is written via the I / O bus 53 to the handwritten data storage area 5 of the hard disk device 58.
8a. At this time, the writing data is stored in the writing data storage area 58a of the hard disk device 58 in association with the time data that the CPU 50 ticks in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You.

At the same time, the voice of the operator of the character input tablet 6 is input from the microphone 8 provided on the handwritten character input tablet 6, input to the sound 67 of the client terminal device 5, and The audio data storage area 58c of the hard disk device 58 via the O bus 53
Is stored. At this time, the audio data is stored in the audio data storage area 58c of the hard disk device 58 in association with the time data ticked by the CPU 50 in synchronization with a clock signal generated by a clock generation circuit (not shown) of the client terminal device 5. You. Therefore, the writing data and the audio data are respectively associated with the time data, so that the writing data and the audio data can be associated with each other.

For example, when the operator uses the pen 7 on the pressure-sensitive pad 6a of the character input tablet 6, the "business results", "sales", "domestic 15,329 million yen",
When writing handwritten characters as "export 90.6 billion yen"
"In terms of operating results, sales in the first half were 105,930 million yen, up 3.5% from the same period of the previous year, a record high in the first half. "In the beginning, it was a business result," said the sales increased by 11.4% to 15,329 million yen and export sales increased by 2.2% to 90.6 billion yen. However, sales in the first half increased 3.5% year-on-year to 1,05
930 million yen, a record high in the interim period. Domestic sales increased 11.4% year-on-year to
Export sales were ¥ 329 million, up 2.2% year on year to ¥ 90.6 billion. ] As audio data,
The audio data storage area 5 is associated with the time data.
8c.

In the client terminal device 5, the handwritten data (including time data) stored in the handwritten data storage area 58a and the voice data (including time data) stored in the voice data storage area 58c are stored in the LAN card 59. Via the Internet 3 via the router 62,
It is transmitted to the handwriting data processing server 2.

In the handwriting data processing server 2, processing according to the flowchart shown in FIG. 13 is performed. First, from the client terminal device 5 via the Internet 3,
When writing data and voice data, which are time-series stroke data, are sent via the router 32 (S121),
The data is input to the LAN 29 and is stored in the writing data storage area 28a and the audio data storage area 28d of the hard disk drive 28 via the I / O bus 23 (S123).

Next, character recognition of the handwritten data stored in the handwritten data storage area 28a is performed (S12).
5). For this character recognition, an online recognition method is used. Specifically, based on the stroke data stored in the writing data storage area 28a, online recognition is performed with reference to an online recognition dictionary stored in the character recognition dictionary storage area 28c of the hard disk drive 28 (S125). . More specifically, the CPU 20 compares the time-series data of strokes stored in the online recognition dictionary stored in the character recognition dictionary storage area 28c with the time-series data of strokes stored in the handwriting data storage area 28a. Characters are identified by comparison, and character recognition data composed of text data is created.

Next, the character recognition data is associated with the time data (S127). In this process, first, the time data associated with the writing data as the basis of the character recognition data is associated with the character recognition data recognized from the writing data, and a character recognition data storage area (not shown) of the RAM 21 is stored. To memorize. This makes it possible to associate the character recognition data with the voice data via the time data. Next, the character recognition candidate is displayed on the CRT 30 as shown in FIG. 10 (S129).

Next, when a character of an unknown part among the character recognition candidates displayed on the screen of the CRT 30 is indicated by the pointer of the mouse 36 (S131), the voice data corresponding to the character recognition candidate is controlled by the CPU 50. The sound is reproduced by the sound card 33 and output from the speaker 34 (S133). For example, as shown in FIG.
If the operator recognizes that the character recognition of the part displayed as "reading" on the screen 30 is wrong, the operator indicates "reading" with the pointer of the mouse 36 (S113).
1) The sound data corresponding to the recognition candidate "reading" is controlled by the CPU 50 for 30 seconds before and after the sound card 33.
And is reproduced from the speaker 34 (S1).
33).

If the recognition candidate "reading" is to be corrected (S135: YES), the operator inputs the correct character "sales", which is the fair copy data, from the keyboard 31 and corrects it (S137). Then, the text data of the handwritten data is stored in the handwritten data storage area 28b of the hard disk device 28 (S139), and the operator of the handwritten data processing server 2 uses the mouse 36 to
When you click the send button (not shown) displayed in 0,
The fair copy data is transmitted from the handwriting data processing server 2 to the client terminal device 5 via the Internet 3 (S141).

In the client terminal device 5, the handwritten data received from the handwriting data processing server 2 is stored in the handwritten data storage area 58b of the hard disk device 58 and displayed on the LCD 60 as shown in FIG.

As described above, in the writing data processing system 1 according to the sixth embodiment, the writing data of the handwritten characters input to the handwritten character inputting tablet 6 is processed via the Internet 3 by the writing data processing. After being transmitted to the server 2 and being subjected to the character recognition processing in the writing data processing server 2, the operator inputs a correct character to the keyboard 31 while listening to the voice data for a strange character recognition portion.
To create a copy-written data composed of text data and return it to the client terminal device 5 via the Internet 3.
Can obtain the fair copy data composed of the correct text data.

Note that the present invention is not limited to the above-described embodiment, and it is needless to say that the present invention can be implemented in various modes without departing from the gist of the present embodiment. For example,
In the above-described embodiment, the tablet 6 for inputting handwritten characters is connected to the client terminal device 5 and connected to the writing data processing server 2 via the Internet 3 so as to create the clear-cut text data. As shown in FIG. 14, a handwriting character input tablet 137 may be built in the handwriting data processing device 200.

More specifically, the writing data processing device 200 includes a CPU 12 that controls the writing data processing device 200.
0, the CPU 120 has a RAM 121 for temporarily storing various data, and a ROM for storing BIOS and the like.
122 and an I / O bus 1 that mediates data transfer
23, and a hard disk device 128 is connected to the I / O bus 123.
The handwriting data storage area 12 stores stroke data of handwritten characters received from the handwriting character input tablet 137 (hereinafter referred to as “writing data”).
8a and a fair copy data storage area 1 for storing fair copy data
28b, a character recognition dictionary storage area 128c for storing a character recognition dictionary for character recognition of handwritten data to create quality copy data, and a voice data storage area 128d for storing voice data input from the microphone 138. , A word dictionary storage area 128e storing a word dictionary, and another information storage area 128f storing other information
And a program storage area 128g in which various programs executed by the handwriting data processing device 200 are stored.

The I / O bus 123 has a video controller 124, a key controller 125, and a CD-
The ROM drive 126, the sound card 133, and the mouse controller 135 are connected, the video controller 124 is connected to the CRT 130, the key controller 125 is connected to the keyboard 131, and the sound card 133 is connected to the speaker 134 and the microphone. 138
Is connected, and the mouse controller 135
36 are connected. The CD-ROM drive 1
The CD-ROM 127 inserted into the CD-ROM 26 stores application programs that are various programs executed by the writing data processing device 200. When the application program is installed, the application program is stored in the CD-ROM 12
7, the program is set up in the hard disk drive 128 and stored in the program storage area 128g.

The processing of the flowcharts shown in FIGS. 9, 11, 12, and 13 may be performed by an integrated structure such as the writing data processing device 200. In this case, accurate handwritten data can be obtained from the handwritten data by the handwritten data processing device 200 as one device. Needless to say, the CRT 30 connected to the handwriting data processing server 2 can display any of handwriting data, recognition candidate data, and clean copy data.

[0146]

As described in detail above, in the handwriting data processing system according to the first aspect of the present invention, when handwriting data of handwritten characters is input to the handwriting data input means connected to the terminal device, the data is transmitted to the terminal side. Means sends the writing data to the server. In the server, the server-side receiving means receives the handwritten data, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the display means displays the handwritten data stored in the handwritten data storage means, The operator refers to the handwritten data displayed on the display means, inputs text data from the text data input means, and the input text data is transmitted to the terminal device by the text data transmission means, and is provided in the terminal device. The terminal-side receiving means receives the text data from the server. Therefore, in the terminal device, the operator of the server can obtain the accurate text data in which the text data has been input from the text data input means with reference to the handwritten data displayed on the display means.

Further, in the handwriting data processing device according to the second aspect of the present invention, when handwriting data of handwritten characters is input to handwriting data input means connected to the terminal device, the terminal side transmitting means transmits the handwriting data to the server. To the server,
The server-side receiving means receives the handwritten data sent from the terminal device, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the text data creating means is stored in the handwritten data storage means. The text data is created by character recognition of the handwritten data, the text correction means corrects the text data created by the text data creation means, and the copy-written text data transmission means sends the text data corrected by the text correction means to the terminal device. Send,
The terminal-side receiving means provided in the terminal device can receive the text data from the server. Therefore, in the terminal device, it is possible to obtain accurate text data whose characters are recognized by the server and corrected by the text correction unit.

In the handwriting data processing apparatus according to the third aspect of the present invention, in addition to the effect of the second aspect, in the server, the text data creating means may be configured to write the handwritten data received by the server-side receiving means. The character recognition is performed to create a character recognition candidate, and the display unit can display the handwritten data received by the server-side receiving unit and the character recognition candidate created by the text data creating unit. Therefore, the server operator can compare the handwritten data displayed on the display means with the character recognition candidates.

In the handwriting data processing system according to the fourth aspect of the present invention, in addition to the effect of the third aspect, in the server, the operator refers to the handwriting data displayed on the display means. The text correction means
The character recognition candidate can be modified.

Further, in the handwriting data processing system according to the fifth aspect of the present invention, handwriting data of handwritten characters is inputted to the handwriting data input means connected to the terminal device, and the voice data input means receives the handwriting data when the handwriting data is inputted. The voice is input,
The writing / voice data associating means associates the writing data input from the writing data input means with the audio data input from the voice data input means, and the terminal-side transmitting means uses the writing / voice data associating means. The associated writing data and voice data are transmitted to the server. In the server, the server-side receiving means receives the handwritten data and the voice data sent from the terminal device, the handwritten data storage means stores the handwritten data received by the server-side receiving means, and the voice data storage means is the server. The side receiving means stores the received voice data, the text data creating means, based on the handwritten data stored in the handwritten data storage means and the voice data stored in the voice data storage means,
Create text data, and the text data transmission means:
The text data created by the text data creating means is transmitted to the terminal device, and the terminal-side receiving means can receive the text data from the server on the terminal device side. Therefore, in the terminal device, it is possible to obtain accurate text data created based on the writing data and the voice data in the server.

In the writing data processing system of the invention according to claim 6, in addition to the effect of the writing data processing system according to claim 5, in the text data creating means, the character recognition candidate generating means includes a writing data storage means. A character recognition candidate is generated by character recognition of the handwritten data stored in the means, and a character recognition candidate accuracy calculation means obtains a certainty of the character recognition candidate generated by the character recognition candidate generation means, and a character / speech data association means Is a character recognition candidate generated by the character recognition candidate generating means corresponding to the writing data, and a predetermined range of voice data corresponding to the writing data targeted when the character recognition candidate generating means generates the character recognition candidate, The associating and degree calculating means includes a character recognition unit that stores the character recognition in the voice data in a predetermined range associated with the character recognition candidate by the character / speech data associating means. The degree of utterance of the candidate is calculated, and the first copy-edited text data creating means utters the certainty of the character recognition candidate obtained by the character recognition candidate accuracy calculating means and the character recognition candidate calculated by the degree calculating means. The text data can be created based on the degree of occurrence. Therefore, in the terminal device, the server can obtain accurate text data created based on the likelihood of the character recognition candidate and the degree at which the character recognition candidate calculated by the degree calculation unit is uttered.

Further, in the handwriting data processing system according to the present invention, in addition to the effect of the handwriting data processing system according to the sixth aspect, the voice / text conversion means of the degree calculation means may include a voice of a predetermined range. The data is converted into text data, and the frequency calculating means calculates the frequency at which the character recognition candidate is included in the text data converted by the voice / text converting means, and the frequency calculated by the frequency calculating means is the degree. Can be. Therefore, in the terminal device,
The server calculates the likelihood of the character recognition candidate and the frequency at which the character recognition candidate is included in the text data converted by the voice / text conversion means, and creates the frequency calculated by the frequency calculation means as the degree. Accurate text data can be obtained.

Further, in the handwriting data processing system according to the present invention, in addition to the effect of the handwriting data processing system according to the sixth aspect, in the degree calculating means, the voice pattern synthesizing means may use the voice pattern synthesizing means as a recognition candidate voice pattern. And the spotting means obtains the likelihood that the audio pattern synthesized by the audio pattern synthesizing means is included in the audio data in the predetermined range, and the likelihood obtained by the spotting means can be the degree. . Therefore, in the terminal device, the server can obtain accurate text data created using the likelihood of the character recognition candidate and the likelihood obtained by the spotting means as the degree.

According to a ninth aspect of the present invention, in addition to the effect of the handwriting data processing system according to any one of the sixth to eighth aspects, the display of the first clear-text data creating means is provided. The means displays character recognition candidates, and the first reproducing means can reproduce a predetermined range of voice data corresponding to the character recognition candidates specified by the character recognition candidate designating means. Therefore, the operator can correct the character recognition candidate by confirming the sound reproduced by the first reproducing means by the correcting means.

According to a tenth aspect of the present invention, in addition to the effect of the handwritten data processing system according to any one of the fifth to ninth aspects, the handwritten data display means of the text data creating means may include: The writing data stored in the writing data storage means can be displayed. The writing data indicating means indicates writing data displayed by the writing data displaying means, and the second reproducing means can reproduce a predetermined range of voice data corresponding to the writing data specified by the writing data indicating means. Accordingly, the text data can be created by the second copy-written text data creating unit while the operator checks the handwritten data displayed by the display unit.

In the handwriting data processing server according to the eleventh aspect, the server-side receiving means receives the handwriting data, the handwriting data storage means stores the handwriting data received by the server-side receiving means, and the display means Displaying the handwritten data stored in the handwritten data storage means, the operator refers to the handwritten data displayed on the display means, inputs text data from the text data input means, and the input text data is text data. The data can be transmitted to the terminal device by the transmission means. Therefore, accurate text data can be returned to the terminal device.

[0157] In the handwriting data processing server according to the twelfth aspect, the server-side receiving means receives the handwriting data sent from the terminal device, and the handwriting data storage means receives the handwriting data received by the server-side receiving means. The text data creating means character-recognizes the writing data stored in the writing data storage means to create text data, and the text correcting means corrects the text data created by the text data creating means, The data transmitting means can transmit the text data corrected by the text correcting means to the terminal device. Therefore, it is possible to transmit, to the terminal device, accurate text data whose characters are recognized by the server and corrected by the text correction unit.

[0158] In the handwriting data processing server according to the thirteenth aspect, in addition to the effect of the twelfth aspect, the text data creating means recognizes the handwriting data received by the server-side receiving means by character recognition. Thus, a character recognition candidate is created, and the display means can display the handwritten data received by the server-side receiving means and the character recognition candidate created by the text data creating means. Therefore, the server operator can compare the handwritten data displayed on the display means with the character recognition candidates.

In the handwriting data processing server according to the fourteenth aspect, in addition to the effect of the thirteenth aspect, the operator can modify the text by referring to the handwriting data displayed on the display means. By means, the character recognition candidate can be corrected.

Further, in the handwriting data processing server according to the present invention, the server-side receiving means receives the handwriting data and the voice data sent from the terminal device, and the handwriting data storage means uses the server-side receiving means. Storing the handwritten data received, the voice data storage means storing the voice data received by the server-side receiving means, and the text data creating means storing the handwritten data and the voice data stored in the handwritten data storage means. Based on the audio data
Create text data, and the text data transmission means:
The text data created by the text data creating means can be transmitted to the terminal device. Therefore, in the terminal device, it is possible to obtain accurate text data created based on the writing data and the voice data in the server.

In the writing data processing server according to the sixteenth aspect, in addition to the effect of the fifteenth aspect, in the text data creating means, the character recognition candidate generating means is stored in the writing data storage means. The character recognition candidate accuracy calculating means obtains the likelihood of the character recognition candidate generated by the character recognition candidate generation means, and the character / speech data association means generates the character / speech data correspondence means. Character recognition candidates generated by the character recognition candidate generation means corresponding to the data and a predetermined range of voice data corresponding to the writing data targeted when the character recognition candidate generation means generates the character recognition candidates, The degree calculating means outputs the character recognition candidate in voice data in a predetermined range associated with the character recognition candidate by the character / voice data associating means. The first clear copy text data creating means calculates the degree of certainty of the character recognition candidate obtained by the character recognition candidate accuracy calculating means and the degree of utterance of the character recognition candidate calculated by the degree calculating means. Text data can be created based on the text data. Therefore,
Accurate text data can be created based on the likelihood of the character recognition candidate and the degree of utterance of the character recognition candidate calculated by the degree calculation means.

In the handwriting data processing server according to the seventeenth aspect, in addition to the effect of the sixteenth aspect, the voice / text conversion means of the degree calculation means may include:
The voice data in the predetermined range is converted into text data, and the frequency calculating means calculates the frequency at which the character recognition candidate is included in the text data converted by the voice / text converting means, and calculates the frequency calculated by the frequency calculating means. Degree. Therefore, the certainty of the character recognition candidate and the text data converted by the voice / text conversion means include:
It is possible to calculate the frequency at which the character recognition candidate is included, and to create accurate text data created using the frequency calculated by the frequency calculation means as the degree.

Further, in the handwriting data processing server according to the eighteenth aspect, in addition to the effect of the sixteenth aspect, in the degree calculation means, the voice pattern synthesis means synthesizes the voice pattern of the recognition candidate with the character. The likelihood of the voice pattern synthesized by the voice pattern synthesizing means being included in the audio data in the predetermined range by the spotting means can be obtained, and the certainty obtained by the spotting means can be determined as the degree. Therefore, it is possible to create accurate text data using the certainty of the character recognition candidate and the certainty obtained by the spotting means as the above-mentioned degree.

In the writing data processing server according to the nineteenth aspect, in addition to the effect according to any one of the sixteenth to eighteenth aspects, the display means of the first clear-text data creating means may include character recognition. The candidate is displayed, the first reproducing means reproduces a predetermined range of the voice data corresponding to the character recognition candidate designated by the character recognition candidate designating means, and the operator reproduces the predetermined range by the correcting means. The voice recognition can be performed to correct the character recognition candidate. Therefore, accurate text data can be created.

According to a twentieth aspect of the handwriting data processing server of the present invention, in addition to the effects described in any one of the fifteenth to nineteenth aspects, the handwriting data display means of the text data creating means is provided in the handwriting data storage means. The stored writing data is displayed, the writing data indicating means indicates writing data to be displayed by the writing data displaying means, and the second reproducing means outputs voice data corresponding to the writing data specified by the writing data indicating means. The predetermined range is reproduced, and the text data can be created by the operator while checking the handwritten data displayed by the display means by the second copy-written text data creating means. Therefore, accurate text data can be created.

In the writing data processing apparatus according to the present invention, the writing data is input from the writing data input means, and the voice data at the time of inputting the writing data is input from the voice data input means. The data associating means associates the writing data input from the writing data inputting means with the audio data input from the audio data inputting means, and the text data creating means associates the writing data with the writing / audio data associating means. Text data is created based on the handwritten data and the voice data. Therefore, it is possible to create accurate text data based on the writing data and the voice data with one device.

Further, in the handwriting data processing device according to the present invention, the handwriting data is inputted from the handwriting data input means. In addition to the effect of the present invention, the text data creating means has character recognition. The candidate generation means,
Character recognition candidates are generated by performing character recognition on the handwritten data input from the handwriting data input means, the character recognition candidate accuracy calculation means obtains the likelihood of the character recognition candidates generated by the character recognition candidate generation means, and the degree calculation means The degree to which the character recognition candidate is uttered in the voice data in a predetermined range with respect to the time at which the writing data corresponding to the character recognition candidate is input is calculated, and the probability of the character recognition candidate being determined by the fine-text data creation means. The text data can be created based on and the degree calculated by the degree calculating means. Therefore, one device can create accurate text data based on the probability of the character recognition candidate and the degree calculated by the degree calculating means.

In the handwriting data processing apparatus according to the twenty-third aspect, in addition to the effects of the twenty-second aspect,
In the degree calculating means, the voice recognizing means converts a predetermined range of voice data into text data, and the frequency calculating means calculates a frequency at which a character recognition candidate is included in the text data converted by the voice recognizing means. The frequency calculated by the calculation means can be set as the degree. Therefore, one device can create accurate text data based on the probability of the character recognition candidate and the frequency calculated by the frequency calculation means.

In the handwriting data processing device according to the twenty-fourth aspect, in addition to the effect of the twenty-second aspect,
In the degree calculating means, the voice pattern synthesizing means synthesizes a voice pattern of a character recognition candidate, and the spotting means obtains a certainty that the synthesized voice pattern is included in the voice data in the predetermined range, The degree of certainty determined by the above can be set as the above-mentioned degree. Therefore, with one device, it is possible to create accurate text data based on the certainty of the character recognition candidate and the certainty obtained by the spotting means.

In the writing data processing apparatus according to the twenty-fifth aspect, in addition to the effects of the twenty-first to twenty-fourth aspects, in the text data creating means, the character / speech data association means may include The character recognition candidate data generated by the generation unit and the voice data are associated via writing data, the display unit displays the character recognition candidate, and the instruction unit instructs the character recognition candidate displayed by the display unit, The first reproducing means reproduces a predetermined range of the voice data corresponding to the character recognition candidate specified by the specifying means, and the operator confirms the voice reproduced by the first reproducing means by the correcting means to check the character recognition candidate. Can be modified. Therefore, a predetermined range of the voice data corresponding to the character recognition candidate specified by the instructing means is reproduced, and the operator confirms the voice reproduced by the first reproducing means by the correcting means to create accurate text data. be able to.

In the writing data processing apparatus according to the twenty-sixth aspect, in addition to the effects of the twenty-first to twenty-fourth aspects, in the text data creating means, the writing data display means is input by the writing data input means. When the handwritten data is displayed and the handwritten data indicating means indicates the handwritten data to be displayed by the handwritten data indicating means, the second reproducing means sets a predetermined range of the voice data corresponding to the handwritten data indicated by the handwritten data indicating means. Reproduced, accurate text data can be created by the second copy-written text data creating means while the operator checks the handwritten data displayed by the display means.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram of a writing data processing system 1 according to a first embodiment of the present invention.

FIG. 2 is a block diagram illustrating a device configuration of a writing data processing server 2;

FIG. 3 is a block diagram showing a device configuration of a client terminal device 5;

FIG. 4 is a flowchart of a process executed by a handwriting data processing server 2.

FIG. 5 is a diagram showing a state in which handwritten characters are written on the handwritten character input tablet 6;

FIG. 6 is a diagram illustrating a CRT 30 of the writing data processing server 2;
It is a figure which shows the display screen of.

FIG. 7 shows an LCD 60 of the client terminal device 5.
It is a figure which shows the display screen of.

FIG. 8 is a flowchart of a process performed by a handwriting data processing server 2 of the handwriting data processing system 1 according to the second embodiment.

FIG. 9 is a flowchart of a process performed by a handwriting data processing server 2 of a handwriting data processing system 1 according to the third embodiment.

FIG. 10 is a CRT of the writing data processing server 2;
30 is a diagram showing a display screen 30. FIG.

FIG. 11 is a flowchart of a process performed by a handwriting data processing server 2 of a handwriting data processing system 1 according to the fourth embodiment.

FIG. 12 is a flowchart of a process performed by a handwriting data processing server 2 of a handwriting data processing system 1 according to the fifth embodiment.

FIG. 13 is a flowchart of a process performed by a handwriting data processing server 2 of a handwriting data processing system 1 according to a sixth embodiment.

FIG. 14 is a block diagram illustrating a device configuration of a writing data processing server 200 according to a modification.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Writing data processing system 2 Writing data processing server 3 Internet 5 Client terminal device 6 Handwriting character input tablet 6a Pressure sensitive pad 8 Microphone 20 CPU 21 RAM 22 ROM 23 I / O bus 24 Video controller 25 Key controller 26 CD-ROM drive 27 CD-ROM 28 Hard Disk Device 29 LAN Card 30 CRT 32 Router 33 Sound Card 34 Speaker 35 Mouse Controller and Roller 36 Mouse 200 Writing Data Processing Device

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 7/173 610 H04N 7/173 630 630 G10L 3/00 571Q F-term (Reference) 5B064 AA07 AB04 BA05 EA10 FA01 FA16 5C064 BA07 BB10 BC16 BC18 BC20 BC23 BD02 BD09 BD13 5D015 LL07

Claims (26)

[Claims] 1. A handwritten data processing system for connecting a terminal device and a server via a network, converting handwritten data transmitted from the terminal device into text data at the server, and returning the text data to the terminal device. A terminal device, a handwritten data input unit for inputting handwritten data, a terminal side transmitting unit for transmitting the handwritten data input to the handwritten data input unit to the server, and a terminal side receiving unit for receiving text data from the server The server comprises: a server-side receiving unit that receives the writing data sent from the terminal device; a writing data storage unit that stores the writing data received by the server-side receiving unit; and the writing data storage. Display means for displaying the writing data stored in the means; and operating with reference to the writing data displayed on the display means. A writing data processing system, comprising: text data input means for inputting text data by an author; and text data transmission means for transmitting text data input from the text data input means to the terminal device. 2. A handwritten data processing system for connecting a terminal device and a server via a network, converting handwritten data transmitted from the terminal device into text data by the server, and returning the text data to the terminal device. A terminal device, a handwritten data input unit for inputting handwritten data, a terminal side transmitting unit for transmitting the handwritten data input to the handwritten data input unit to the server, and a terminal side receiving unit for receiving text data from the server The server comprises: a server-side receiving unit that receives writing data sent from the terminal device; a writing data storage unit that stores the writing data received by the server-side receiving unit; and the writing data storage. Text data creating means for creating text data by character recognition of the handwritten data stored in the means; A handwritten data process comprising: text correction means for correcting text data created by text data creation means; and text data transmission means for transmitting the text data corrected by the text correction means to the terminal device. system. 3. The server further comprises a display unit, wherein the text data creating unit performs character recognition on the handwritten data received by the server-side receiving unit to create a character recognition candidate. 3. The handwritten data processing system according to claim 2, wherein the handwritten data received by the server-side receiving means and the character recognition candidates created by the text data creating means are displayed. 4. The handwriting according to claim 3, wherein the operator can correct the character recognition candidate by the text correction means with reference to the writing data displayed on the display means. Data processing system. 5. A handwriting data processing system for connecting a terminal device and a server via a network, converting handwriting data transmitted from the terminal device into text data at the server, and returning the text data to the terminal device, The terminal device includes a handwriting data input unit for inputting handwriting data, a voice data input unit for inputting a voice at the time of the handwriting data input, and a handwriting data input from the handwriting data input unit and an input from the voice data input unit. Writing / speech data associating means for associating with the recorded speech data, terminal-side transmitting means for transmitting the writing data and speech data associated with the writing / speech data associating means to the server, and text from the server. Terminal receiving means for receiving data, wherein the server receives the data transmitted from the terminal device. Server-side receiving means for receiving handwritten data and voice data; handwritten data storage means for storing handwritten data received by the server-side receiving means; and voice data storage for storing voice data received by the server-side receiving means. Means, text data creating means for creating text data based on the handwritten data stored in the handwritten data storage means and the voice data stored in the voice data storage means, A writing data processing system comprising: text data transmitting means for transmitting text data to the terminal device. 6. The text data generating means, wherein the character recognition candidate generating means generates a character recognition candidate by performing character recognition on the writing data stored in the writing data storage means, and the character recognition candidate generating means generates the character recognition candidate. Character recognition candidate accuracy calculation means for determining the likelihood of a character recognition candidate; character recognition candidates generated by the character recognition candidate generation means corresponding to the writing data; and the character recognition candidate generation means generating the character recognition candidates And the audio data in a predetermined range corresponding to the writing data targeted at the time of
Character / speech data associating means for associating via the handwritten data; and wherein the character recognition candidate is vocalized in the predetermined range of speech data associated with the character recognition candidate by the character / speech data associating means. Degree calculating means for calculating the degree of being performed, the likelihood of the character recognition candidate obtained by the character recognition candidate accuracy calculating means, and the degree to which the character recognition candidate calculated by the degree calculating means is uttered. 6. The handwritten data processing system according to claim 5, further comprising: first copy-written text data creating means for creating text data based on the text data. 7. The degree calculating means includes: a voice / text converting means for converting the voice data in the predetermined range into text data; and the text data converted by the voice / text converting means includes the character recognition candidate. The writing data processing system according to claim 6, further comprising: a frequency calculating unit configured to calculate a frequency of the writing data, wherein the frequency calculated by the frequency calculating unit is the degree. 8. The voice data in the predetermined range, wherein the degree calculation means includes a voice pattern synthesis means for synthesizing a voice pattern of the character recognition candidate, and a voice pattern synthesized by the voice pattern synthesis means. 7. The writing data processing system according to claim 6, further comprising spotting means for obtaining certainty, wherein the certainty obtained by said spotting means is used as said degree. 9. The first copy-written text data creating means, a display means for displaying the character recognition candidate, a character recognition candidate designating means for designating a character recognition candidate to be displayed by the display means, and the character recognition candidate. First reproducing means for reproducing a predetermined range of voice data corresponding to the character recognition candidate indicated by the instruction means, and an operator correcting the character recognition candidate by confirming the sound reproduced by the first reproducing means. 9. The writing data processing system according to claim 6, further comprising a correction unit that performs the correction. 10. The handwriting data display means for displaying handwriting data stored in the handwriting data storage means, the handwriting data instruction means for indicating handwriting data to be displayed by the handwriting data display means, A second reproducing unit that reproduces a predetermined range of audio data corresponding to the handwritten data instructed by the handwritten data instructing unit; and text data while an operator confirms the handwritten data displayed by the display unit. 10. A method according to claim 5, further comprising a second means for creating a copy-written text data.
The written data processing system according to any one of the above. 11. A writing data processing server which connects a terminal device via a network, converts writing data transmitted from the terminal device into text data, and returns the text data to the terminal device. Server-side receiving means for receiving the coming writing data, writing data storage means for storing the writing data received by the server-side receiving means, display means for displaying the writing data stored in the writing data storage means, Text data input means for inputting text data by an operator with reference to the handwritten data displayed on the display means, and text data transmission means for transmitting the text data input from the text data input means to the terminal device And a writing data processing server. 12. A writing data processing server which connects a terminal device via a network, converts writing data transmitted from the terminal device into text data, and returns the text data to the terminal device. Server-side receiving means for receiving incoming writing data; writing data storage means for storing the writing data received by the server-side receiving means; and text data obtained by character recognition of the writing data stored in the writing data storage means. Text data creating means for creating text data, text correcting means for correcting the text data created by the text data creating means, and a clear text data transmitting means for transmitting the text data corrected by the text correcting means to the terminal device. A writing data processing server, comprising: 13. The apparatus further comprising a display unit, wherein the text data creating unit recognizes characters of the handwritten data received by the server-side receiving unit, and creates a character recognition candidate. 13. The writing data processing server according to claim 12, wherein the writing data received by the receiving unit and the character recognition candidates created by the text data creating unit are displayed. 14. An operator refers to the handwritten data displayed on the display means, and
14. The writing data processing server according to claim 13, wherein the character recognition candidate can be modified. 15. A writing data processing server which connects a terminal device via a network, converts writing data transmitted from the terminal device to text data, and returns the text data to the terminal device. Server-side receiving means for receiving incoming handwritten data and voice data associated with the handwritten data; handwritten data storage means for storing handwritten data received by the server-side receiving means; and Voice data storage means for storing received voice data; text data generating means for generating text data based on the handwritten data stored in the handwritten data storage means and the voice data stored in the voice data storage means And a text for transmitting the text data created by the text data creation means to the terminal device. Handwritten data processing server, characterized in that a chromatography data transmission means. 16. The text data creation means, wherein the character recognition candidate creation means for creating a character recognition candidate by performing character recognition on the handwritten data received by the server-side reception means, and the character created by the character recognition candidate creation means Character recognition candidate accuracy calculation means for determining the likelihood of a recognition candidate; character recognition candidates generated by the character recognition candidate generation means corresponding to the handwritten data; and the character recognition candidate generation means generating the character recognition candidates. The audio data in a predetermined range corresponding to the writing data targeted at the time,
The character / speech data associating means for associating via the writing data, and the character recognition candidate is uttered in the voice data of the predetermined range associated with the character recognition candidate by the character / speech data associating means. A degree calculating means for calculating the degree of character recognition, a probability of the character recognition candidate obtained by the character recognition candidate accuracy calculating means, and a degree at which the character recognition candidate calculated by the degree calculating means is uttered. 16. The handwriting data processing server according to claim 15, further comprising: a first copy-written text data creating unit that creates text data by using the method. 17. A speech / text conversion unit for converting the voice data in the predetermined range into text data, wherein the text recognition result includes the character recognition candidate. 17. The writing data processing server according to claim 16, further comprising: a frequency calculating unit configured to calculate a frequency of the writing data, wherein the frequency calculated by the frequency calculating unit is the degree. 18. The voice data in the predetermined range, wherein the degree calculating means includes a voice pattern synthesizing means for synthesizing a voice pattern of the character recognition candidate, and a voice pattern synthesized by the voice pattern synthesizing means. 17. The writing data processing server according to claim 16, wherein the spotting means for obtaining the certainty, and the certainty obtained by the spotting means is the degree. 19. The first text-written text data creating means, a display means for displaying the character recognition candidates, a character recognition candidate designating means for designating the character recognition candidates displayed by the display means, and the character recognition candidate. First reproducing means for reproducing a predetermined range of voice data corresponding to the character recognition candidate indicated by the instruction means, and an operator correcting the character recognition candidate by confirming the sound reproduced by the first reproducing means. 19. The writing data processing server according to claim 16, further comprising a correction unit that performs the correction. 20. A handwritten data display means for displaying the handwritten data stored in the handwritten data storage means, and a handwritten data instructing means for indicating handwritten data to be displayed by the handwritten data display means. Second playback means for playing back a predetermined range of voice data corresponding to the handwritten data instructed by the handwritten data instructing means; and text data created by an operator while checking the handwritten data displayed by the display means. 20. The handwriting data processing server according to claim 15, further comprising: a second clear-text data generator. 21. Writing data input means for inputting writing data, voice data input means for inputting voice data at the time of inputting the writing data, writing data input from the writing data input means, and inputting the voice data. Writing / speech data associating means for associating audio data input from the means, and text data creating means for creating text data based on the writing data and the audio data associated by the writing / speech data associating means A writing data processing device comprising: 22. A character recognition candidate generating unit configured to generate a character recognition candidate by character-recognizing writing data input from the writing data input unit, wherein the character data generating unit generates the character recognition candidate. A character recognition candidate accuracy calculating means for determining the likelihood of the character recognition candidate, and wherein the character recognition candidate is uttered during a predetermined range of voice data for a time at which writing data corresponding to the character recognition candidate is input. A degree calculation means for calculating the degree of presence of the character, and a fair copy text data creation means for creating text data based on the likelihood of the character recognition candidate and the degree calculated by the degree calculation means. The writing data processing device according to claim 21. 23. The degree calculating means, wherein: a voice recognizing means for converting the voice data in the predetermined range into text data; and a frequency of including the character recognition candidate in the text data converted by the voice recognizing means. 23. The writing data processing apparatus according to claim 22, further comprising: a frequency calculating unit that calculates the frequency, wherein the frequency calculated by the frequency calculating unit is the degree. 24. The degree calculating means, comprising: voice pattern synthesizing means for synthesizing the voice pattern of the character recognition candidate; and spotting for determining the likelihood that the synthesized voice pattern is included in the voice data in the predetermined range. 23. The writing data processing apparatus according to claim 22, further comprising: means for determining the likelihood obtained by the spotting means as the degree. 25. A text / speech data associating means for associating the character recognition candidate data generated by the character recognition candidate generating means with the voice data via handwritten data, Display means for displaying recognition candidates; instruction means for instructing character recognition candidates to be displayed by the display means; first reproduction for reproducing a predetermined range of voice data corresponding to the character recognition candidates instructed by the instruction means 25. The apparatus according to claim 21, further comprising: a correction unit configured to correct the character recognition candidate by confirming a voice reproduced by the first reproduction unit by an operator. Writing data processing device. 26. A handwritten data display means for displaying the handwritten data input by the handwritten data input means, and a handwritten data instructing means for indicating handwritten data displayed by the handwritten data display means. Second playback means for playing back a predetermined range of audio data corresponding to the handwritten data instructed by the handwritten data instructing means; and text data while an operator confirms the handwritten data displayed by the display means. 25. The handwriting data processing device according to claim 21, further comprising a second copy-formed text data creating means for creating.