JP3319203B2 - Document filing method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document filing method and apparatus, and more particularly, to a document filing system for registering and retrieving document image information, which can be applied to the case of inputting and registering document image information.

[0002]

2. Description of the Related Art Conventionally, in a document filing system for registering and retrieving document image information, various methods have been proposed for easily inputting registration items to a database used for retrieving document image information. As a document filing system capable of inputting a registration item to a database used for searching for such document image information, for example, there is one disclosed in Japanese Patent Application Laid-Open No. 4-77885. This is a method of recognizing characters by image recognition data such as a business card read by a scanner or the like, and recognizing an item by using a character recognizing means. It is.

FIG. 8 is a display showing a correction operation in the document filing system. In the figure, C1 is a cursor of the character data currently designated, C2 is a selection display of a recognition candidate list for the character data indicated by the cursor C1, and C3 is a selection display of an item name list indicated by the cursor C1. Show. The cursor C1 is moved to the position of the character data to be corrected, a correct character is selected by C2, and a correct item name is selected and corrected by C3. After sequentially performing such post-correction, the read image data and the recognized character data are registered in the database.

[0004]

However, in the display and correction operation in the above-described document filing system, a list of item names of one document is displayed, and after correcting them, a list of items of another document is displayed. Therefore, in order to correct all the documents, a complicated operation of updating and correcting the display of the screen each time is necessary, and there has been a problem that the efficiency of the correction work is low. In the document filing system described above,
Only documents that have a limited number of items in a limited format such as business cards and slips are handled, and there is no convenience such as inputting and correcting multiple types of documents at once, The usability of users was still insufficient.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and it is possible to collectively input a plurality of types of documents with a simple configuration and operation, and to impose a burden on a user in correction work. It is an object of the present invention to obtain a document filing method and apparatus capable of reducing the number of times and improving the usability.

[0006]

According to a document filing method according to the present invention, a series of image data obtained by collectively reading a plurality of types of documents from an image reading apparatus is converted into a document corresponding to each of the read plurality of documents. A document image data input process to be input as image data, a pattern recognition process for pattern-recognizing and coding the document image data input in the document image data input process, and generating a coded data, each corresponding to a plurality of documents. A first display step of collectively displaying document image data and coded data to be displayed on the display unit, and determining whether post-correction of the coded data is necessary based on a display result of the first display step. A correction process for post-correction of the coded data based on the determination result, and a second display process for displaying the coded data corrected in the correction process on a display unit , And a database registration processing step of registering in a database in bulk document image data and the coded data displayed on the display unit, the document image data
The input process is based on the document image data and the data corresponding to the document type.
Extract the image data of the database registration item area and
The recognition process consists of the data extracted in the document image data input process.
Code the image data in the database registration item area
And the first and second display steps are:
Image data in the database registration item area and this image data
The database entry values that encode the data
And database registration items arranged in a table in a batch
To display .

Further, in the document filing method according to the next invention, the document image data input step includes a step of comparing a series of image data with predetermined standard document information and dividing the read image into a plurality of read documents. And a document discriminating step of comparing the plurality of documents divided in the document dividing step with the standard document information to discriminate the type of the document and transmitting the same as document image data.

In the document filing method according to the next invention, the document image data input step further includes, for each document type, a description of a database registration item value when the document is registered in the database in addition to the fixed document information. A document item definition process in which a database registration item region indicating a position is defined in advance, and a database registration item extraction process of extracting image data of a database registration item region corresponding to a document type from document image data corresponding to a document. The image data of the database registration item area extracted in the database registration item extraction process is coded in the pattern recognition process, the coded database registration item value is assigned, and the document image data is transferred to the database in the database registration process. It is to register.

In the document filing method according to the next invention, the first and second display steps include the steps of storing document image data and coded database registration item values of a plurality of types of documents, and document and database registration items. The data is displayed collectively in an arrayed table format, and the correction process is to correct the database registration item value coded corresponding to the document in the document and database registration item table.

[0010] In the document filing method according to the next invention, the first and second display steps switch between display and operation according to the type of document.

A document filing method according to the next invention is characterized in that a first display mode in which the same type of document is displayed collectively for a plurality of different types of documents, and that different types of documents are mixed and displayed simultaneously. The second display mode is switched in response to a request from the user.

In the document filing method according to the next invention, the second display mode includes a fixed display mode in which display positions of database registration item values different for each type of document are fixed and displayed, and a second display mode for each type of document. A left-justified display mode in which display positions of different database registration item values are displayed left-justified is switched according to a user's request.

[0013] In the document filing method according to the next invention, the correction step includes an image data display area for always enlarging and displaying document image data corresponding to a database registration item value in a document to be corrected, and a coded image data display area. A pattern data display area for constantly displaying the pattern data of the database registration item value in an enlarged manner.

[0014] In the document filing method according to the next invention, the correction process includes the steps of: document image data of the entire page including the database registration item value in the document to be corrected; and coded pattern data of the database registration item value. Are always displayed at a specified display magnification.

In the document filing method according to the next invention, the display of the document image data of the entire page in the fixed display area is performed by changing the area on the document image data corresponding to the database entry to be corrected. An area or color is displayed by a frame, and an area on the document image data corresponding to the database registration item value to be corrected is moved to a minimum necessary to fit in the display area.

In the document filing method according to the next invention, when the user corrects the correction process to be a document type different from the document type determined by the document determination process, the document type is corrected. Only the document is reprocessed according to the processing procedure corresponding to the document type defined by the document item definition process.

Further, in the document filing method according to the next invention, the database registration processing step includes a keyword extraction step of cutting out words from coded data obtained by coding the entire document by the character recognition step and extracting the extracted words as keywords. When a keyword is specified as a registration item, the keyword extracted in the keyword extraction process is added and registered in the database.

In the document filing method according to the next invention, an area for extracting a keyword from a database entry item value is limited, and a keyword is extracted from coded data obtained by coding only a specified area by a character recognition process. Then, the extracted keyword is added and registered in the database.

The document filing method according to the next invention further comprises a keyword extraction range designation step for designating a keyword extraction range in accordance with the contents of the document. A keyword is extracted from a specified extraction range, and the extracted keyword is added to the extracted keyword and registered in a database.

A document filing apparatus according to another aspect of the present invention inputs a series of image data obtained by reading a plurality of types of documents collectively from an image reading apparatus as document image data corresponding to the read plurality of documents. Document image data inputting means, pattern recognition means for pattern-recognizing and coding the document image data input by the document image data inputting means, and coded data, and document image data respectively corresponding to a plurality of documents. Display means for collectively displaying coded data on the display unit, and determining whether post-correction of the coded data is necessary based on the display result of the display unit, and post-correcting the coded data based on the determination result. Correction means for supplying to the display means, and the document image data and coded data displayed on the display unit are collectively registered in the database. And a database registration processing means,
The document image data input means converts the document image data into a document.
Image data in the database registration item area according to the type
Extracting and pattern recognizing means, document image data input means
Copy the image data of the database registration item area extracted in
Into a database entry item value.
Image data in the database registration item area and this image data
The database entry values that encode the data
And database registration items arranged in a table in a batch
To display .

Further, in the document filing apparatus according to the next invention, the document image data input means compares the series of image data with predefined standard document information and divides the image data into a plurality of read documents. And a document discriminating means for comparing the plurality of documents classified by the document dividing means with the standard document information to determine the type of the document and transmitting the same as document image data.

In the document filing apparatus according to the next invention, the document image data input means further includes, for each type of document, a database registration item value when registering the document in a database, in addition to the standard document information. Document item defining means for defining a database registration item area indicating a position in advance, and database registration item extracting means for extracting image data of a database registration item area corresponding to a document type from document image data corresponding to a document. The image data of the database registration item area extracted by the database registration item extraction means is coded by the pattern recognition means, the coded database registration item value is added, and the document image data is transferred to the database by the database registration processing means. It is to register.

In the document filing apparatus according to the next invention, the display means collects the document image data and the coded database registration item values of a plurality of types of documents in a tabular form in which the documents and database registration items are arranged. The correction means corrects the coded database registration item value corresponding to the document in the document and database registration item table.

[0024]

A series of image data obtained by reading a plurality of types of documents collectively from an image reading apparatus is input as document image data corresponding to each of the read documents.
Generates coded data by pattern recognition of the document image data, collectively displays the document image data and coded data corresponding to a plurality of documents on the display unit, and post-corrects the coded data as necessary Then, the document image data and the coded data displayed on the display unit are collectively registered in the database. Thus, a plurality of types of documents can be input collectively with a simple configuration and operation, and the user's work load in correction work can be reduced and usability can be improved.

Further, a series of image data is compared with predetermined standard document information, divided into a plurality of read documents, and the divided plural documents are compared with the standard document information to determine the type of the document. Determine. Thus, a plurality of documents can be reliably input to a plurality of types collectively with a simple configuration and operation.

Further, from the document image data corresponding to the document, the image data of the database registration item area corresponding to the type of the document is extracted, and the image data of the database registration item area is coded in a pattern recognition process, and the coding is performed. The document image data is registered in the database by giving the database entry item value. Thereby, a plurality of types of documents can be reliably input collectively with a simple configuration and operation.

Further, the document image data and the coded database registration item values of a plurality of types of a plurality of documents are collectively displayed in a table format in which the documents and the database registration items are arranged. Modify the database entry value coded for the document. Thereby, the user's work load in the correction work can be reduced and the usability can be improved.

The display and operation of the first and second display steps are switched according to the type of the document. As a result, the user can easily recognize the type of the document, and the user's work load in the correction work can be reduced and the usability can be improved.

Further, for a plurality of different types of documents,
A first display mode for displaying the same type of documents collectively, and a second display mode for simultaneously displaying different types of documents mixedly.
Is switched in response to a user request.
Thus, if the user selects the first display mode,
You can compare the contents of multiple documents of the same type, and if you select the second display mode, you can compare the contents of different documents,
Since these can be selected as necessary, the work load on the user in the correction work can be reduced and the usability can be improved.

The second display mode includes a fixed display mode in which the display position of a database registration item value that differs for each type of document is fixed and displayed, and a display position of a database registration item value that differs for each type of document. The display mode is switched between left-justified display mode and left-justified display mode according to the user's request.
Thus, if the user selects the fixed display mode, it is possible to easily recognize the presence or absence of database registration items for each type of document, and if the user selects the left-justified display mode, more database registration items in the same document Can be displayed and recognized, and these can be selected as necessary, so that the work load on the user in the correction work can be reduced and the usability can be improved.

In the correction step, the document image data corresponding to the database registration item value in the document to be corrected is always enlarged and displayed in the image data display area, and the coded database registration item value is displayed in the pattern data display area. Is always enlarged. This allows
The enlarged and displayed pattern data to be corrected and its document image data can be compared and recognized, and the user's work load in the correction work can be reduced and the usability can be improved.

In the correction process, the document image data of the entire page including the database registration item value in the document to be corrected and the coded database registration item value pattern data are stored in the fixed display area. , Always display at the specified display magnification. As a result, the pattern data to be corrected displayed at the specified display magnification and the document image data can be compared and recognized, and the user's work load in the correction work can be reduced and the usability can be improved.

Further, the display of the document image data of the entire page in the fixed display area is performed by surrounding the area on the document image data corresponding to the database entry item value to be corrected with a frame or displaying a color. The area on the document image data corresponding to the database entry item value to be corrected is moved to the minimum necessary to fit in the display area. As a result, the document image data corresponding to the database registration item value to be corrected can be reliably recognized, and the user's work load in the correction work can be reduced and the usability can be improved.

When the user corrects the correction process so that the document type is different from the document type determined by the document determination process, only the document whose document type has been corrected is defined by the document item definition process. The document is re-processed according to the processing procedure corresponding to the type of the received document. Thus, even if a document is erroneously determined, it can be easily corrected, and a plurality of types of documents can be reliably input collectively.

In the keyword extraction process of the database registration process, a word is cut out from coded data obtained by coding the entire document by the character recognition process and extracted as a keyword. When the keyword is specified as a database registration item, The extracted keyword is assigned and registered in the database. As a result, keywords can be easily extracted and registered in the database, and the usability of the user can be improved.

Further, the area for extracting the keyword from the database entry item value is limited, the keyword is extracted from the coded data obtained by coding only the designated area by the character recognition process, and the extracted keyword is assigned. To register in the database. As a result, the database registration item value can be easily extracted as a keyword and registered in the database, and the usability of the user can be improved.

In the keyword extraction range designating step, a keyword extraction range is designated in accordance with the contents of the document, and keywords are extracted from the designated extraction range in the coded data obtained by encoding the entire document by the character recognition step. The extracted keyword is added to the extracted keyword and registered in the database. As a result, valid keywords can be extracted and registered in the database, and the usability of the user can be improved.

Further, a series of image data read from a plurality of types of documents collectively by an image reading device is input as corresponding document image data by a document image data input means, and pattern recognition is performed by a pattern recognition means. Coded data is generated, and the display means collectively displays the document image data and the coded data corresponding to the plurality of documents on the display unit, and the correction means as necessary based on the display result of the display unit. The document image data and the coded data displayed on the display unit are collectively registered in the database by the database registration unit. Thus, a plurality of types of documents can be input collectively with a simple configuration and operation, and the user's work load in correction work can be reduced and usability can be improved.

Further, in the document image data input means, a series of image data is compared with predefined standard document information by the document classification means, and divided into a plurality of read documents.
The plurality of documents classified by the document discriminating means are compared with standard document information to determine the type of the document, and transmitted as document image data. Thus, with a simple configuration and operation, a plurality of documents in a plurality of types can be reliably input collectively.

Further, in the document image data input means, a database entry item area indicating the entry position of a database entry item value for each document type is defined in advance by the document item definition means in addition to the standard document information. The item extracting means extracts the image data of the database registration item area from the document image data corresponding to the document, adds the coded database registration item value by the pattern recognition means, and registers the document image data in the database. Thereby, a plurality of types of documents can be reliably input collectively with a simple configuration and operation.

The display means collectively displays the document image data and the coded database registration item values of a plurality of types of documents in a table format in which the documents and the database registration items are arranged. Correct the database entry value coded corresponding to the document in the table of document and database entry. Thereby, the user's work load in the correction work can be reduced and the usability can be improved.

[0042]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

Embodiment 1 FIG. 1 shows the configuration of a document filing system according to the present invention incorporated in a personal computer or a workstation as a whole.
Is a control unit which comprises a CPU (Central Processing Unit) and controls the operation of the entire system. The control unit 1 includes a document reading unit 2, a document classifying unit 3, a document discriminating unit 4, a database registration item extracting unit 5, a character recognizing unit 6, a batch display unit 7, a correcting unit 8, a keyword extracting unit 9, a database registering unit. The control unit 10 instructs control and transmits and receives data to and from the storage unit 11.

The document reading means 2 is a software driver which operates corresponding to the document reading section 12 comprising an optical image reading device having a sheet feeder. As a result, a plurality of types of documents are placed on the document reading unit 12, and image data obtained by scanning all at once is input to the document filing system. Document classification means 3
Compares a series of image data corresponding to a plurality of types of documents input through the document reading unit 2 with a document image pattern predefined in the document item definition unit 13 using a pattern matching method or the like. , And recognizes the first page of the document and classifies it as document image data corresponding to one document. The document image data obtained as a result is sent to the document discriminating means 4 again by the document item defining means 13.
Is compared with the document image pattern defined in the above, and the type of the document is determined.

In practice, in this embodiment, FIG.
A plurality of documents of a plurality of types are read and input collectively with the top identification page PH shown in FIG. On the top identification page PH, the top page mark HM of the rectangular area painted in advance
And a document type mark SM composed of a plurality of rectangular areas each of which specifies the type of the document by the user by painting. Therefore, in the document item definition means 13, a document image pattern corresponding to the first page mark HM and a document image pattern having a document type mark SM corresponding to the type of document are defined in advance. As a result, the document classifying means 3 and the document discriminating means 4 can classify the input image data for each document and determine the type of the document by an easy pattern matching technique.

When the above-described head identification page PH (FIG. 2) is used, the presence or absence of a filled rectangular area is detected in place of the pattern matching technique to divide the document and determine the type of the document. In this case, the processing of the document classification unit 3 and the document determination unit 4 can be further reduced. Also, without using the head identification page PH, the head document image pattern is defined in advance and stored in the storage unit 11 for a plurality of types of documents to be input, and pattern matching is performed using this document image pattern. The document may be classified and the type of the document may be determined. In this case, the user may place only the document on the document reading device 12 and read the document, thereby improving the usability when inputting the document.

The document item definition means 13 includes, in addition to the above-described plural types of document image patterns used for document classification and document type discrimination, the item names of database registration items corresponding to each document type, and The description position in the document corresponding to the item name is defined as a database registration item area. Database registration item extraction means 6
Extracts the image data of the contents of the database registration item area according to the document type from the document image data determined by the document determination means 4. In the case of this embodiment, in the database registration item area, coordinate information of the area is numerically defined as a description position in the document, and document image data within the range of the coordinate information is cut out as image data of the database registration item area. . This database registration area may also be defined as a document image pattern. In this case, the image data of the database registration item area can be cut out using the pattern matching method as described above.

The character recognizing means 6 recognizes a character pattern of the image data of the database registration item area extracted by the database registration item extracting means 5 by a character recognition technique and converts it into a character code. Is sent to the batch display means 7. The character recognition patterns are characters, alphanumeric characters, symbols, patterns, and the like. The batch display means 7 arranges documents and database registration items vertically and horizontally, and arranges image data of database registration items of a plurality of types of documents and character string data obtained by converting the image data into two lines, respectively. And collectively display them on the display unit 14.
At the same time, the correction means 8 displays the document image data of the corresponding page of the document containing the database registration item to be edited in the image data viewer which is a fixed display area, and displays a frame on the corresponding area to be edited. For example, various display information is generated and sent to the display unit. The display unit 14 presents display information generated by the collective display unit 8 to a user, and includes a CRT (cathode ray tube) display or the like.

The correction means 8 responds to various correction operations instructed by the user through the operation unit 15 for the various information displayed on the display unit 14 and makes desired corrections to the internal data. Is transmitted to the batch display means 7 and is reflected on the display content projected on the display unit 14. In addition to the correction operation of the recognized characters, the user can operate the division of the document and the correction of the error of the type. In this case, depending on the type of the corrected document, A desired database registration item extraction area is obtained from the document item definition means 13, reprocessed through the database registration item extraction means 5 and the character recognition means 6 described above, and added and displayed again by the collective display means 7. .
The operation unit 14 includes a keyboard and a pointing device such as a mouse for inputting a correction desired by the user to the information displayed on the display unit 14 to the correction unit 10.

The keyword extracting means 9 includes the character recognizing means 6
A keyword useful for document retrieval is extracted from the character string data of the database registration item coded by the correction means 8 and corrected by the correction means 8. Further, according to the type of the document, the document image data of the region of the entire document or the region such as the abstract as the designated document category is extracted from the character string data coded by the character recognizing means 6, and all the words are cut out. May be extracted. Database registration means 10
Is added to the database registration item extracted by the database registration item extracting means 5 and coded by the character recognizing means 6 and corrected by the correcting means 8 and the keyword extracted by the keyword extracting means 9 as document information. The document image data determined in step 4 is registered in the database of the storage unit 11. The storage unit 14 stores a document image pattern of the document item definition unit 13, a document registered by the database registration unit 13, and the like, and includes a semiconductor storage device and a disk device.

FIG. 3 shows the flow of document filing processing in the document filing system of this embodiment having such a configuration. A plurality of types of documents 20 are placed on the document reading unit 12 and input collectively through the document reading unit 2, and a series of image data corresponding to the plurality of documents is converted into documents by the document sorting unit 3. Divided,
The type of each document is determined by the document determination unit 4. For example, the definition 22A for the document A obtained from the document item definition means 13 is applied to the document image data 21A determined to be the document A, and the document image data 21A is processed according to the procedure defined therein.

From each document image data 21, the image data 23 of the database registration item area is cut out by the database registration item extraction means 5 according to the definition 22 of the document item definition means 13. As a result, the image data 23 of each database registration item area cut out is converted into a character code by the character recognizing means 6, and character string data 24 composed of database registration item values is created. These are displayed by the batch display / correction interface 25 and post-corrected as necessary, and the document image data and the character string data are registered in the database 26 constructed in the storage unit 11.

Here, the batch display / modification interface 25
Will be described in detail. The information for collectively displaying the database registration items of all the documents is generated by the collective display means 7 and displayed on the display unit 14. The user can confirm the information and operate the operation unit 15 as necessary while confirming the information. Correction can be performed interactively. In this case, only one cursor is displayed on the operation unit 15 at all times, and the database registration item at the position where the cursor is displayed is a correction target at that time.

In the database registration items to be corrected, the image data and the coded character string data are displayed in an enlarged manner, so that the user can correct the coded character string data while checking the expanded image data. it can. The image data of the corresponding page of the document including the database registration item to be corrected is always displayed in the image data viewer, and the database registration item area corresponding to the database registration item is displayed with a frame, or the color is inverted. The user can easily recognize the database registration item to be corrected on the image data.

If the size of the image data of the corresponding page is too large to fit in the image data viewer, scroll in the minimum necessary range to display the area corresponding to the database registration item to be corrected. When correcting the character string data of the coded database registration item, if the selected character has two or more character recognition candidates, a list thereof is displayed, and the character can be replaced with the character selected by the user. However, if there is no correct character in the list of character recognition candidates, the correct character is input from a keyboard or the like using kana-kanji conversion or the like.

FIG. 4 shows the display of the batch display / correction interface 25 for correction displayed on the display unit 14. This is configured based on the collective display data generated by the collective display means 7. The correction instruction input from the operation unit 15 is interpreted by the correction means 8, the correction is performed, and the result is displayed. , And reflected on the display unit 14, and thereafter, by repeating this, the user performs a correction work on each database registration item of each document as necessary.
In the case of this embodiment, in the display frame of the display unit 14, in addition to the main display area for displaying documents and database registration items in a table format, the display areas for the image data viewer 30 and the character string data viewer 31 described above are provided. Are arranged.

In the main display area, the document type label 32 is a label indicating the type of the document, indicates the result determined by the document determination means 4, and the data displayed on the same line is a database registration item of the same document. Indicates that there is.
The database registration item label 33 is a label indicating a database registration item for searching for a document. The label indicates a result extracted by the database registration item extracting means 5, and data indicated by the same digit is the same database of each document. Indicates a registered item. The cells corresponding to each document are an image data cell 34 and a character string data cell 35.
In two stages.

The image data cell 34 indicates the image data of the area of the database image registration item of the document image data of the document. The image data to be displayed at this time is reduced to a necessary minimum so that the display can be accommodated in the image data cell 34. The character string data cell 35 is
The character string data obtained by encoding the image data of the area corresponding to the database registration item of the document image data in the document by the character recognition means 6 is shown. This character string data cell 3
5, the character string data can be modified. At some point, only one of the character string data cells 35 must be
And the image data cell 34 in which the cursor 36 exists
And the character string data cell 35 are referred to as a current cell, and are to be corrected at that time.

The image data viewer 30 is a fixed display area for displaying image data of a corresponding page in a document corresponding to the current cell. In the image data of the image data viewer 30, a frame 37 is displayed in a region corresponding to the current cell. If the frame 37 is out of the display area, the image of the display portion is moved by a necessary minimum, and the frame 37 is displayed in the image data viewer at the maximum. The character string data viewer 31 is a fixed display area for displaying character string data of the current cell, and has a normal editor function. A cursor 38 is always displayed on the character string data viewer 31 so that character string data can be corrected.

As a modification of the character string data, the cursor 36 on the current cell or the character string data viewer 31
Alternatively, when correcting one character corresponding to 38, as shown in FIG. 5, a recognition candidate selection screen showing a list of recognition candidate characters pops up and displayed. In this figure, image data 40 for one character to be corrected is displayed, and
Recognition candidate characters 41 are displayed one by one with classifications such as kanji, hiragana, katakana, and alphanumeric characters. Therefore, the user selects a character to be corrected and selected by moving a pointing device such as a mouse or a cursor. If no correction is to be made, an instruction to display the end 42 is inputted.

As a result, the recognition candidate selection screen is closed, and the characters in the corresponding character string data are corrected at the positions of the cursors 38 and 36 on the character string data viewer 31 and on the current cell. When the character desired to be corrected is not displayed on the recognition candidate selection screen, the user can correct the character corresponding to the cursors 36 and 38 and the character string including the character by using an ordinary editor interface. Actually, when one of the cursor 38 on the character string data viewer 31 and the cursor 36 on the current cell moves, the other cursor moves so as to always keep the same position on the character string data. Further, when a correction is made in either the character string data viewer 31 or the current cell, the other character string data is always subjected to the same correction.

With this configuration, a database registration item area to be extracted is defined for each type of document, and image data of a plurality of types of documents read collectively based on this definition is collectively determined. By extracting necessary database registration items and encoding them into character string data by a character recognition unit, database registration items to be used for searching for document images can be input efficiently. Furthermore, by displaying and correcting the image data of the database registration items and the coded character string data for a plurality of types of documents collectively, the user's work load at the time of the correction operation can be reduced.

Embodiment 2 FIG. In the batch display correction interface of the first embodiment, the image data viewer 30 and the character string data viewer 31 are displayed in addition to the display of the document and the database registration items. As shown in FIG. 6, only characters and database registration items are displayed in a table format. In this case, the display of the current cell where the cursor 36 is present is always enlarged and displayed, and when the cursor 36 moves to another cell, the display of the source cell is returned to the normal size and the display of the destination cell is displayed. Enlarge the display.

In this way, the same effect as that of the first embodiment can be realized. In addition, if the display area is the same as that of the first embodiment, the database registration items corresponding to more documents can be displayed. Thus, the user's work load during the correction operation can be further reduced. When displaying database registration items corresponding to the same number of documents as in the first embodiment, the image data viewer 30 and the character string data viewer 3 are displayed.
The display area can be reduced by the amount that 1 is not displayed,
The display screen of the display unit 14 can be used effectively.

Embodiment 3 FIG. In the batch display / correction interface according to the first and second embodiments, a plurality of types of documents are collectively displayed in the order of batch reading. However, in this embodiment, only data of the same document type is displayed. And correct them as a unit. When the correction of one document type is completed, correct another type of document, and repeat the process to read all types of multiple documents. Make corrections. In addition, the display positions of the respective documents may be rearranged based on the types of the documents. In this way, the same effects as those of the first and second embodiments can be realized. In addition, when compared with the first and second embodiments, the displayed documents of the same type are compared with each other. Since the correction operation can be performed, the work load of the user on the correction operation can be further reduced.

Embodiment 4 FIG. In the batch display / correction interface of the first to third embodiments described above, a blank is displayed for an item in which there is no value of a certain database registration item of each document and a cell at that position does not exist. In the embodiment, as shown in FIG. 7, there is no value of the database registration item of each document, and the cells are displayed left-justified as much as there is no cell at that position. In this case, the item name displayed on the database registration item label 33 is the database registration item name corresponding to the document corresponding to the current cell. Each time the current cell moves, the item name displayed in the database registration item The item name displayed on the label 33 is changed. With this configuration, the same effects as those of the above-described first to third embodiments can be achieved. In addition, compared to the first to third embodiments, the database registration item corresponding to the document to be corrected can be obtained. By increasing the display amount, the user's work load for the correction operation can be further reduced.

Embodiment 5 FIG. In the image data viewer 30 in the batch display correction interface according to the first embodiment, the case has been described in which the displayed image data includes the moved database registration item area in accordance with the movement of the current cell. In this embodiment, the display area of the image data is moved using a pointing device such as a mouse,
By specifying an arbitrary point in the database registration item area in the image data, the cursor 36 of the document and database registration item table is moved. Further, the image data viewer 30 may be provided with an operation display for switching the display of the document image data to another sheet. By doing so, in addition to achieving the same effect as in the first embodiment, the user can visually perform a correction operation while viewing the display of the image data viewer 30, and the user can perform the correction operation at the time of the correction operation. Work load can be further reduced.

[0068]

As described above, according to the present invention, a series of image data obtained by reading a plurality of types of documents at once from an image reading apparatus is converted into document image data corresponding to each of the read documents. Input, pattern-recognize the document image data to generate coded data, and collectively display the document image data and coded data corresponding to each of a plurality of documents on a display unit. By post-correcting, and by registering the document image data and the coded data displayed on the display unit collectively in the database, a plurality of types of documents can be input collectively with a simple configuration and operation, A document filing method that can reduce the work load of the user in the correction work and improve the usability can be realized.

According to the next invention, a series of image data is compared with predefined standard document information, divided into a plurality of read documents, and the divided plural documents are compared with the standard document information. Then, by determining the type of the document, a document filing method capable of securely inputting a plurality of documents in a plurality of types with a simple configuration and operation can be realized.

According to the next invention, the image data of the database registration item area corresponding to the type of the document is extracted from the document image data corresponding to the document, and the image data of the database registration item area is subjected to the pattern recognition process. Document filing that allows multiple types of documents to be input collectively with a simple configuration and operation by coding and adding the coded database registration item values and registering the document image data in the database The method can be realized.

According to the next invention, the document image data and the coded database registration item values of a plurality of types of documents are collectively displayed in a table format in which the documents and database registration items are arranged. By correcting the database registration item value coded corresponding to the document in the database registration item table, a document filing method that can reduce the work load on the user in the correction work and improve the usability can be realized.

According to the next invention, the display and operation are switched according to the type of the document in the first and second display steps, so that the user can easily recognize the type of the document. Thus, it is possible to realize a document filing method that can reduce the work load of the user in the correction work and improve the usability.

According to the next invention, a first display mode for displaying a plurality of documents of the same type collectively for a plurality of different types of documents, and a second display mode for simultaneously displaying documents of different types in a mixed manner. Is switched in response to the user's request, so that if the user selects the first display mode, the contents of a plurality of documents of the same type can be compared. By selecting the mode, the contents of different documents can be compared, and these can be selected as necessary. Thus, a document filing method that can reduce the work load on the user in the correction work and improve the usability can be realized.

According to the next invention, the second display mode includes a fixed display mode in which the display position of the database registration item value that is different for each document type is fixed and displayed, and a different database registration for each document type. When the user selects the fixed display mode by switching between the left-justified display mode that displays the display position of the item value left-justified according to the user's request, the database registration for each document type You can easily recognize the presence or absence of items, and if you select the left-justified display mode, you can display and recognize more database registration items in the same document, you can select these as needed, and thus A document filing method that can reduce user work load and improve usability can be realized.

According to the next invention, in the correction process, the document image data corresponding to the database registration item value in the document to be corrected is always enlarged and displayed in the image data display area, and the code is displayed in the pattern data display area. The enlarged pattern data of the database registration item values is always displayed in an enlarged manner, so that the enlarged and displayed correction target pattern data and its document image data can be compared and recognized, thereby reducing the user's work load in the correction work. It is possible to realize a document filing method that can reduce and improve the usability.

According to the next invention, in the correction process, the document image data of the entire page including the database registration item value in the document to be corrected and the coded database registration item are displayed in the fixed display area. By always displaying the value pattern data at the specified display magnification, the pattern data to be corrected displayed at the specified display magnification and its document image data can be compared and recognized, and the user in the correction work can be recognized. A document filing method which can reduce the work load of the document and improve the usability can be realized.

According to the next invention, the display of the document image data of the entire page in the fixed display area is performed by enclosing the area on the document image data corresponding to the database registration item value to be corrected with a frame. Alternatively, the document image corresponding to the database registration item value to be corrected is displayed by displaying the color and moving the area on the document image data corresponding to the database registration item value to be corrected to the display area by the minimum necessary. A document filing method capable of reliably recognizing data, reducing the user's work load in correction work, and improving usability can be realized.

According to the next invention, when the user corrects the correction process so that the document type is different from the document type determined by the document determination process, only the document whose document type has been corrected is a document. By re-processing according to the processing procedure corresponding to the document type defined by the item definition process, even if a document is erroneously determined, it can be easily corrected, and multiple types of documents can be reliably collected at once. A document filing method that can be input can be realized.

According to the next invention, the keyword extraction step of the database registration processing step extracts a word from coded data obtained by encoding the entire document by the character recognition step, extracts the word as a keyword, and registers the keyword as a database registration item. Is designated, by adding the extracted keyword and registering it in the database, the keyword can be easily extracted and registered in the database, and a document filing method that can improve the usability of the user can be realized.

According to the next invention, the keyword extraction area is limited from the database registration item values, and the keyword is extracted from the coded data obtained by encoding only the designated area by the character recognition process. By assigning the extracted keyword and registering it in the database, the database registration item value can be easily extracted as a keyword and registered in the database, and a document filing method that can improve the usability of the user can be realized.

According to the next invention, the keyword extraction range is specified in the keyword extraction range specifying step in accordance with the contents of the document, and the entire document is specified in the coded data obtained by coding the character recognition step. By extracting keywords from the extracted extraction range, adding the extracted keywords and registering them in the database, valid keywords can be extracted and registered in the database, and a document filing method that can improve the usability of the user can be realized. .

According to the next invention, the document image data input means inputs a series of image data obtained by reading a plurality of types of documents collectively from the image reading apparatus as the corresponding document image data, The recognition means generates the coded data by pattern recognition, the display means collectively displays the document image data and the coded data corresponding to the plurality of documents on the display unit, and the correction means displays the display result on the display unit. Based on the necessity, post-correction is provided to the display means, and the document image data and the coded data displayed on the display unit are collectively registered in the database by the database registration means. A document filing device capable of collectively inputting a plurality of types of documents, reducing the user's work load in correction work, and improving usability. It can be realized.

According to the next invention, in the document image data input means, the series of image data is compared by the document classification means with the predetermined standard document information, divided into a plurality of read documents, and the document is determined. By comparing the plurality of documents classified by the means with the standard document information and determining the type of the document, and sending the same as document image data, the plurality of documents can be reliably classified into a plurality of types with a simple configuration and operation. A document filing device that can collectively input data to a document can be realized.

Further, according to the next invention, in the document image data input means, the document item definition means further includes a database registration item area indicating a description position of a database registration item value for each document type in addition to the standard document information. Predefined
Extracting the image data of the database registration item area from the document image data corresponding to the document by the database registration item extraction means, adding the database registration item value coded by the pattern recognition means, and registering the document image data in the database Accordingly, a document filing apparatus capable of reliably inputting a plurality of types of documents collectively with a simple configuration and operation can be realized.

According to the next invention, the display means collectively displays the document image data and the coded database registration item values of a plurality of types of documents in a table format in which the documents and the database registration items are arranged. In the correction means, by correcting the database registration item value coded corresponding to the document in the table of the document and the database registration items, the user's work load in the correction work can be reduced and the usability can be improved. A document filing device can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of a document filing system according to the present invention.

FIG. 2 is a schematic diagram showing a head identification page used in batch input of the document filing system of FIG. 1;

FIG. 3 is a schematic diagram illustrating a processing flow of the document filing system of FIG. 1;

FIG. 4 is a schematic diagram illustrating batch display and correction operations in the document filing system of FIG. 1;

FIG. 5 is a schematic diagram illustrating a recognition candidate selection screen in the batch display and correction operation in FIG. 3;

FIG. 6 is a schematic diagram illustrating batch display and correction operations in a document filing system according to a second embodiment of the present invention.

FIG. 7 is a schematic diagram illustrating batch display and correction operations of a document filing system according to a fourth embodiment of the present invention.

FIG. 8 is a schematic diagram illustrating display and correction operations in a conventional document filing system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Control part 2 Document reading means 3 Document classification means 4 Document discriminating means 5 Database registration item extraction means 6 Character recognition means 7 Batch display means 8 Correction means 9 Keyword extraction means 10 Database registration means 11 Storage part 12 Document reading part 13 Document items Definition unit 14 Display unit 15 Operation unit 20 Document 21 Document image data 22 Definition 23 Image data of database registration item area 24 Character string data 25 Batch display correction interface 26 Database 30 Image data viewer 31 Character string data viewer 32 Document type label 33 Database Registration item label 34 Image data cell 35 Character data cell 36 Cursor 37 Frame 38 Cursor 40 Image data for one character 41 Recognition candidate character 42 End display

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Yoichi Fujii 5-1-1, Ofuna, Kamakura-shi Mitsubishi Electric Corp. Personal Information Equipment Development Laboratory (72) Inventor Yuzo Maruta 5-1-1, Ofuna, Kamakura-shi No. Mitsubishi Electric Corporation Personal Information Equipment Development Laboratory (56) References JP-A-6-52236 (JP, A) JP-A-5-303619 (JP, A) JP-A 1-1111285 (JP, A) JP-A-3-263182 (JP, A) JP-A-6-68165 (JP, A) JP-A-5-54120 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 1/00 G06K 9/03

Claims (16)

(57) [Claims] 1. A document image data input step of inputting a series of image data obtained by reading a plurality of types of documents collectively from an image reading device as document image data respectively corresponding to the read plurality of documents; A pattern recognition step of pattern-recognizing and coding the document image data input in the document image data input step and generating coded data; and the document image data and the coded data respectively corresponding to the plurality of documents. A collective display on the display unit, and the necessity of post-correction of the coded data is determined based on a display result of the first display process, and the code is determined based on the determination result. A correction process for post-correction of the coded data, a second display process for displaying the coded data corrected in the correction process on the display unit, and a display process on the display unit. A database registration processing step of collectively registering the indicated document image data and the coded data in a database .
From the database registration item area corresponding to the type of the above document
Extracting the image data, and performing the pattern recognition process in the document image data input process.
The image data of the database registration item area extracted in
The first and second display steps are performed by coding the data into a database registration item value.
Code the image data of the item area and this image data
The registered database entry values with the document and database
A document filing method characterized by displaying registered items collectively in a table format . 2. The document image data input step includes the steps of: comparing the series of image data with predetermined standard document information; dividing the read image into a plurality of read documents; 2. The document filing method according to claim 1, further comprising: a document discriminating step of comparing the divided documents with the fixed document information to determine the type of the document and transmitting the document type as the document image data. Method. 3. The document image data input process further includes, for each type of the document, in addition to the fixed document information, a database registration indicating a description position of a database registration item value when the document is registered in the database. A document item definition process in which an item area is defined in advance, and a database registration item extraction process of extracting the image data of the database registration item region according to the type of the document from the document image data corresponding to the document. The image data of the database registration item area extracted in the database registration item extraction step is coded in the pattern recognition step, and the coded database registration item value is added. 3. The document image data is registered in the database. Document filing method described. 4. The document filing method according to claim 3 , wherein display and operation of the first and second display processes are switched according to the type of the document. 5. A method for a plurality of documents of different types,
A first display mode in which the same type of documents are displayed together and a second display mode in which different types of documents are mixed and displayed at the same time are switched according to a user request. The document filing method according to claim 4 . 6. The second display mode includes a fixed display mode in which a display position of the database registration item value different for each document type is fixedly displayed, and a database registration item different for each document type. 6. The document filing method according to claim 5 , wherein the display mode is switched between a left-justified display mode in which the display position of the value is displayed left-justified according to a user's request. 7. An image data display area for constantly enlarging and displaying said document image data corresponding to said database registration item value in said document to be corrected, said coded database registration item 4. The document filing method according to claim 3 , further comprising a pattern data display area for constantly displaying the value pattern data in an enlarged manner. 8. The method according to claim 1, wherein the correcting step includes: the document image data of the entire page including the database registration item value in the document to be corrected; and the coded pattern of the database registration item value. 4. The document filing method according to claim 3 , further comprising a fixed display area that is always displayed at a specified display magnification. 9. The display of the document image data of the entire page in the fixed display area is performed by enclosing an area on the document image data corresponding to the database entry item value to be corrected with a frame. 9. The document according to claim 8 , wherein a color is displayed, and an area on the document image data corresponding to the database registration item value to be corrected is moved to a minimum necessary to fit in the display area. Filing method. 10. When the user corrects the correction process so that the type of the document is different from the type of the document determined by the document determination process, only the document in which the type of the document is corrected is the document. 4. The document filing method according to claim 3 , wherein reprocessing is performed in accordance with a processing procedure corresponding to the type of the document defined by an item definition process. 11. The database registration processing step includes a keyword extraction step of extracting a word from the coded data obtained by coding the entire document by the character recognition step and extracting the extracted word as a keyword. If a keyword is specified,
4. The document filing method according to claim 3, wherein the keyword extracted in the keyword extraction step is added and registered in the database. 12. An area for extracting the keyword from the database entry item value is limited, and the keyword is extracted from the coded data obtained by coding only a designated area by the character recognition process. 12. The document filing method according to claim 11 , wherein the keyword is added and registered in the database. 13. A keyword extraction range designating step for designating a range in which the keyword is to be extracted in accordance with the contents of a document, and a designated extraction range of the coded data obtained by encoding the entire document by the character recognition process. 12. The document filing method according to claim 11 , further comprising extracting the keyword, adding the extracted keyword, and registering the extracted keyword in the database. 14. A document image data input means for inputting a series of image data obtained by reading a plurality of types of documents collectively from an image reading device as document image data respectively corresponding to the read plurality of documents; Pattern recognition means for pattern-recognizing and coding the document image data input by the document image data input means and generating coded data; and the document image data and the coded data respectively corresponding to the plurality of documents. Display means for collectively displaying the coded data on the display unit, determining whether or not post-correction of the coded data is necessary based on the display result of the display unit, and post-correcting the coded data based on the determination result. Correction means for supplying the document image data and the coded data displayed on the display unit to the database Database registration processing means for registering, wherein the document image data input means
From the database registration item area corresponding to the type of the above document
Extracting image data, wherein the pattern recognition means comprises:
The image data of the database registration item area extracted in
The data is coded into a database registration item value, and the display means displays an image of the database registration item area.
Data and database that coded this image data
Array of registration item values and document and database registration items
A document filing apparatus characterized by displaying in batches in a table format . 15. The document image data input means compares the series of image data with predetermined standard document information and classifies the read document into a plurality of read documents. 15. The document filing device according to claim 14 , further comprising: a document discriminating unit that compares the divided documents with the fixed document information to determine the type of the document and sends the document image data as the document image data. apparatus. 16. The document image data input means further includes, for each type of the document, in addition to the standard document information, a database registration indicating a description position of a database registration item value when the document is registered in the database. A document item definition means in which an item area is defined in advance; and a database registration item extraction means for extracting the image data of the database registration item area corresponding to the type of the document from the document image data corresponding to the document. The image data of the database registration item area extracted by the database registration item extraction means is coded by the pattern recognition means, and the coded database registration item value is added. claim the document image data and registers to the database 1 6. The document filing apparatus according to 5 .