JP4147014B2 - Document processing system, method, program, and storage medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention compares a document (form) image read from a scanner or the like with a pre-registered image (registration form) and corrects a positional deviation or a size difference between the images, a method, and a program And a storage medium.
[0002]
[Prior art]
An input form image read from a scanner is compared with a plurality of registered forms (registration forms) registered in a database in advance to specify the format of the form image.
[0003]
[Problems to be solved by the invention]
However, comparing the input form image read from the scanner with the registration form registered in the database in advance and specifying a similar registration form from the database means that the input form image read from the scanner has a tilt or the like, There was a problem that the matching accuracy deteriorated.
[0004]
In addition, if the read input form image has a positional deviation, inclination, or difference in resolution at the time of scanning, the character recognition accuracy is improved when character recognition processing of characters included in a predetermined area of the input form image is performed. There was a problem of getting worse.
[0005]
[Means for Solving the Problems]
In order to solve the above problems, a document processing system of the present invention is a document processing system that registers document image data and an index in a document management unit that manages document image data. A template creation unit for obtaining an index structure of the database of the document management unit, creating a template based on the obtained index structure and connection information to the database of the document management unit, and storing the created template; Based on the template, a worksheet creation means for creating a worksheet having items for use in index registration; Storage means for storing a plurality of registration forms; An area registration means for setting an area on the registration form and associating the set area with an item of the worksheet; Feature extraction means for extracting feature information from the input document image data, comparing the feature information of the extracted document image with the feature information of each of the stored registration forms, A difference between the registration form and the document image based on the form identification means for identifying the registration form determined to match the comparison result, the feature information of the identified registration form, and the feature information of the document image A difference extracting means for extracting the image, an image correcting means for correcting the document image based on the extracted difference, and the corrected document image based on an area set in the identified registration form. Area extracting means for extracting the corresponding area, recognition means for recognizing the image of the extracted area, and the recognition result by the recognition means , The item corresponding to the area As an index, Worksheet registration means for registering in the worksheet, determination means for determining whether or not to continue to input other document image data, and if the determination means determines to continue input, the other document image data While performing processing by the feature extraction unit, the form identification unit, the difference extraction unit, the image correction unit, the area extraction unit, the recognition unit, the worksheet registration unit, and the determination unit, If the determination means determines that the input is not continued, the index registered in the worksheet is Registration means for registering with the document image data together with the document image data.
[0006]
In order to solve the above problems, a document processing method of the present invention is a document processing method for registering document image data and an index in a document management unit that manages document image data, A template creation unit obtains an index structure of the database of the document management unit, creates a template based on the obtained index structure and connection information to the database of the document management unit, and stores the created template A template creation step, a worksheet creation step in which the worksheet creation means creates a worksheet having items to be used for index registration based on the template, and an area registration means are stored in the storage means. An area registration step of setting an area on the registration form, associating the set area with the item of the worksheet, and a feature extraction unit, A feature extraction step for extracting feature information from the input document image data; Form identification means Feature information of the extracted document image; Above A form identification step for comparing the feature information of each of the plurality of registered forms stored in the storage means and identifying the registered form determined to match the comparison results; The difference extraction means A difference extracting step of extracting a difference between the registered form and the document image based on the identified registration form feature information and the document image feature information; Image correction means An image correcting step for correcting the document image based on the extracted difference; Area extraction means An area extraction step of extracting a corresponding area of the corrected document image based on an area set in the identified registration form; Recognition means A recognition step for recognizing the image of the extracted area; Worksheet registration means The recognition result in the recognition step , The item corresponding to the area As an index, If the worksheet registration step to be registered in the worksheet, the determination unit determines whether to continue inputting other document image data, and the determination step determines to continue the input, One of repeating the feature extraction step, the form identification step, the difference extraction step, the image correction step, the area extraction step, the recognition step, the worksheet registration step, and the determination step for the document image data of When the determination unit determines that the input is not continued, the registration unit displays the index registered in the worksheet, And a registration step of registering in the document management unit together with the document image data.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 shows a system configuration diagram in the present embodiment.
[0008]
Here, on a network, a user terminal (100, 101, 102), an MFP (multifunction peripheral) (110) having a fax / scanner / printer / copy function, a document registration system (120), a file server (130), A document management system (140) is connected.
[0009]
Each user terminal uses the file server 130 and the document management system 140 to share various documents. There are various types of documents such as a document created by a general document creation application, a document read by a scanner, and a document received by fax.
[0010]
In this embodiment, the MFP 110 having a plurality of functions is used as the document input device. However, a single-function device such as a scanner device or a facsimile device may be used instead of the MFP.
[0011]
The file server 130 manages mainly documents stored temporarily. For example, a document image read by the scanner unit of the MFP 110 can be temporarily stored, and the user terminal or the document registration system 120 can acquire the document image.
[0012]
The document management system 140 mainly manages documents that need to be stored for a long period of time. The document transmitted from the document registration system is stored together with an index, and a user terminal is requested to transmit a desired document. In this case, the document can be searched and transmitted based on an index or the like. In the document management system, a plurality of indexes can be attached to one document, and the plurality of indexes are classified and managed for each index item. In the present embodiment, the document management system 140 is configured by a single document management apparatus, but may be configured by a combination of a plurality of devices. A plurality of document management systems 140 may exist.
[0013]
The document registration system 120 automatically or manually specifies and acquires a document (fax received document image, scanner-read document image, etc.) generated by the MFP 110 or a user terminal, or a document stored in a file server. An index (attribute information / search information) of the document is automatically created based on information obtained when the document is acquired or information obtained by analyzing the document, and the document and the index are combined. It has a function of registering in the document management system 140 (desired directory). In addition, it has a function of adding various indexes manually entered by the user and registering the document in the document management system, a function of re-editing attribute information of the document already registered in the document management system, and the like. In the present embodiment, the document registration system 120 is configured by a single document registration apparatus, but may be configured by a combination of a plurality of devices. In addition, using multiple document registration devices and giving authority to perform processing for each device so that multiple users can perform indexing and confirmation operations, etc., coordinate processing with multiple document registration devices By doing so, the document registration system may be configured.
[0014]
In the present embodiment, the document registration apparatus stores a CPU that performs actual processing operations, a RAM that reads a program and uses it as a work area, a program for executing processing corresponding to a flowchart to be described later, and various data. Storage media (hard disk, ROM, removable disk (floppy disk, CD-ROM, etc.)), keyboard and pointing device for various operations, display for displaying documents to be processed, and network connection It is a computer configured with a network interface. The program to be executed by the CPU may be supplied from the storage medium or may be read from an external device via a network. In addition, although this embodiment shall be implement | achieved when a computer runs a program, you may make it comprise the one part or all by hardware.
[0015]
FIG. 2 is a block diagram showing the structure of the document registration system 120.
[0016]
<Document input processing to the document registration system>
First, document input processing to the document registration system 120 (document acquisition processing by the document registration system) will be described.
[0017]
The device control unit 210 obtains document data from a device (MFP) 110 having a fax function, a scanner function, and the like, and inputs it to subsequent processing. Note that the device control unit 210 performs processing to download document data by periodically monitoring the device or to receive document data transmitted by the push method from the device side.
[0018]
Reference numeral 130 denotes a folder (storage device) in which a document in a file server or the like is stored, and the folder control unit 212 acquires document data in the folder of the file server. For example, the folder control unit 212 periodically monitors a predetermined folder on the file server, and performs processing for acquiring the document when there is a new document.
[0019]
Reference numeral 201 denotes an operation unit in this system, which can set and operate various systems. Further, when a user operating the document registration system designates a document, the user terminal or file server can be used. The document data can be imported (acquired), and input processing of the document data acquired by the manual designation can be performed via the file control unit 214.
[0020]
An input control unit 220 can control the input from the device control unit 210, the folder control unit 212, the file control unit 214, and the like.
[0021]
<Internal processing of document registration system>
Next, processing in the document registration system will be described.
[0022]
Reference numeral 221 denotes a job control unit that performs control related to a job to be described later.
[0023]
A document data processing unit 222 performs an image format conversion process and the like.
[0024]
Reference numeral 223 denotes an index processing unit that performs control related to index processing to be described later.
[0025]
A data management unit 224 saves and manages data on templates, worksheets, jobs, and the like, which will be described later, in a data storage area 250.
[0026]
Reference numeral 225 denotes an OCR control unit that performs character recognition using the dictionary data 251.
[0027]
A form (form) analysis unit 226 can automatically identify various forms such as a form format using the registered form data 252.
[0028]
Reference numeral 240 denotes a mail processing unit. When the automatic processing is performed when the document is automatically acquired and the registration processing is performed, the result is notified to a destination registered in advance. Used to do.
[0029]
<Access processing to the document management system in the document registration system>
Next, an interface with the document management system in the document registration system will be described.
[0030]
Reference numeral 227 denotes a document management system control unit, which registers documents and indexes, and documents already registered in the document management system via the access control unit 230 provided according to the type of the document management system 140. Perform processing such as index change.
[0031]
A database 253 stores access data necessary for accessing the document management system. A data area 254 temporarily stores data to be registered in the document management system. This area is used as a buffer when document data processing is required.
[0032]
<< Basic Flow of Document Registration System (Fig. 3) >>
FIG. 3 is a basic flowchart in the document registration system 120.
[0033]
First, a processing flow when a document (image) is manually acquired, an index (attribute information) is input, and the document and the index are registered in the document management system will be described.
[0034]
In step S100, the index structure of the database of the document management system is acquired, and a template (model) for creating a worksheet is created from the index structure. This template includes access information to the document management system and schema information (item information that can be used as an index) created from the index structure. Details of the template creation processing will be described later with reference to FIG.
[0035]
In step S101, a worksheet is created from the template. This creation process is described in detail in FIG. In this worksheet, fields (items) and the like for actually registering index information are set. In this embodiment, management of acquired documents, index information added to the documents, and the like are performed and managed in a series of processes (as a management table) in units of the worksheets. In this example, a worksheet is created from the template created in step S100. However, step S100 is omitted, a template created and saved in advance is selected, and a worksheet is created from the selected template. You may do it.
[0036]
It is also possible to select a worksheet that has already been created / registered and input document information and index information to the worksheet. In this case, in step S110, the worksheet selected by the user is extracted from the worksheets stored in the data area 250, and the process proceeds to step S102.
[0037]
In step S102, the worksheet created in step S101 or the worksheet selected in step S110 is displayed on the display.
[0038]
In step S103, the document acquired from the device is input.
[0039]
In step S104, the document information is registered in the worksheet. Specifically, the input document name (document ID) is registered in the item (schema information) of the document name (document ID) of the worksheet, and the document registered in the worksheet can be identified.
[0040]
In step S105, necessary index (attribute) information is input and managed for each item of the worksheet. The index information may be input manually by the user, or information obtained by automatically performing OCR or the like on the document may be input. Details of the index information input will be described later with reference to FIG.
[0041]
If it is necessary to input another document in step S106, the process returns to step S103.
[0042]
In step S107, the data in the worksheet acquired / set by the above processing is released (registered) to the document management system. Details of the release process will be described later with reference to FIG.
[0043]
Next, when a document input process is set in advance on an input activation button on a GUI (graphical user interface) (FIG. 8) and the input activation button is pressed, the process is executed. The process will be described.
[0044]
When it is detected in step S120 that the input activation button has been pressed, a process preset for the button is determined. In step S121, a document acquisition process corresponding to the set condition (for example, designated in advance) is determined. Process to acquire a document newly stored in the specified folder, or to acquire a document image by reading a document with a scanner specified in advance, and register it in a preset worksheet To do. The execution of the input activation button process will be described later with reference to FIG.
[0045]
In step S122, an index is input to the worksheet. Details of the index input process will be described later with reference to FIG.
[0046]
In step S123, the data in the worksheet acquired / set by the above process is released (registered) to the document management system. Details of the release process will be described later with reference to FIG.
[0047]
In addition, a document acquisition setting, index setting, and release setting are designated and created and saved as a job in advance, and the job selected by the user from the plurality of saved jobs (step S130) is executed (step S130). Step S131) can also be performed. Details of the job selection / execution process will be described later with reference to FIG.
[0048]
Although details will be described later, it is also possible to automatically input an index using a form when inputting an index in steps S105 and S122. In step S140, processing for registering the form is performed, and in step S141, an area is set in the form. Details of the form registration process will be described later with reference to FIG. 6, and details of the area registration process will be described later with reference to FIG.
[0049]
In step S150, various setting processes (job setting process (FIG. 12), input activation button setting process (FIG. 8), etc.) for executing the above processes are performed.
[0050]
The processing described above will be described in detail below.
[0051]
<Template creation process (FIG. 4)>
The template creation process (step S100) will be described in detail with reference to FIG.
[0052]
In step S200, a connection is first made to the document management system of the document registration destination.
[0053]
In step S201, an index structure of a database in the document management system (schema information (information that can be used as a search index) set in a folder in which a document is stored) is acquired.
[0054]
In step S202, a template that collectively manages connection information to the document management system and schema information is created, and identification information (template name) for identifying the created template is added and saved.
[0055]
<Worksheet creation process (FIG. 5)>
The worksheet creation process (step S101) for creating a worksheet from a saved template will be described in detail with reference to FIG.
[0056]
In step S300, first, the user is allowed to select a template corresponding to the document management system of the document registration destination. Note that this selection step may be omitted when a worksheet is created directly from the template created in FIG.
[0057]
In step S301, if the schema information of the document management system stored in the template does not need to be indexed, is editing such as selecting a field corresponding to the schema information as an input target? Judge whether the instruction was given.
[0058]
If it is determined that field editing is to be performed, the field editing operation is performed in step S302. Basically, this is an operation for selecting a necessary field from item areas (fields) of a template that can be used as schema information. In some cases, an extension of the area (field) can be instructed.
[0059]
In step S303, a predetermined setting such as field editing is confirmed, and a worksheet is created from the selected template.
[0060]
In step S304, identification information (worksheet name) for identifying the created worksheet is added and saved.
[0061]
<Form registration process (Fig. 6)>
A form registration process (step S140) for registering a form for use in a process for identifying an input document and a process for extracting information from an image as an index will be described in detail with reference to FIG.
[0062]
In step S400, image data used as a reference at the time of form analysis is acquired.
[0063]
In step S401, the previously obtained image data is registered in a form management sheet for managing various forms.
[0064]
In step S402, the image data is passed to the form analysis unit 226 that performs form analysis processing (analyzes position and size information about the frames and tables constituting the form, written character information, etc.) Data form analysis processing is performed, the form analysis result is registered in the form management sheet in association with the image data, and managed as registered form data 252.
[0065]
In step S403, the newly added registration form data is added with identification information (form registration name) and saved.
[0066]
<Area registration processing (FIG. 7)>
The area registration process (step S141) for setting an area for performing character recognition, barcode recognition, etc. on the registration form will be described in detail with reference to FIG. This area is used for recognizing the document area corresponding to the area set in the registration form and registering the recognition result as an index in the worksheet when indexing the document. .
[0067]
In step S500, the user first selects a form for setting an area.
[0068]
In step S501, the user is allowed to specify an area to be a form analysis target area.
[0069]
In step S502, the analysis mode of the area is designated. Detailed conditions for improving the recognition rate are set as the analysis mode. For example, designation of alphanumeric character recognition, kana / kanji character recognition, analysis engine such as barcode recognition, and the like can be designated for each area. You can also set the correspondence between areas and template items (worksheet items) here. When analyzing and registering document images, the recognition results for each area are displayed in the corresponding worksheet. The item is entered as an index. The association between the area and the worksheet item (template item) may be changed when the index input process is performed.
[0070]
In step S503, it is determined whether another area setting instruction has been issued. If another area is set, the process returns to step S501. If the area setting is completed, the process proceeds to step S504.
[0071]
In step S504, the setting information regarding the set area is registered as additional information for the registration form data.
[0072]
<Input activation button setting (FIG. 8)>
The input activation button setting process (step S150) will be described in detail with reference to FIG. When the input activation button is pressed, input processing from various data sources (folders, scanners, etc.) is performed at once.
[0073]
In step S600, the user is allowed to select a worksheet for registering a document acquired from the set data source when the input activation button is pressed.
[0074]
In step S601, the user is made to select a data source (folder, scanner, etc.) from which document data is acquired.
[0075]
In step S602, it is determined whether the scanner is designated as the data source. If the scanner is designated, the scanner reading mode (reading resolution, reading area, etc.) is set in step S603.
[0076]
In step S604, it is determined whether the folder is designated as a data source. If a folder is designated, folder access settings (access conditions such as access right settings) are performed in step S605.
[0077]
In step S610, the set condition is registered in association with the input activation button, and the input activation button is displayed as a GUI. Also, a plurality of input activation buttons can be registered and displayed by changing the button name and shape according to the setting conditions.
[0078]
<Input activation button execution processing (FIG. 9)>
The execution process (step S121) when the input activation button is pressed will be described in detail with reference to FIG.
[0079]
In step S700, the condition (setting mode) set for the pressed input activation button is read.
[0080]
In step S710, it is determined whether the scanner is designated. If the scanner is set, the process proceeds to step S711 to perform scanning according to the scanner reading condition set as the setting mode to obtain a document image. In step S712, the document is saved in the worksheet.
[0081]
In step S720, it is determined whether or not document acquisition from a folder is set. If acquisition from a folder is set, the process proceeds to step 721, and a condition for acquiring a document from the folder (for example, a newly input document) In step S722, the content of the folder is checked. In step S722, whether there is a document satisfying the condition is determined. In step S723, the document is acquired, and in step S724, the document is stored in the worksheet. On the other hand, if it is determined in step S722 that no document satisfies the condition, the process proceeds to step S725.
[0082]
Step S725 is a determination process used when a condition for acquiring a document by periodically checking (polling) the folder is set, and when the cancellation of the periodic folder check is instructed or originally. If the setting conditions for periodic folder confirmation have not been made, the process ends. If not, the process returns to step S721 again to periodically confirm the contents of the designated folder.
[0083]
<Index input process (FIG. 10)>
The index input process (S105, S122, S1121, etc.) for inputting a document index to the worksheet will be described in detail with reference to FIG.
[0084]
In step S800, it is determined whether or not automatic processing such as a job has been specified. If automatic processing is selected, the processing proceeds to step S815. If not, worksheets and forms are sequentially selected.
[0085]
If it is not automatic processing, in step S801, the user selects the worksheet by designating the worksheet.
[0086]
In step S802, it is determined whether the index is input manually by the user or the index is input using a registration form. When it is instructed by the user to manually input index information, manual index input is performed in step S803, and then it is determined in step S804 whether there are other documents to be processed, and there are other documents. Returns to step S800, and if not, the process ends.
[0087]
On the other hand, if it is determined in step S802 that an instruction to input an index using the registered form has been given, the process proceeds to step S810 to determine whether to use the automatic form detection function. If it is determined that the use of the automatic form detection function has been instructed, in step S811, a form detection process for detecting a form that matches the input document from a plurality of registered forms is performed. If a matching form is detected by the form detection process, the process proceeds from step S812 to step S820. If a matching form is not detected, the process proceeds from step S812 to step S813. On the other hand, if it is determined in step S810 that use of the automatic form detection function is not instructed and manual form selection is instructed, the process proceeds to step S813. In step S813, a desired form is manually selected from a plurality of registered forms according to a user instruction.
[0088]
In step S820, an area set in the registration form selected automatically or manually is searched.
[0089]
In step S821, it is determined whether or not the indexing process using the information in the area corresponding to all areas set in the registration form has been completed in the document. If the indexing process has been completed, the process proceeds to step S804. On the other hand, if not completed, the process proceeds to step S822.
[0090]
In step S822, area analysis is performed according to the area analysis conditions set in the registration form. Here, it is also possible for the user to actively adjust the area range set as the area analysis condition. The adjustment of the area range is performed by displaying the area range set as the analysis condition of the set area in a state where the input document image is displayed, with a rectangle of a predetermined color superimposed. The area range can be adjusted by adjusting the rectangle. Further, the character image smoothing process (black pixel block edge smoothing process) is performed on the image data cut out in accordance with the region range. The character image here includes all of the meanings transmitted in the form of the bit arrangement of the bitmap such as barcode data and pictographs. The smoothing process of the character image is included in the OCR control unit 225 in the form of library software. By performing this smoothing process, jaggies (dot variations) that occur when image processing such as tilt correction, shift correction, and magnification correction is performed are reduced, and the recognition rate of the character recognition process can be improved.
[0091]
In step S823, it is determined whether the area analysis process has succeeded without any abnormality. If the process ends abnormally, the process proceeds to step S824, where it is determined whether the process is in the automatic process mode. If the process is not in the automatic process mode, the user manually inputs an index in step S825. Do. Here, the provisional process is, for example, a process of adding an NG mark indicating that the area analysis is unsuccessful, or outputting a value defined as another default as an index.
[0092]
In step S830, the result of area analysis processing is saved in the worksheet, and the process returns to step S820 to search for another unprocessed area.
[0093]
On the other hand, if it is determined in step S800 that automatic processing such as a job has been designated, in step S815, a worksheet associated with the automatic processing is selected, and preprocessing for automatic indexing processing is performed. Here, the automatic processing is premised on that a worksheet is set in advance. If the registration form is also designated in advance in the automatic processing, the process proceeds to step S820 assuming that the form is detected in step S816. If the setting is such that the registration form is automatically detected and selected in the automatic process, the form detection process similar to step S811 is performed in this pre-process, and a matching registration form is detected. If the form is detected in step S816, the process proceeds to step S820. On the other hand, if the registered form cannot be determined, the process proceeds to post-processing in step S817, and the document is displayed on the worksheet as a document that could not be processed. In step S804, it is determined whether to process the next document.
[0094]
<Release (data registration) processing (FIG. 11)>
A process (S107, S123, S1131, etc.) for releasing (registering) data (document information and index information) managed in the worksheet in the document management system will be described in detail with reference to FIG.
[0095]
In step S900, it is determined whether automatic release is designated. If it is determined that the release is automatic release, the process proceeds to step S901. If it is determined that the automatic release is not performed, the process proceeds to step S910.
[0096]
In the case of automatic release, in step S901, automatic release setting information (connection information to the registration destination document management system, registration destination folder information, etc.) is first acquired from the worksheet, and in step S902, the document management system A connection process to the storage folder is performed, and the process proceeds to step S920 to prepare data to be released.
[0097]
If it is not automatic release, in step S910, first, from the document managed in the worksheet, the selection of the document to be registered and the handling (deletion) of the document stored in the document registration system after registration in the document management system are performed. It can be set. To select a document to be registered, a plurality of registration target document IDs managed in the worksheet and an index corresponding to each document are displayed in a table format of the worksheet, and a registration process is performed from the list. A document can be selected. During the selection process, the document ID and index selected on the worksheet are highlighted so that the user can easily understand that the document is selected.
[0098]
In step S911, login processing is performed to the document management system as a registration destination, and a folder for storing the document is designated.
[0099]
In step S912, options can be set at the time of document registration, and an execution log related to document registration processing can be set, and conversion settings for the document format at the time of document registration can be set.
[0100]
In step S920, document information to be registered in the document management system is temporarily stored in the internal output data area 254 to prepare for registration.
[0101]
In step S930, it is determined whether a format conversion of the document image to be registered or a process for adding a result of OCR analysis of the document image is instructed. If it is determined that the instruction is instructed, in step S931, according to the instruction. Perform image processing.
[0102]
In step S940, the prepared document data and the like are registered in the folder of the registration destination document management system.
[0103]
In step S950, it is determined whether deletion of the document stored in the document registration system is instructed after registration of the document in the document management system. If deletion is instructed, deletion of the document is performed in step S951. The process is performed and the release process is terminated.
[0104]
<Job creation processing (FIG. 12)>
With reference to FIG. 12, a process (S150) for defining each setting of the document acquisition process, the index process, and the release process and creating it as one job will be described in detail.
[0105]
In step S1000, it is determined whether automatic document acquisition job settings have been specified. If so, the process proceeds to step S1001 to set a worksheet to be used, specify a scanner or folder as a source device, and the like. Set for acquisition.
[0106]
In step S1010, it is determined whether setting of automatic processing for index input has been instructed. If setting is performed, index processing such as setting of a worksheet to be used, designation of a registered form or designation of automatic search for a registered form is performed in step S1011. Set the items required to automatically process
[0107]
In step S1020, it is determined whether an automatic release (document registration) setting has been instructed, and if so, in step S1021, the worksheet to be processed, the document registration destination (document management system, folder), and the post-registration are registered. Set information necessary for automatic release processing such as document handling.
[0108]
In step S1030, the setting information set in steps S1001, S1011, and S1021 is added with identification information (job name) and stored.
[0109]
It is possible to register and save a plurality of jobs combining various settings.
[0110]
<Job execution processing (FIG. 13)>
The job execution process (S130, 131) will be described in detail with reference to FIG.
[0111]
In step S1100, the user selects a desired job from a plurality of registered and stored jobs.
[0112]
In step S1101, the conditions set for the selected job are read and job execution processing is started.
[0113]
In step S1110, it is determined whether an automatic document acquisition job is set. If it is determined that a job is set, in step S1111, document acquisition processing is performed from the device according to the setting. In step S1112, the acquired document is converted into a worksheet. Manage and save with.
[0114]
In step S1120, it is determined whether or not automatic index input processing is set. If it is determined that index input is set, automatic index processing is executed in step S1121 according to the set conditions. In the automatic index processing, when automatic form detection is performed, image processing such as various zooming and misalignment correction is performed, and information described on the image is acquired according to area (region) settings. Further, by performing smoothing processing on the image information and performing character recognition (OCR) processing, index data based on text data can be automatically created.
[0115]
In step S1130, it is determined whether automatic release (document registration) processing is set. If it is determined that it is set, in step S1131, automatic release processing is performed to the document management system according to the set conditions.
[0116]
By setting a job from document acquisition to document registration as a job, a series of processing can be performed at once.
[0117]
<Extended device control unit (FIG. 14)>
In the present embodiment, the device control shown in FIG. 2 is further performed, and the method for acquiring a document from the device can be extended to support a plurality of forms. The extended device control will be described with reference to FIG.
[0118]
The device control unit A (210) accesses the device side from the document registration system and acquires a document. Using this control unit A, it is possible to acquire document data from a plurality of devices (110, 111) compatible with an equivalent access acquisition method.
[0119]
The device control unit B (211) obtains a transmitted document by accessing from the device side, not from the document registration system. Also in this case, it is possible to acquire a document from a plurality of devices (112, 113) corresponding to the equivalent access method.
[0120]
In addition, the device control unit B is provided with a data folder 300 to support simultaneous document transfer processing from a plurality of devices. Furthermore, this data folder can be configured as a folder having a hierarchy, and document transfer processing by designating a folder path from the device side is also possible. By adopting a general-purpose file transfer protocol as an access protocol in this case, the device side can be easily handled and a document can be transferred from a general computer.
[0121]
Furthermore, when acquiring document data, in addition to the document image, additional (attribute) information related to the document data (for example, owner information of each document, transmission source information in the case of a fax reception document, information on a connected device, etc. ) And the data can be used in the system.
[0122]
The input control unit 220 can handle the document data acquired by each device control unit so as not to be confused.
[0123]
<Extended Document Management System Control Unit (FIG. 15)>
In the present embodiment, it is possible to expand so that document data acquired from a device or the like and its index (additional information) can be registered in a plurality of document management systems that manage documents in various different formats. FIG. 15 is a diagram for explaining the correspondence when the document management system A and the document management system B that manage documents in different formats coexist on the network.
[0124]
Here, the access means and the file format of a document that can be handled are generally different depending on the type of the document management system. Therefore, for the document management system A (140), the access control unit A (230), An access control unit B (231) is provided for the document management system B (141). The document registration system 120 can cope with a case where a plurality of document management systems that manage documents in different formats coexist by changing the access control unit to be used according to the document management system of the registration destination.
[0125]
Information necessary for accessing each document management system is stored as management data 253, and is extracted and used as necessary.
[0126]
<Supplement to Release Process in FIG. 11 (FIG. 16)>
A supplement regarding the release (registration) processing (step S940) to the document management system will be described with reference to FIG.
[0127]
Step S1200 is a setting process for starting the release process, and confirms whether or not an end notification is necessary, and obtains an end notification destination information acquisition process.
[0128]
Step S1210 is a process for confirming whether there is any remaining data to be registered in the document management system. If all the data to be registered has not been transmitted, the process proceeds to step S1211, and if there is no remaining data, the process proceeds to step S1220.
[0129]
In step S1211, registration processing is performed for transmitting and registering data to be registered in the document management system.
[0130]
In step S1212, it is determined whether or not the transmitted data is registered. If it is confirmed that the data has been registered, the process returns to step S1210. If the registration fails, the process proceeds to step S1213.
[0131]
Step S1213 is processing in the case where an error occurs in data registration in the document management system. Data relating to a document in which a registration error has occurred is registered as a registration error document data in a predetermined worksheet, and then manual re-registration processing is performed. Can be done.
[0132]
In step S1220, based on the setting in S1200, it is determined whether it is necessary to notify the user of registration completion. If not, the process ends.
[0133]
If it is determined that it is necessary to notify the end, in step S1221, it is determined whether there is an error in the result of the data registration process. If there is no error, in step S1222, a new document is sent to the pre-registered registration notification mail address. If it is determined that an error has occurred, on the other hand, if it is determined that an error has occurred, in step SS1223, the error information is sent to the pre-registered error notification email address. It has become.
[0134]
<Automatic Document Registration Processing for Documents Sent from Device (FIG. 17)>
FIG. 17 describes a flow for automatically processing the document data sent from the device until the document registration. In FIG. 17, document registration automatic processing to the document management system and form registration automatic processing to the document registration system are described as registration forms.
[0135]
In step S1300, various items necessary for executing automatic document processing are set in advance, and the set automatic document processing is activated.
[0136]
The setting items here include settings related to document input processing, folders used for form registration processing, setting of attribute information to be used, setting of worksheets for processing corresponding to each attribute, automatic index input There is a setting of the form used for performing.
[0137]
Examples of attribute information include the following. The attribute information is used for automatic processing discrimination and index input.
[0138]
・ Device type
Folder path (file server 130, device controller data folder 300)
・ Document owner information
・ Sender information in FAX received documents
・ Area analysis results, etc.
In step S1301, it is detected whether or not a document is input. If there is a document input, the process proceeds to step S1302. In the present embodiment, document data input from a device via a device control unit and document data input from a device via a folder control unit can be collectively handled. In the present embodiment, it is assumed that a document transmitted by push from a device is input via a data folder of the device control unit. Whether or not a new document has been input to the folder is detected by periodically checking the folder in the same manner as in S721 to S725 in FIG.
[0139]
In step S1302, the route through which the input document has entered is confirmed.
[0140]
In step S1310, it is determined whether the document is input via the form registration folder. If it is determined that the document is input via the form registration folder, the form registration process is automatically performed in step S1311. The document entered is registered as a form.
[0141]
In step S1320, various attribute information is analyzed, and it is confirmed whether or not a process corresponding to the obtained attribute information is defined. For example, it is assumed that a document input via a predetermined folder A or a document input from a predetermined device is defined to perform automatic processing for registering in a folder X of a predetermined document management system.
[0142]
As attribute information, information on devices, folders, and the like obtained by the input control unit 220, and information obtained from a predetermined area targeted by form analysis by calling a process corresponding to automatic index processing at this stage and so on.
[0143]
In step S1321, it is determined whether or not automatic processing corresponding to the attribute information of the input document has been detected. If not, the process proceeds to step S1323 to temporarily store the input document. Can be handled manually.
[0144]
On the other hand, when the corresponding process is detected (for example, when the registration process to the folder X of the predetermined document management system is detected), the process proceeds to step S1324, and information necessary for the automatic index process and the automatic release process is read. In step S1330, automatic index input is performed using area information corresponding to the form as necessary, and in step S1340, automatic release processing for registering the document and index in the document management system is performed.
[0145]
In step S1350, it is determined whether a request to end automatic document processing has been made. If so, the processing is terminated. If not, the processing returns to step S1301, and the processing is continued.
[0146]
<Configuration of form analysis unit (FIG. 18)>
FIG. 18 is a detailed configuration diagram of the form analysis unit (form processing unit).
[0147]
The form analysis unit includes an image processing engine 1801, an image feature extraction engine 1802, an image feature comparison engine 1803, an image processing setting 1804, an image feature registration engine 1805, an image feature processing engine 1806, an image feature difference extraction engine 1807, and an image feature registration management. 1808, each module of the data registration database 1809.
[0148]
When creating and registering a form (model) based on the read image, first, the image processing engine 1801 performs image processing such as inclination correction on the image, and then the image feature extraction engine 1802 performs the image processing. (For example, the size and position of the frame / table shape) are extracted and stored in the data registration DB 1809 by the image feature registration management 1808 as a registration form.
[0149]
The process of searching for a registration form that matches the input image among a plurality of registration forms is performed by first performing image processing such as tilt correction on the input document image in the image processing engine, A feature extraction engine 1802 extracts feature data (frame, table shape, etc.) of the document image. The extracted feature data and the feature data of the registration form registered in the data registration DB 1809 are compared by the image feature comparison engine 1803 to obtain a registration form that matches the input document image. The image feature difference extraction engine 1807 obtains the difference between the original position of the document image and the registration form, the scaling factor, and the like, and the image feature processing engine 1805 executes image processing so as to correct the difference and scaling factor of the document image. To do.
[0150]
Then, the document image that has been subjected to image processing in this way and the information of the identified registration form are sent to the OCR control unit, and the OCR control unit performs the image processing based on the area information set in the registration form. After the region image is extracted from the document image on which is executed, the extracted region image is subjected to the smoothing process, the character recognition process is performed, and the character recognition process result is stored in the worksheet as an index.
[0151]
<Example of image processing (FIG. 19)>
FIG. 19 shows an example when image data is processed by the image processing engine 1801.
[0152]
Reference numeral 1911 denotes image data when a document printed on paper is read from a scanner or the like, and the document has been read at an angle. By performing inclination correction by the image processing engine and correcting the origin position, a document image having no inclination such as 1901 can be obtained. Reference numerals 1902 and 1912 indicate positions on the document image corresponding to the area areas set in the registration form. Desired information cannot be obtained with an inclined image such as 1911. Desired information is obtained in the image subjected to the image processing. Reference numeral 1913 denotes the relative position of the area 1902 of the image 1901 after image processing, and it can be seen that the positions of 1912 and 1913 are greatly different.
[0153]
<Configuration of image feature comparison engine (FIG. 20)>
FIG. 20 shows a configuration diagram of an image feature comparison engine used when identifying a registration form.
[0154]
The image feature comparison engine 2000 recognizes the characters described in the document image and compares them with the characters described in the registration form. The image feature comparison engine 2000 detects the registration form, and the structure of the table frame of the document image. A form recognition engine 2003 that detects the registration form by comparing the size and the like with the table frame information of the registration form, and recognizes the barcode described in the document image and compares it with the barcode information of the registration form. Image features are compared by combining a barcode engine 2004 that detects a form, and the like, and a matching registered form is obtained. The image feature comparison engine management unit 2001 controls these engines.
[0155]
In this embodiment, an engine of OCR, a form, and a barcode is used. However, some combinations among them may be used, and other engines 2005 may be added.
[0156]
<Details of form registration process (FIG. 21)>
FIG. 21 shows the details of the process of registering an image as a registration form (S402) in which a form analysis process is performed to extract and register form information.
[0157]
From step S2100, processing for analyzing and registering an image as a form is started. In step S2101, image data is automatically or manually acquired from the device or a predetermined folder.
[0158]
In step S2102, the color depth of the captured image data is checked, and if it is a multi-valued image (color image), the process proceeds to step S2103, the binarization threshold is set to an optimum value based on the color depth distribution, and step S2104 is performed. In step S2105, the image data is subjected to color reduction processing to be converted into a monochrome binary image. On the other hand, if the captured image is not a multi-valued image but a binary image, the process advances to step S2105.
[0159]
In step S2105, it is checked whether the document image data has an inclination or a twist, and it is determined whether to correct these. In the case of an image input from most devices, such as an image read from a scanner, the image is often read at an angle, so the process advances to step S2106. On the other hand, if the input image data has no inclination, the process advances to step S2109. Whether or not to perform image processing in step S2105 may be determined in advance for each type of image acquisition destination device (scanner, FAX, folder, etc.), and the ruled line direction included in the acquired image. Alternatively, the orientation of the character image may be determined to determine whether inclination correction is necessary.
[0160]
In step S2106, it is determined whether there is a black portion on the outer edge of the captured document image data. When the background of the surface to be scanned by the scanner is black, when the document is tilted, or when the document size is smaller than the scan surface, as shown in 1911 of FIG. Is included. If there is a black part, the process advances to step S2108, and the boundary between the black part and the original is known. Therefore, the inclination of the original is obtained based on the boundary line, and the inclination of the document image is corrected based on the inclination angle. On the other hand, if there is no black portion, the process advances to step S2107 to determine the inclination angle based on the arrangement of character strings in the document image or the ruled line direction of the table frame so that the characters of the document image data are upright. Correct the tilt.
[0161]
In step S2109, a method for extracting features from the corrected document image data and an area for executing a predetermined process are designated. Here, feature data is extracted from the document image data. The image feature information extraction engines that can be specified here are an OCR engine 2002, a form recognition engine 2003, a barcode engine 2004, and other 2005 as shown in FIG.
[0162]
In step S2110, the engine setting specified in step S2109, the area, processing for the area (or the entire document) (such as smoothing processing), and an image obtained by extracting features from the document image data using the specified engine. Save the feature data as a registration form.
[0163]
<Details of Comparison Search Processing and Registration Processing Execution (FIG. 22)>
FIG. 22 is a flowchart showing details of the comparison search process (S811) for comparing the acquired document data with the registration form and searching for a matching registration form, and the execution of the set registration process (S822).
[0164]
In step S2201, document image data is acquired manually or automatically from a device or the like.
[0165]
In step S2202, the color depth of the acquired image data is checked, and in the case of a binary image, the process proceeds to step S2203. In the case of a multi-valued image (color image), the process proceeds to step S2202, and the binarization threshold value is optimized. In step S2203, the image data is subjected to color reduction processing and converted into a monochrome binary image.
[0166]
In step S2204, it is checked whether or not the image data document is tilted or twisted, and it is determined whether or not to correct these. In the case of an image input from most devices, such as an image read from a scanner, in many cases, the image is read at an angle, so the process advances to step S2205. On the other hand, if the input image data has no inclination, the process advances to step S2208.
[0167]
In step S2205, it is determined whether there is a black portion on the outer edge of the captured document image data. When the background of the surface to be scanned by the scanner is black, when the document is tilted, or when the document size is smaller than the scan surface, as shown in 1911 of FIG. Is included. If there is a black part, the process proceeds to step S2207, where the boundary between the black part and the original is known. The inclination of the original is obtained based on the boundary line, and the inclination of the document image is corrected based on the inclination angle. On the other hand, if there is no black portion, the process advances to step S2206 to determine the inclination angle based on the arrangement of the character strings in the document image or the ruled line direction of the table frame so that the characters of the document image data stand upright. Correct the tilt.
[0168]
In step S2208, image features are extracted by the same method as the registered form feature extraction method (the method specified in step S2109).
[0169]
In step S2209, the image feature extracted in step S2208 is compared with the image feature information of the registration form, and as a result, a registration form that matches (or resembles) the image feature information is identified.
[0170]
In step S 2210, the document image is corrected for a shift between the origin of the document image and the origin of the identified registered form, magnification change, and the like.
[0171]
In step S2211, the area image of the area is extracted from the corrected image, and the process registered in the area is executed. For example, if processing is registered to perform OCR on the area and use it as an index, the area image of the area is transferred to the OCR control unit, and smoothing processing is performed to perform character recognition processing (barcode recognition processing may be used. ) And control to use the recognition result as an index.
[0172]
<Detailed processing of comparison search / image correction (FIG. 23)>
FIG. 23 is a diagram showing details of the comparison search / image correction processing in steps S2209 to S2210.
[0173]
In step S2301, image feature information on the registration form is compared with information obtained by extracting image features (S2208) from a document image sent from a device or the like.
[0174]
In step S2302, if it is determined in step S2301 that the image features have the same image characteristics, the process advances to step S2303, and the degree of coincidence between the registered form determined to be the same and the document image is calculated from each image feature. The registration form ID and the degree of coincidence are temporarily stored in the database. On the other hand, if it is not determined that they are the same, the process proceeds to step S2304. (In this embodiment (FIG. 23), the configuration of the frame table ruled line is used as each feature, and if the configuration of the frame table ruled line has a similar configuration, it is determined that they match.)
In step S2304, it is checked in advance whether or not all registered forms have been compared and searched. If all the comparisons have been completed, the process proceeds to step S2305. If all the comparisons have not been completed, the process returns to step S2302. Compare with the following registration form.
[0175]
In step S2305, based on the result stored in the database in step S2303, data relating to the registered form having the highest degree of matching is extracted.
[0176]
In step S2306, the origin of the document image and the origin of the registration form are obtained.
[0177]
In step S2307, the origins obtained in step S2306 are compared, and if it is determined that the origins are the same in both the X and Y coordinates, the process proceeds to step S2310, and if it is determined that they are different, the process proceeds to step S2308.
[0178]
In step S2308, a difference is obtained from each origin position.
[0179]
In step S2309, the deviation of the document image is corrected based on the difference.
[0180]
In step S2310, the size of each image is compared based on the image feature of the registration form extracted in step S2305 and the feature of the document image, and the ratio of enlargement / reduction is determined.
[0181]
In step S2311, it is determined whether there is enlargement / reduction based on the result obtained in step S2310. If it is determined that there is, the process proceeds to step S2312, and if not, the process proceeds to step S2313.
[0182]
In step S2312, the information corrected in step S2309 is further corrected according to the ratio obtained in step S2311.
[0183]
【The invention's effect】
As described above, according to the present embodiment, the document image is corrected based on the identified registration form, the area image is extracted and recognized, and the recognition result is used as an index. , Recognition accuracy is high. Further, since the recognition accuracy is increased, index input can be performed efficiently.
[0184]
Also, the recognition accuracy can be further improved by smoothing the corrected image.
[Brief description of the drawings]
FIG. 1 is a system configuration diagram of an embodiment of the present invention.
FIG. 2 is a block diagram showing the structure of the document registration system 120.
Fig. 3 Basic flow of document registration system
FIG. 4 Template creation processing
[Figure 5] Worksheet creation processing
[Figure 6] Form registration process
FIG. 7: Area registration process
[Figure 8] Input activation button setting
FIG. 9: Input activation button execution processing
FIG. 10: Index input processing
FIG. 11: Release (data registration) processing
FIG. 12 is a job creation process.
FIG. 13 is a job execution process.
FIG. 14 is an extended device controller.
FIG. 15 shows an extended document management system control unit.
FIG. 16: Supplement of release processing
FIG. 17 is an automatic document registration process for a document transmitted from a device.
FIG. 18 shows the configuration of the form analysis unit.
FIG. 19 is an image processing example.
FIG. 20 is a block diagram of an image feature comparison engine.
FIG. 21: Details of form registration processing
FIG. 22 shows details of comparison search processing and registration processing execution.
FIG. 23 shows detailed processing of comparison search and image correction.
[Explanation of symbols]
100 to 102 user terminals
110 MFP
120 Document registration system
130 File Server
140 Document Management System

Claims (10)

A document processing system that registers document image data and an index in a document management unit that manages document image data,
A template creation unit for obtaining an index structure of the database of the document management unit, creating a template based on the obtained index structure and connection information to the database of the document management unit, and storing the created template;
Based on the template, a worksheet creation means for creating a worksheet having items for use in index registration;
Storage means for storing a plurality of registration forms;
An area registration means for setting an area on the registration form and associating the set area with an item of the worksheet;
Feature extraction means for extracting feature information from the input document image data;
A form identifying means for comparing the extracted feature information of the document image with the feature information of each of the stored plurality of registered forms and identifying a registered form that is determined to match the comparison result;
Difference extraction means for extracting a difference between the registration form and the document image based on the feature information of the identified registration form and the feature information of the document image;
Image correcting means for correcting the document image based on the extracted difference;
Area extracting means for extracting a corresponding area of the corrected document image based on the area set in the identified registration form;
Recognizing means for recognizing the image of the extracted area;
Worksheet registration means for registering the recognition result by the recognition means in the worksheet as an index of items corresponding to the area ;
Determining means for determining whether or not to continue inputting other document image data;
When it is determined that the input is continued by the determination unit, the feature extraction unit, the form identification unit, the difference extraction unit, the image correction unit, the area extraction unit, and the recognition are performed on the other document image data. While executing the processing by the means, the worksheet registration means and the judgment means,
A document processing system comprising: a registration unit that registers an index registered in the worksheet together with the document image data in the document management unit when the determination unit determines that the input is not continued . Furthermore, it has smoothing means for smoothing the image of the area extracted by the area extraction means,
The document processing system according to claim 1, wherein the recognition unit performs a recognition process on the smoothed area image. The document processing system according to claim 1, wherein the difference extracted by the difference extraction unit is at least one of a deviation of an origin position and a scaling factor . It said recognition means, the document processing system according to any of claims 1 to 3, the image of the extracted area, characterized by performing at least one of the character recognition processing and a bar code recognition processing. Furthermore, it has an inclination correction means for correcting the inclination of the input document image data,
The feature extraction unit, a document processing system according to any of claims 1 to 4, characterized in that said from the slope corrected document image data, extracts the feature information. The tilt correction means detects a black portion of an edge of the input document image, obtains a tilt angle based on a boundary line between the black portion and the document portion, and performs tilt correction according to the tilt angle. The document processing system according to claim 5 . The document processing system according to claim 5 , wherein the tilt correction unit performs tilt correction based on a direction in which a character image in the input document image stands upright. A document processing method for registering document image data and an index in a document management unit that manages document image data,
A template creation unit obtains an index structure of the database of the document management unit, creates a template based on the obtained index structure and connection information to the database of the document management unit, and stores the created template A template creation step;
A worksheet creating step for creating a worksheet having items to be used for index registration based on the template;
An area registration unit that sets an area on a registration form stored in the storage unit and associates the set area with an item of the worksheet;
A feature extraction step in which feature extraction means extracts feature information from the input document image data;
Registration form is form identification means, wherein the feature information of the extracted the document image is compared with the plurality of registration form each feature information stored in the storage means, it is determined that the comparison result matches A form identification step to identify
Difference extracting means, wherein the feature information of the identified registration form, on the basis of the characteristic information of the document image, the difference extracting a difference between the registration form and the document image,
Image correction means, based on the extracted difference, the image correcting step of correcting the document image,
Area extraction means, on the basis of the area set in the identified registration form, the area extracting step of extracting the corresponding area of the corrected document image,
Recognizing means, a recognition step of recognizing processing the image of the extracted area,
Worksheet registration means for registering the recognition result in the recognition step as an index of an item corresponding to the area in the worksheet,
A determining step for determining whether or not to continue inputting other document image data;
If it is determined that the input is continued in the determination step, the feature extraction step, the form identification step, the difference extraction step, the image correction step, the area extraction step, and the recognition are performed on the other document image data. While repeating the step, the worksheet registration step, and the determination step, if the determination unit determines that the input is not continued, the registration unit displays the index registered in the worksheet together with the document image data. A document processing method comprising: a registration step of registering in a document management unit. A computer of a document processing system for registering the document image data and an index in a document management unit that manages document image data,
A template creation unit for obtaining an index structure of the database of the document management unit, creating a template based on the obtained index structure and connection information to the database of the document management unit, and storing the created template;
Worksheet creation means for creating a worksheet having items for use in index registration based on the template,
Storage means for storing multiple registration forms;
Area registration means for setting an area on the registration form and associating the set area with the item of the worksheet;
Feature extraction means for extracting feature information from the input document image data;
Form identification means for comparing the extracted feature information of the document image with the feature information of each of the plurality of stored registration forms, and identifying a registration form that is determined to match the comparison result;
Difference extraction means for extracting a difference between the registration form and the document image based on the feature information of the identified registration form and the feature information of the document image;
Image correcting means for correcting the document image based on the extracted difference;
Area extracting means for extracting a corresponding area of the corrected document image based on the area set in the identified registration form;
Recognition means for recognizing the image of the extracted area;
Worksheet registration means for registering the recognition result by the recognition means in the worksheet as an index of items corresponding to the area,
Determining means for determining whether or not to continue to input other document image data;
When it is determined that the input is continued by the determination unit, the feature extraction unit, the form identification unit, the difference extraction unit, the image correction unit, the area extraction unit, and the recognition are performed on the other document image data. While executing the processing by the means, the worksheet registration means and the judgment means,
A registration unit that registers the index registered in the worksheet together with the document image data in the document management unit when the determination unit determines that the input is not continued;
A computer-executable program that functions as a computer. A computer-readable storage medium storing the program according to claim 9 .

Images