JP7095259B2 - Document processing equipment and programs - Google Patents

Description

The present invention relates to a document processing apparatus and a program.

There are document management devices and software that manage image data read by a scanner and document data created by an application on a personal computer (PC). For example, document handling software such as DocuWorks ™ provided by the applicant and Adobe Acrobat ™ of Adobe Systems, Inc. is an example. This type of document management device is an image data file obtained by scanning a paper document (for example, bitmap format, TIFF format, JPEG format), or an application file created by various applications such as a word processor or a spreadsheet. Can be captured and managed.

This type of document management device handles files with different data formats created by various applications as data described in a language that defines the appearance of the page, such as a page description language. In addition, this type of document management device can also handle files in the image data format generated by a scanner, a digital camera, or the like.

Further, in this type of document management device, as exemplified in Patent Document 1, a plurality of document files are bundled to form one document file, or one document file is combined into a plurality of document files in page units. Some have the function of disassembling ("disassembling"). A document file in which an application document and an image document are bundled into one is a file in which a page described in a language such as a page description language and a page in an image data format are mixed.

The device disclosed in Patent Document 2 is an image processing device having a function of displaying a thumbnail image in which elements of document data are thumbnailed, and is a storage device for storing document data and a storage device stored in the storage device. It has a thumbnail control object addition operation function that adds a thumbnail control object for controlling the thumbnail image of the document data to this document data and saves the added thumbnail control object in the storage device.

Japanese Unexamined Patent Publication No. 10-12489 Japanese Unexamined Patent Publication No. 2008-42359

Character recognition is a process that targets image data, and character recognition cannot be performed for data that is not in the image data format. In the process of character recognition for a document in which pages in image data format and pages in non-image data format are mixed, character recognition is performed only for pages in image data format, and character recognition is also performed for pages not in image data format. If no processing is done, the user does not even know that the document contained pages for which character recognition could not be performed.

It is an object of the present invention to provide more information about pages that are not in the image data format in the character recognition process, as compared to the case where no processing is performed on the pages that are not in the image data format in the document.

The invention according to claim 1 is when a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to perform character recognition. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. In the document processing device , the type 2 page processing is a notification process for notifying the user that the document includes a page to which the character recognition cannot be applied when the document contains the type 2 page. There is a document processing device .

In the invention according to claim 2 , in the notification processing, an inquiry as to whether or not to acquire character information from the second type page is further made to the user, and the document processing apparatus is a user's inquiry to the inquiry. The document processing apparatus according to claim 1 , which controls whether or not to execute a process of acquiring character information from the second type page according to an answer.

The invention according to claim 3 is the document processing apparatus according to claim 1, wherein the type 2 page processing is a processing for acquiring character information from the type 2 page.

The invention according to claim 4 is the document processing apparatus according to claim 3 , wherein the type 2 page processing is a processing for extracting text data included in the type 2 page.

The invention according to claim 5 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. The type 2 page processing is a document processing device , which is a process of converting the type 2 page into image data and executing the character recognition on the image data.

The invention according to claim 6 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. A document processing device, wherein the document processing device performs a process of extracting text data included in the second-class page, converts the second-class page into image data, and displays the characters on the image data. It is a document processing device that executes a process selected by the user among the processes for executing recognition as the second type page process.

The invention according to claim 7 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. In the document processing device, the type 2 page processing is based on the text data extracted from the type 2 page and the result of character recognition for the image data obtained by converting the type 2 page. This is a document processing device that is a process for obtaining character information that is a processing result for a type 2 page.

According to the eighth aspect of the present invention, in the second-class page processing, for a certain place in the second-class page, the first character corresponding to the place in the text data extracted from the second-class page and the first character. When both the second character corresponding to the place in the result of the character recognition for the image data obtained by converting the second type page exist, the first character is adopted as the character information about the place. The document processing apparatus according to claim 7 .

According to the invention of claim 9 , in the type 2 page processing, there is no character corresponding to the place in the text data extracted from the type 2 page for a certain place in the type 2 page, and the above-mentioned invention. A claim that, when there is a character corresponding to the place in the result of the character recognition for the image data converted from the second type page, the character in the result of the character recognition is adopted as character information about the place. The document processing apparatus according to 7 or 8 .

According to the invention of claim 10 , in the type 2 page processing, there is a character corresponding to a certain place in the type 2 page in the text data extracted from the type 2 page. Whether or not to adopt the character in the text data as character information about the place when there is no character corresponding to the place in the result of the character recognition for the image data converted from the second type page. The document processing apparatus according to any one of claims 7 to 9 , which is controlled according to a user's instruction.

The invention according to claim 11 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. In the document processing device, when the document includes the type 2 page, the notification process for notifying the user that the page to which the character recognition cannot be applied is included, and the character from the type 2 page. It is a document processing apparatus having a process for acquiring information and a means for presenting a setting screen for receiving a setting from a user as to which of the two is executed as the second type page process.

The invention according to claim 12 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , The character recognition is executed for the first type page included in the document, and the second type page processing, which is a process different from the character recognition, is executed for the second type page included in the document. In the document processing device, when the document includes the second type page, the notification process for notifying the user that the page to which the character recognition cannot be applied is included, and the second type page include the notification process. Two or more of the process of extracting the text data to be generated and the process of converting the second-class page into image data and executing the character recognition on the image data are selected as the second-class processing options. It is a document processing device having a means for presenting a setting screen for letting a user select.

The invention according to claim 13 is a case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , A means for executing the character recognition for the type 1 page included in the document, and executing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. This is a program for operating a computer as a user , and the user indicates that the second-class page processing includes a page to which the character recognition cannot be applied when the second-class page is included in the document. It is a program that is a notification process to notify to .
The invention according to claim 14 is in the case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , A means for executing the character recognition for the first type page included in the document, and executing the second type page processing which is a process different from the character recognition for the second type page included in the document. The type 2 page processing is a program for converting the type 2 page into image data and executing the character recognition for the image data. Is.
The invention according to claim 15 is a case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , A means for executing the character recognition for the first type page included in the document, and executing the second type page processing which is a process different from the character recognition for the second type page included in the document. The second-class page processing is a program for operating the computer as, and the second-class page processing includes the text data extracted from the second-class page and the result of the character recognition for the image data obtained by converting the second-class page. , Is a program that obtains character information that is a processing result for the second type page.
The invention according to claim 16 is a case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , A means for executing the character recognition for the type 1 page included in the document, and executing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. , Is a program for operating the computer as, and notifies the user that the computer includes a page to which the character recognition cannot be applied when the document contains the second type page. To function as a means of presenting a setting screen that accepts from the user the setting of which of the processing and the processing of acquiring character information from the second type page is executed as the second type page processing. It is a featured program.
The invention according to claim 17 is a case where the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed. , A means for executing the character recognition for the first type page included in the document, and executing the second type page processing which is a process different from the character recognition for the second type page included in the document. , A program for operating the computer as, notifying the user that the computer includes a page to which the character recognition cannot be applied when the document contains the second type page. Two of a process, a process of extracting the text data included in the type 2 page, and a process of converting the type 2 page into image data and executing the character recognition on the image data. The program is characterized in that the above functions as a means for presenting a setting screen for allowing the user to select the above as an option for the second type of processing.

In the configuration 1 of the reference example, when the first document including the page not in the image data format is bundled with one or more other documents into one document, the page not in the image data format is imaged. It is a document processing apparatus having a bundling means for converting into a data format and preventing the document as a result of the bundling process from including pages other than the image data format.

Configuration 2 of the reference example is a document as a result of the bundling process when the first document including a page not in the image data format is bundled with one or more other documents into one document. It further has a second bundling means for leaving pages that are not in the image data format, and a means for accepting from the user whether to use the bundling means or the second bundling means when the bundling process is instructed. The document processing apparatus according to the configuration 1 of the reference example.

In the configuration 3 of the reference example, when the first document including the page not in the image data format is bundled with one or more other documents into one document, the page not in the image data format is imaged. It is a program for operating a computer as a bundling means for converting into a data format and preventing the document resulting from the bundling process from including pages other than the image data format.

According to the invention according to claim 1 or 13 , the information obtained by the type 2 page processing is obtained as compared with the case where the character recognition is simply not executed for the type 2 page that is not in the image data format in the document. It can be provided to the user.

Further , the user can know that the document targeted for character recognition contains pages to which character recognition cannot be applied.

According to the second aspect of the present invention, the user can instruct to acquire character information by another method from a page to which character recognition cannot be applied.

According to the third aspect of the present invention, it is possible to provide the user with character information acquired from the second type page by a method different from character recognition.

According to the invention of claim 4 , the text data included in the type 2 page, which is character information having higher accuracy than the character information obtained by character recognition, is provided to the user as the character information of the type 2 page. Can be done.

According to the invention of claim 5 , even when the image that can be seen as characters by the human eye when the type 2 page is displayed is expressed in a format other than the text data format in the data of the type 2 page. , You can get the information of the character.

According to the invention of claim 6 or 14 , it is possible to extract text data from the type 2 page, convert the type 2 page into image data for character recognition, or use a method desired by the user. can.

According to the invention of claim 7 or 15 , the text data extracted from the type 2 page or the result of character recognition for the image data converted from the type 2 page is adopted as compared with the case of adopting either one. More accurate information can be provided.

According to the invention of claim 8 , when the text data for the same place in the type 2 page and the character recognition result are different, the accurate character information of the text data can be adopted.

According to the invention of claim 9 , character information can be obtained about a place where a character is recognized by character recognition although there is no character in the text data in the type 2 page.

According to the invention of claim 10 , the user's intention is to decide whether or not to adopt the character information of the text data in the place where the characters of the text data are present but the characters are not recognized by the character recognition in the type 2 page. Can be reflected.

According to the invention according to claim 11, 12, 16 or 17 , the user can specify which process is to be performed as the second type page process.

According to the invention according to the configuration 1 or 3 of the reference example, a document to which character recognition can be applied to all pages can be obtained as a document as a result of the bundling process.

According to the configuration 2 of the reference example, as the document as a result of the bundling process, a document to which character recognition can be applied to all pages is obtained, or a document in which pages other than the image data format included in the document before bundling are left is left. It can be decided according to the user's wishes.

It is a figure which illustrates the schematic structure of the document handling system to which the character recognition processing control of an embodiment is applied. It is a figure which shows the example of the document list screen provided by a document handling system. It is a figure which shows the example of the setting screen about a certain setting item of a document handling system. It is a figure which shows the example of the control procedure which a processing control unit executes when the user instructed to execute character recognition (OCR). It is a figure which shows the example of the screen which the processing control part displays after performing character recognition in the document which contains an application page. It is a figure which shows another example of the control procedure which a process control part executes when the user instructed to execute character recognition (OCR). It is a figure which shows the example of the confirmation screen which the processing control part displays when performing character recognition in the document which contains an application page. It is a figure which shows another example of the confirmation screen which the processing control part displays when performing character recognition in the document which contains an application page. It is a figure which shows an example of the automatic selection process of the text acquisition method from the application page by a process control unit. It is a figure which shows another example of the automatic selection process of the text acquisition method from the application page by a process control unit. It is a figure which shows still another example of the automatic selection process of the text acquisition method from the application page by a process control unit. It is a figure which shows still another example of the automatic selection process of the text acquisition method from the application page by a process control unit. It is a figure which shows another example of the setting screen about one setting item of a document handling system.

FIG. 1 illustrates a schematic configuration of a document handling system 10 to which the character recognition processing control according to the present invention is applied. The document handling system 10 allows viewing, editing, annotating (colored markers, sticky notes, stamps, etc.) of registered document files, such as DocuWorks ™ provided by the applicant and Adobe Acrobat ™ of Adobe Systems. It is software that provides processing functions such as (addition). This software realizes the function of the document handling system 10 by being executed by a computer such as a PC (personal computer) or a server. The document handling system 10 may be a personal system installed on a PC, or may be constructed as a server that provides a document handling service to a remote user.

The document handling system 10 includes a file import unit 12, a document processing unit 14, a UI processing unit 16, a setting management unit 18, and a character recognition processing unit 20 as functional modules.

The file import unit 12 is responsible for importing (that is, registering) a file generated outside the document handling system 10 into the document handling system 10. The files to be imported include files in the application-specific data format (called application files) generated by application software such as word processors and table calculations, bitmaps generated by scanners and digital cameras, TIFF, JPEG, etc. There are image files in the image data format of.

When an application file is input as an import target, the file import unit 12 converts it into a document file described in a specific data description language used by the document handling system 10. The data description language used is PDL (page description language) or another kind of language such as a page description language that can describe the appearance of a document (image) (PDF data format is also an example). Document files described in this data description language include multiple types such as text, vector graphics (graphics described in vector representation), and continuous gradation images (data formats such as bitmap, TIFF, and JPEG). Objects can be included. A document file described in this data description language is hereinafter referred to as an "application document". In addition, the individual pages that make up the application document are called "application pages". The app page can contain several objects such as text, vector graphics, and continuous tone images. The app page may include objects in the image data format, but the data format of the app page itself is defined by its data description language, not the image data format.

Further, when an image file is input as an import target, the file import unit 12 imports the file as an "image document". An image document is a document file in which each page included in the document is data in an image data format such as bitmap, TIFF, or JPEG. Each page that constitutes an image document is called an "image page".

In this way, the file capture unit 12 plays a role similar to that of a printer driver or a virtual printer.

The document file imported by the file import unit 12 is stored in one of the folders under the control of the document handling system 10 (for example, the one specified or preset by the user).

The document processing unit 14 executes a process instructed by the user for a document file stored in a folder under the control of the document handling system 10. For example, when the user is instructed to browse the document file, the document processing unit 14 opens the document file and displays an image of the page of the document file on the screen. Further, the document processing unit 14 adds or deletes annotations to the opened document file in response to an instruction from the user. In addition, the document processing unit 14 has the functions of "disassembling" and "bunching" documents. The disassembly process is a process of dividing one document file into a first document file consisting of pages before the specified page and a second document file consisting of pages after the next page of the specified page. be. Further, the bundling process is a process of bundling a plurality of document files specified by the user into one document file. The document file after bundling is a file having both the page group of the first document file and the page group of the second document file before bundling. When the application document and the image document are bundled, in one example, a document file in which the application page and the image page are mixed is generated. The document file has, for each included page, type information indicating whether the page is an application page or an image page as one of the attribute information of the page.

The UI processing unit 16 provides a UI (user interface) screen for operating a document file, and accepts a user's operation on the screen. The UI processing unit 16 generates, for example, the document management screen 100 illustrated in FIG. 2 and displays it on the display device of the computer on which the document handling system 10 is installed. On the illustrated document management screen 100, icons 102, 104, and 106 of each document file in the open folder are listed and displayed. In this example, the icons 102, 104, and 106 of each document file are thumbnail images of the first page of the document file, respectively. For example, the user can instruct the bundling process of these two document files by superimposing the icon of one document file on the icon of another document file by a drag and drop operation. It is also possible to select the process to be performed on the document file from the menu on the document management screen 100 (not shown) or the context menu called by an operation such as right-clicking the mouse on the icon.

The setting management unit 18 accepts input of various setting items of the document handling system 10 and holds the values of the input setting items. Examples of setting items managed by the setting management unit 18 include processing of application pages when bundling documents, processing of application pages during OCR (optical character recognition, hereinafter also simply referred to as character recognition) processing, and the like.

FIG. 3 illustrates a setting screen 200 displayed by the UI processing unit 16 in order to accept the settings of these items. On this setting screen 200, there are two options, the first option "bundle as it is" and the second option "convert the application page into an image page and bundle" in the option column 202 for processing the application page when bundling documents. Is shown. The first option is the process of bundling the application pages included in the document file to be bundled as they are. The application page remains in the bundled document file obtained as a result of this process. On the other hand, the second option is a process of converting the application pages included in the document files to be bundled into image pages and then bundling them. The bundled document file obtained as a result of this process consists of image pages and does not include application pages.

When the second option of the option column 202 is selected, when the user performs an operation of bundling the document files, the document processing unit 14 displays all the application pages included in those document files as image pages by the rasterizer 28 or the like. Convert to. Since the document file of the bundled result obtained in this case consists of an image page and does not include an application page, character recognition processing can be applied to all pages.

The circular figure at the beginning (left end) of each option in the option column 202 is a radio button. The user selects a desired radio button by clicking or touching the mouse.

Further, on the setting screen 200, the first option "display a message that OCR could not be performed" and the second option "convert to an image page and perform OCR" are displayed in the option column 204 for processing the application page at the time of OCR. "Execute" and the third option "Extract the text in the application page and use it as the OCR result" are shown. Here, OCR (character recognition) is a process of recognizing characters included in an image (image data) by pattern matching or the like, and cannot be directly applied to data other than image data. Therefore, OCR can be executed on the image page in the document file, but OCR cannot be executed on the application page. The first option in the option column 204 is a process of executing OCR processing only on the image page in the document file to be OCR and displaying a message on the screen indicating that a page to which OCR cannot be applied is included after the end of OCR. Is. The second option is a process of converting the application page in the document file into an image page and then executing OCR for all the pages. The third option is a process of extracting the text data included in the text object in the application page and outputting the extracted text data as a character recognition result.

In the illustrated example, the choice field 204 contained three choices, but the choice field 204 does not have to contain all three choices and includes choices other than those three. You may.

Returning to the description of FIG. 1, the character recognition processing unit 20 is a functional module responsible for OCR (character recognition) processing. The character recognition processing unit 20 includes a processing control unit 22, a character recognition engine 24, a text extraction unit 26, and a rasterizer 28. The processing control unit 22 controls the character recognition engine 24, the text extraction unit 26, and the rasterizer 28 to perform control for realizing the functions of the character recognition processing unit 20. The character recognition engine 24 executes character recognition processing based on a known OCR algorithm on the image data. The text extraction unit 26 extracts text data from the application page. The rasterizer 28 converts the application page described in the data description language into raster data (bitmap image).

FIG. 4 shows an example of a control procedure executed by the processing control unit 22 when the user is instructed to execute character recognition (OCR). In this procedure, the processing control unit 22 first examines the attributes of each page in the document file designated as the target of character recognition, and acquires the type information of each page (S10). This makes it possible to know whether each page in the document file is an application page or an image page. The processing control unit 22 determines whether or not the document file includes one or more application pages based on the information acquired in S10 (S12). When the result of this determination is No, that is, all the pages in the document file are image pages, a known character recognition process is executed for all the pages (S22). The character recognition result data of each page is saved in association with the corresponding page.

When it is determined in S12 that the document file contains one or more application pages, the processing control unit 22 sets processing setting items (see reference numeral 204 in FIG. 3) for the application page during character recognition (OCR) processing. Obtained from 18. Then, it is determined whether the setting item is the above-mentioned first option (message display), second option (OCR after image conversion), or third option (text extraction) (S14).

When the value of the setting item indicates the first option (menu display), the processing control unit 22 causes the character recognition engine 24 to process each image page in the document file, and each of them obtained by this is processed. The data of the character recognition result of the page is saved (S16). If the document file is composed only of application pages, no characters are recognized in S16. When the character recognition process of S16 is completed, the process control unit 22 displays that the document file designated as the target this time contains a page to which character recognition cannot be applied (S18).

FIG. 5 shows an example of the screen 300 displayed at this time. The user who has instructed character recognition understands from this screen that some pages in the document file instructed this time do not have the character recognition result. Further, further information about the page group to which character recognition could not be applied may be displayed on this screen 300. Such further information includes, for example, a list of page numbers to which character recognition could not be applied, the percentage of pages to which character recognition could not be applied to the total number of pages in the document, and the like. Further, when the document file contains only the application page and does not include the image page, a screen indicating that the page capable of OCR is not included may be displayed instead of the screen illustrated in FIG. ..

Further, on the screen 300, the user may be inquired whether or not to try to acquire the text information from the page (application page) to which the character recognition cannot be applied. There are two methods for acquiring text from the application page, such as converting to an image and recognizing characters, and acquiring the text data of the text object in the application page (see another example below for details). As a measure for the page to which character recognition cannot be applied on the screen 300, among the options of doing nothing, converting to an image and recognizing the character, and acquiring the text data of the text object in the application page. You may let the user select the process from. Then, when the user selects one of the text acquisition methods, the processing control unit 22 acquires the text information from each application page by that method.

When it is found in S14 that the value of the setting item is the second option (OCR after image conversion), the processing control unit 22 causes the rasterizer 28 to process each application page in the document file, so that each application page is processed. Is converted into an image page (S20). As a result, all the target document files become image pages. The processing control unit 22 causes the character recognition engine 24 to perform character recognition processing on all of these image pages (S22).

The flow of converting all the application pages in the document file to image pages and then recognizing all the pages in the document file is just an example. Instead of this, while processing the document file page by page in page order, if it is an application page, it may be converted into an image page and then character recognition is executed, and if it is an image page, character recognition may be simply performed.

In addition, each application page in the document file may be replaced with the image page obtained in S20. In this case, all the pages of the document file stored in the document handling system 10 after the character recognition process are image pages. It should be noted that the user may be inquired at the time when it is determined in S12 that the application page is included, and whether or not such replacement is performed during the character recognition process may be inquired in advance as one of the setting items of the setting management unit 18. You may be able to do it.

When it is found in S14 that the value of the setting item is the third option (text extraction), the processing control unit 22 causes the text extraction unit 26 to extract the text data in each application page in the document file (S24). .. Then, the text data extracted from the application page is saved as the character recognition result corresponding to the application page.

In the above explanation, the case where a certain document file is selected as the target of character recognition is taken as an example, but this is only an example. Instead, when some pages in the document file are selected for character recognition, or in a specific area within a page of all or selected parts of the document file (eg, such). The same processing may be performed even when only (the user specifies the positions of the four corners of the rectangle on the outer circumference of the area) is selected as the target of character recognition. In this way, the point that the target of character recognition may be the entire document file or a part thereof (that is, a part of a page group or a part of a page) is described later. The same applies to variations.

In the above example, the handling of the application page is determined according to the preset settings. As another example, after starting the character recognition process, when it is found that the target document file has the application page, the user may be inquired about the handling of the application page. The procedure according to this example is shown in FIG. In the procedure of FIG. 6, it is not necessary to register the settings for the processing of the application page in character recognition in the setting management unit 18 in advance.

In the procedure of FIG. 6, first, the processing control unit 22 examines the attributes of each page in the document file designated as the target of character recognition, and acquires the type information of each page (S30). The processing control unit 22 determines whether or not all the pages in the document file are application pages based on the information acquired in S10 (S32). When the determination result of S32 is Yes, there is no image page capable of character recognition in the document file, so that the processing control unit 22 includes a page capable of character recognition in the document file specified by the user. A screen indicating that there is no such screen is displayed via the UI processing unit 16 (S34).

If the determination result in S32 is No, the processing control unit 22 determines whether or not the application page is included in the document file (S36). If the result of this determination is No, all the pages in the document file are image pages. In this case, the processing control unit 22 performs character recognition processing by the character recognition engine 24 on all pages of the document file (S38).

When the determination result of S36 is Yes, the processing control unit 22 displays the confirmation screen 320 illustrated in FIG. 7 via the UI processing unit 16 (S40). On the confirmation screen 320, a message indicating that the target of character recognition specified by the user includes a page that cannot recognize characters, an inquiry indicating whether or not the process of recognizing characters may be executed only on the page that can recognize characters, And two buttons ("OK" or "Cancel") for entering the pros and cons of the inquiry are displayed. The user selects the "OK" button when the process of recognizing characters only on the page where the characters can be recognized is sufficient, and selects the "Cancel" button otherwise. The processing control unit 22 determines which is selected by the user (S42), and when "OK" is selected, causes the character recognition engine 24 to process only the image page in the document file (S44). If "Cancel" is selected, S44 is skipped and the process ends. In this case, no character recognition processing is performed on the document file.

In the procedure of FIG. 6 described above, either the character recognition is not performed at all for the document file including the application page, or the character recognition is performed only for the image page in the document file. Such a procedure is just an example. Instead of this, similar to the procedure of FIG. 4, the application page may be subjected to character recognition or a process for obtaining an equivalent result.

For example, when it is determined from the information acquired in S30 that the application page to be processed for character recognition is included, the processing control unit 22 may display the confirmation screen 340 exemplified in FIG. The confirmation screen 340 shows a message that the target of character recognition specified by the user includes a page whose character recognition is not possible, a message requesting selection of processing for the page whose character recognition is not possible, and options for the processing. The option column 342, the "OK" button and the "Cancel" button are displayed. The options shown in the option column 342 are similar to the three options shown in the option column 204 of the setting screen 200 illustrated in FIG. That is, the first option is a process in which character recognition is not performed on a page (application page) where character recognition (OCR) is not possible, and character recognition is performed only on an image page. The second option is a process of converting an unrecognizable page into an image page and performing character recognition, and the third option is to extract the text data contained in the unrecognizable page and extract the text. This is a process that uses data as a character recognition result. The user selects one process desired by himself / herself from the option field 342 of the confirmation screen 340, and presses the "OK" button to instruct the process control unit 22 to execute the process. For example, when the first option is selected, the processing control unit 22 causes the character recognition engine 24 to process only the image page in the document file. When the second option is selected, the processing control unit 22 first causes the rasterizer 28 to process the application page in the document file, and causes the character recognition engine 24 to process the image data obtained as a result. When the third option is selected, the processing control unit 22 causes the text extraction unit 26 to extract text data from each application page, and uses the extracted text data as the character recognition result of the page.

Next, further variations will be described. In the example described above, the user has selected whether to acquire text information from the application page by creating an image and recognizing characters or extracting text from a text object in the application page. .. On the other hand, in the example described below, the processing control unit 22 automatically makes this selection.

FIG. 9 shows an example of the text acquisition method selection process by the process control unit 22. The processing control unit 22 performs the processing of FIG. 9 for each application page in the designated character recognition target. That is, the processing control unit 22 extracts text data from the text object in the application page using the text extraction unit 26 (S50), and the number of characters included in the extracted text data is equal to or higher than a preset threshold value. It is determined whether or not there is (S52). When the number of characters in the extracted text is equal to or greater than the threshold value, the processing control unit 22 saves the extracted text as a character recognition result corresponding to the application page (S54). On the other hand, when the number of characters in the extracted text is less than the threshold value, the processing control unit 22 converts the application page into image data using the rasterizer 28, and the character recognition engine 24 transfers the image data to the image data. Character recognition is performed (S56).

The process of FIG. 9 deals with the fact that the part that looks like a character when the application page is displayed (drawn) is not always expressed as text in the data of the application page. That is, the application applies the application file created by the application to the file importing unit 12 (or a virtual printer having an equivalent function thereof) of the document handling system 10 with a plurality of vector graphics of the characters in the text in the file. It may be disassembled into files and input. This is because, for example, the reproducibility of the appearance of the character is ensured. When a character is decomposed into vector graphics and input to the file import unit 12 in this way, the character is expressed as vector graphics even in the application page of the application document generated by the file import unit 12. Become. Therefore, the character does not exist as text data on the application page, but when the application page is displayed (drawn), an image that looks like the character appears. If many of the "characters" (image objects that look like characters to the eye) in an app page are represented by vector graphics, extracting text from that app page will only yield text with a very small number of characters. I can't.

Therefore, in the procedure of FIG. 9, when the number of characters extracted from the application page is too small (that is, less than the threshold value), the application page is converted into an image and then the character recognition process is performed. If the app page contains more "characters" than the number of characters that can be extracted as text (many of which are expressed in vector graphics), it can be imaged and recognized as characters. It becomes possible to detect many of those "characters".

After S56, even if the number of characters m recognized by the character recognition process and the number of characters n extracted in S50 are compared and it is confirmed that the number of characters sufficiently larger than the number of characters in the extracted text is obtained by character recognition. good. That is, in this example, when the number of characters m is sufficiently larger than the number of characters n (that is, the difference is equal to or more than a certain threshold value), the result of character recognition is adopted as the final character recognition result, and if not, it is extracted. Adopt the text as the final character recognition result.

FIG. 10 shows a processing procedure of another example of automatic selection of the text acquisition method from the application page.

In the procedure of FIG. 10, the processing control unit 22 uses the text extraction unit 26 to extract text data from the text object in the application page (S60). Further, the processing control unit 22 converts the application page into image data using the rasterizer 28, and character recognition is performed on the image data by the character recognition engine 24 (S62) (the processing order of S60 and S62 is shown in the figure. Not limited to order). Then, the processing control unit 22 compares the number of characters between the text extracted in S60 and the text obtained by the character recognition in S62, and adopts the one having the larger number of characters as the final character recognition result. (S64). In consideration of the possibility of character recognition error, in S64, the number of characters obtained by character recognition in S62 is sufficiently larger than the number of characters in the text extracted in S60 (that is, the difference between the two is equal to or greater than the threshold value). ) The character recognition result of S62 may be adopted only in the case, and the text extraction result of S60 may be adopted otherwise.

FIG. 11 shows a processing procedure of still another example of automatic selection of the text acquisition method from the application page.

In the procedure of FIG. 11, the processing control unit 22 uses the text extraction unit 26 to extract text data from the text object in the application page, and in the entire area of the application page, the text area in which the text object exists. Specify the position and dimensions (S70). The position and dimensions of the text area can be obtained from the parameters included in the text object (for example, those that define the text flow range). Further, the processing control unit 22 converts the application page into image data using the rasterizer 28, and performs character recognition on the image in the range other than the above-mentioned text area in the image data (S72). Then, the processing control unit 22 adopts the text extracted in S70, the text obtained by the character recognition in S72, and the final character recognition result for the application page (S74).

FIG. 12 shows a processing procedure of still another example of automatic selection of the text acquisition method from the application page.

This example is an example of processing that takes into consideration that the text contained in the application page is not always visible to the human eye when the application page is displayed (drawn).

For example, if the application sets the text in the created document to be completely transparent, or the text color of the text is set to the same color as the background color, the text is invisible to the human eye. Also, if you obscure a text object with another opaque object, that text object becomes invisible (the "other object" is visible when displayed). If the creator of the application file wants to leave textual information in the file in a way that is invisible to third parties, they may use those methods. In addition, some applications leave a character string deleted as a result of a user's editing operation as a kind of history on the page as transparent characters.

This kind of "invisible text" cannot be recognized even if the application page is converted (that is, drawn) into an image and character recognition is performed, but it is extracted from the application page by the text extraction unit 26. Can be done. Therefore, if an app page containing "invisible text" is included in the target of character recognition processing, either the appearance (no characters are displayed) or the content (text data exists) as a result of the processing. The question is whether to prioritize. When giving priority to appearance, the text extracted from the application page is discarded, and when giving priority to content, the extracted text is adopted as the character recognition result.

In the example of FIG. 12, which of the two is prioritized is left to the judgment of the user instructing the character recognition process. That is, the user registers the value of the setting item indicating which of the two is prioritized in the setting management unit 18. Alternatively, the user selects whether to prioritize the appearance or the content on the confirmation screen displayed by the UI processing unit 16 when the "invisible text" is detected from the application page.

In the procedure of FIG. 12, the processing control unit 22 uses the text extraction unit 26 to extract text data from the text object in the application page (S80). Further, the processing control unit 22 converts the application page into image data using the rasterizer 28, and character recognition is performed on the image data by the character recognition engine 24 (S82). Then, the processes S84 to S92 are performed for each area in which at least one of the characters in the text extracted in S80 or the characters recognized in S82 exists in the entire area of the application page.

Since the OCR algorithm detects the position of the recognized character in the page, the area of each character in the character recognition result can be specified from the information on the position of the character. In addition, the application page contains information indicating the area of the text object in the page (for example, defined as a rectangular area) and information such as the font size, distance spacing, and line spacing of each character in the text object. have. The position of each character in the text object can be calculated from that information. From this information, it is sufficient to identify the area where at least one of the characters extracted from the text object and the characters recognized by the characters exists. The area in which neither the character extracted from the text object nor the character recognized character exists is an area in which there is no character at all and is not subject to the processing of FIG.

In S84, the processing control unit 22 determines whether or not the character extracted by the text extraction unit 26 exists in the area currently attracting attention. If the result of this determination is No, the character recognized by the character recognition process exists in the area. In this case, the processing control unit 22 adopts the recognized character as the character recognition result for the area (S86).

When it is determined that the character extracted by the text extraction unit 26 exists in the area currently attracting attention in S84, the processing control unit 22 determines whether or not the character recognized character exists in the area (S88). .. When the determination result of S88 is Yes, both the characters extracted by the text extraction unit 26 and the characters recognized by the character recognition engine 24 are present in the area. Since the result of character recognition may be incorrect, in this case, the character extracted by the text extraction unit 26 is adopted as the character recognition result for the area (S90).

When the determination result of S88 is No, the extracted text characters exist in the area, but the character-recognized characters do not exist. This is the case of "invisible text" mentioned above. In this case, the processing control unit 22 asks the user whether to prioritize the appearance or the content, or determines based on the setting content registered in the setting management unit 18. When the appearance is prioritized, it is assumed that there are no characters in the area (that is, the character recognition result is "no character"), and when the content is prioritized, the text extraction unit is used as the character recognition result in the area. The text of the area extracted by 26 is adopted (S92).

In the process illustrated in FIG. 12, when the character extracted from the text object in the application page and the character recognized by the character recognition process conflict, basically, the character extracted from the text object which is accurate is adopted. .. However, if the character is not recognized by the character recognition process from the place where the character is extracted from the text object, it is "invisible text" and it is not always possible to give priority to the uniformly extracted text. Therefore, ask the user to make a decision.

The character recognition processing unit 20 may be programmed to always apply any one of the automatic selection processes of the text acquisition method described with reference to FIGS. 9 to 12 to the application page.

Further, the automatic selection of the text acquisition method may be presented to the user as a processing option for the application page, and the user may be made to select it. For example, in the setting screen 200a shown in FIG. 13, in the processing option column 204a for the application page at the time of OCR, in addition to the first to third options shown in FIG. 3, the automatic selection of the text acquisition method is listed as the fourth option. ing. Since the method of the fourth option takes significantly longer to process than the other options, the description of the fourth option in the option column 204a is provided with a proviso that the process takes time. .. A user who wants to obtain as much text information as possible with the highest quality even if it takes a certain amount of time may select this fourth option. When the fourth option of the setting screen 200a is registered in the setting management unit 18 as the method selected by the user, the processing control unit 22 shows the application pages of the character recognition targets in FIGS. 9 to 12. Process the app page using one of the above methods.

The document handling system 10 or the character recognition processing unit 20 exemplified above is realized by, for example, causing a computer to execute a program representing the functions of each of the devices. Here, the computer is, for example, as hardware, a microprocessor such as a CPU, a memory (primary storage) such as a random access memory (RAM) and a read-only memory (ROM), and an HDD controller that controls an HDD (hard disk drive). Various I / O (input / output) interfaces, network interfaces that control for connection with networks such as local area networks, and the like have a circuit configuration connected via, for example, a bus. In addition, portable non-volatile recording of various standards such as a disk drive and flash memory for reading and / or writing to a portable disk recording medium such as a CD or DVD via an I / O interface, for example, on the bus. A memory reader / writer for reading and / or writing to the medium may be connected. A program describing the processing contents of each functional module illustrated above is stored in a fixed storage device such as a hard disk drive via a recording medium such as a CD or DVD, or via a communication means such as a network, and stored in a computer. Will be installed. By reading the program stored in the fixed storage device into the RAM and executing it by a microprocessor such as a CPU, the functional module group illustrated above is realized.

10 document handling system, 12 file import unit, 14 document processing unit, 16 UI processing unit, 18 setting management unit, 20 character recognition processing unit, 22 processing control unit, 24 character recognition engine, 26 text extraction unit, 28 rasterizer.

Claims (17)

When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document .
The second-class page processing is a document processing device that notifies a user that a page to which the character recognition cannot be applied is included when the second-class page is included in the document . In the notification process, the user is further inquired as to whether or not to acquire character information from the second type page.
The document processing device controls whether or not to execute a process of acquiring character information from the second type page in response to a user's response to the inquiry.
The document processing apparatus according to claim 1 . The document processing apparatus according to claim 1, wherein the type 2 page processing is a processing for acquiring character information from the type 2 page. The document processing apparatus according to claim 3 , wherein the type 2 page processing is a processing for extracting text data included in the type 2 page. When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document.
The second-class page processing is a document processing apparatus that converts the second-class page into image data and executes the character recognition on the image data. When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document.
The document processing apparatus has a process of extracting text data included in the type 2 page, and a process of converting the type 2 page into image data and executing the character recognition on the image data. A document processing device that executes a process selected by a user as the second-class page process. When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document.
The second-class page processing is processing for the second-class page based on the text data extracted from the second-class page and the character recognition result for the image data obtained by converting the second-class page. A document processing device that is a process for obtaining the resulting character information. In the type 2 page processing, for a certain place in the type 2 page, the first character corresponding to the place in the text data extracted from the type 2 page and the type 2 page are converted. The seventh aspect of claim 7 , wherein the first character is adopted as character information about the place when both the second character corresponding to the place in the result of the character recognition for the image data are present. Document processing device. In the type 2 page processing, there is no character corresponding to the place in the text data extracted from the type 2 page for a certain place in the type 2 page, and the image obtained by converting the type 2 page. The document processing according to claim 7 or 8 , wherein when there is a character corresponding to the place in the result of the character recognition for the data, the character in the result of the character recognition is adopted as character information about the place. Device. In the type 2 page processing, with respect to a certain place in the type 2 page, there is a character corresponding to the place in the text data extracted from the type 2 page, and the image obtained by converting the type 2 page. When there is no character corresponding to the place in the result of the character recognition for the data, whether or not the character in the text data is adopted as the character information about the place is controlled according to the user's instruction. Item 6. The document processing apparatus according to any one of Items 7 to 9 . When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document.
When the document contains the second type page, a notification process for notifying the user that the page to which the character recognition cannot be applied is included, and
The process of acquiring character information from the second type page and
A document processing apparatus having a means for presenting a setting screen for receiving a setting from a user as to which of the two is executed as the second type page processing. When the execution of character recognition is instructed for a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed, the first type included in the document. A document processing device that executes the character recognition for the type 1 page, and executes the type 2 page processing, which is a process different from the character recognition, for the type 2 page included in the document.
When the document contains the second type page, a notification process for notifying the user that the page to which the character recognition cannot be applied is included, and
The process of extracting the text data contained in the second type page and
A process of converting the second type page into image data and executing the character recognition on the image data.
A document processing apparatus having a means for presenting a setting screen for allowing a user to select two or more of them as options for the second type processing. When a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to execute character recognition, the first type included in the document. To operate the computer as a means for executing the character recognition for the type 1 page and performing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. It ’s a program,
The second-class page processing is a notification process for notifying a user that a page to which the character recognition cannot be applied is included when the second-class page is included in the document . When a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to execute character recognition, the first type included in the document. To operate the computer as a means for executing the character recognition for the type 1 page and executing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. It ’s a program,
The second-class page processing is a program that converts the second-class page into image data and executes the character recognition on the image data. When a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to execute character recognition, the first type included in the document. To operate the computer as a means for executing the character recognition for the type 1 page and performing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. It ’s a program,
The type 2 page processing is processing for the type 2 page based on the text data extracted from the type 2 page and the result of the character recognition for the image data obtained by converting the type 2 page. A program that is a process for obtaining the resulting character information. When a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to execute character recognition, the first type included in the document. To operate the computer as a means for executing the character recognition for the type 1 page and executing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. It ’s a program,
The computer
When the document contains the second type page, a notification process for notifying the user that the page to which the character recognition cannot be applied is included, and
The process of acquiring character information from the second type page and
A means for presenting a setting screen that accepts from the user the setting of which of the two is executed as the second type page processing.
A program characterized by functioning as. When a document in a data format in which a type 1 page in an image data format and a type 2 page in a non-image data format can be mixed is instructed to execute character recognition, the first type included in the document. To operate the computer as a means for executing the character recognition for the type 1 page and executing the type 2 page processing which is a process different from the character recognition for the type 2 page included in the document. It ’s a program,
The computer
When the document contains the second type page, a notification process for notifying the user that the page to which the character recognition cannot be applied is included, and
The process of extracting the text data contained in the second type page and
A process of converting the second type page into image data and executing the character recognition on the image data.
A means for presenting a setting screen for allowing the user to select two or more of them as options for the second type processing.
A program characterized by functioning as.

Images