JP2008084186A - Image processing system and image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing system and an image processing program which can suppress increase in transformation rule information. <P>SOLUTION: The system has an input means for inputting a document image which has two or more areas containing character strings, a split means for dividing the document image inputted by the input means for every area, a recognition means for recognizing the character string from the document image inputted by the input means, a determination means for determining whether the character string which has been recognized by the recognition means is contained in the character string which is defined for every area memorized beforehand for every area divided by the split means, and a converting means for converting the character string recognized by the recognition means to the specific character string corresponding to the character string which is defined for every area memorized beforehand when it is determined that the character string recognized by the recognition means is contained in the character string which is defined for every area memorized beforehand. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing system and an image processing program.

  In recent years, the digitization of paper documents has progressed, and various processes have been performed on documents obtained by digitizing paper documents with a scanner or the like (hereinafter referred to as image data). For example, from a wide variety of image data, a format by image processing (difference in table position and size, etc.) or a character by optical character recognition processing (hereinafter referred to as OCR processing) is recognized, and the recognition result is used. A relationship with other image data may be held as conversion rule information and converted according to the conversion rule information. More specifically, conversion rule information is used in advance for image data for forms, such as different form formats or information (price and price, etc.) indicating the same contents but different expressions for Company A and Company B. And used when converting form image data of company A into form image data of company B.

As a technique related to these, for example, Patent Document 1 discloses a document management apparatus capable of effectively managing document dependency. Also, for example, Patent Document 2 creates, implements and uses a “smart” document generated by a computer to which a function for providing context-sensitive tools, control and help contents to a document user is added. A method and system are disclosed.
Further, for example, Patent Document 3 discloses a method and apparatus that can insert a new element into an XML document based on an insertion position designated on the display document of the XML document.
JP 2003-281118 A Japanese Patent Laid-Open No. 2004-46828 JP 2004-272684 A

  The present invention has been made in the background of such a background art, and an object of the present invention is to prevent an increase in conversion rule information, and an image processing capable of suppressing an increase in conversion rule information. It is an object to provide a system and an image processing program.

  In order to achieve the above object, the invention of claim 1 includes an input means for inputting a document image having a plurality of areas including character strings, a dividing means for dividing the document image input by the input means for each area, A recognition unit for recognizing a character string from the document image input by the input unit, and a character string recognized by the recognition unit for each region divided by the dividing unit is determined for each prestored region. A determination means for determining whether or not the character string is included in the character string, and the character string recognized by the recognition means by the determination means is included in a character string determined for each prestored area. A conversion means for converting the character string recognized by the recognition means into a specific character string corresponding to a character string determined for each of the previously stored areas. It is characterized in.

  In the present invention, as in the invention of claim 2, the character string determined for each of the previously stored areas belongs to a group provided for each character string having the meaning indicated by the specific character. The specific character string may be associated with each group.

  In the present invention, as in the invention of claim 3, the character string determined for each of the previously stored areas may include a character string when the recognition unit erroneously recognizes the character string.

  The present invention changes the specific character string according to the output document image in order to output an output document image that is a document image including the content indicated by the specific character string. You may make it have further the change means to do.

  According to the present invention, as in the invention of claim 5, the area further includes value information corresponding to the character string and indicating a predetermined value, and the recognizing means recognizes the value information and changes the value. The means may change the value recognized by the recognition means according to the output document image.

  In order to achieve the above object, the invention of claim 6 includes an input step of inputting a document image having a plurality of regions including character strings, and a dividing step of dividing the document image input by the input step into regions. A recognition step for recognizing a character string from the document image input by the input step, and a character string recognized by the recognition step for each region divided by the division step is determined for each region stored in advance. A determination step for determining whether or not the character string is included in the character string, and the determination step includes that the character string recognized by the recognition step is included in a character string determined for each prestored area. If it is determined that the character string recognized by the recognition step is a character string determined for each of the previously stored areas It executes processing and a conversion step of converting the specific character string to respond in the computer.

  According to the first aspect of the present invention, since it is possible to have conversion rule information for each area instead of the entire document image, compared to the case where the present configuration is not provided, the same area is used for other document images. Since there is no need to create new conversion rule information when there is an image, an image processing system capable of suppressing an increase in conversion rule information can be provided.

  According to the second aspect of the present invention, the character string having the meaning indicated by the specific character can be unified with the specific character as compared with the case where the present configuration is not provided.

  According to the third aspect of the present invention, it is possible to convert the character string into the specific character string even when the character string is erroneously recognized as compared with the case where the present configuration is not provided.

  According to the fourth aspect of the present invention, the specific character string can be changed according to the output document image as compared with the case where the present configuration is not provided.

  According to the fifth aspect of the present invention, the value indicated by the value information corresponding to the character string and indicating the predetermined value is changed according to the output document image as compared with the case where the present configuration is not provided. be able to.

  According to the invention described in claim 6, since it is possible to have the conversion rule information for each area instead of the entire document image as compared with the case where this configuration is not provided, the same area is used for other document images. Since there is no need to create new conversion rule information when there is an image, an image processing program capable of suppressing an increase in conversion rule information can be provided.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a conceptual module configuration diagram of an embodiment of the present invention.

  The module generally refers to a component such as software or hardware that can be logically separated. Therefore, the module in the present embodiment indicates not only a module in a program but also a module in a hardware configuration. Therefore, the present embodiment also serves as an explanation of a program, a system, and a method. In addition, the modules correspond almost one-to-one with the functions. However, in mounting, one module may be composed of one program, or a plurality of modules may be composed of one program. A plurality of programs may be used. The plurality of modules may be executed by one computer, or one module may be executed by a plurality of computers in a distributed or parallel environment. Hereinafter, “connection” includes not only physical connection but also logical connection.

  The system includes a configuration in which a plurality of computers, hardware, devices, and the like are connected via a network and the like, and includes a case where the system is realized by a single computer, hardware, devices, and the like.

  Next, the outline of the processing according to the present embodiment will be described with reference to FIG. In this embodiment, as shown in the figure, the conversion rule information is stored not for each document image but for each region (table regions a, b, c, d).

  In the present embodiment, as shown in the figure, in different document images A and B, the table areas a, b, c, and d are extracted from the document image A, and the extracted table areas a, b, and c are extracted. , D are converted using the conversion rule information according to the document image B, and rearranged according to the document image B. The types of the table areas a, b, c, and d here are, for example, organization and product.

  By doing so, it is possible to suppress an increase in the conversion rule information even if there are two types of document images A and B as shown in FIG.

  Next, the module configuration according to the present embodiment will be described with reference to FIG. As shown in FIG. 2, the module according to the present embodiment includes an input module 18, a division module 20, a recognition module 22, a determination module 24, a conversion module 26, a change module 28, and an output module 30. .

  The input module 18 is connected to the determination module module 24 and the recognition module 22 and inputs a document image having a plurality of regions including character strings.

  Here, the document image is an electronic document, and more specifically includes a document that is an image input by a scanner, a document image generated by a document generation application (for example, a word processor), and the like. The inputting includes acquiring using a scanner, acquiring a document from a document database, acquiring from an external system via a communication line, and the like. Further, the document to be acquired may be a single page or a multi-page document.

  The input module 18 is connected to the recognition module 22, and the relationship between them will be described later.

  The division module 20 is connected to the input module 18 and the recognition module 22 and divides the document image input by the input module 18 for each area.

  The recognition module 22 is connected to the input module 18, the division module 20, the determination module 24, the conversion module 26, and the change module 28. The recognition module 22 is a module that recognizes a character string from the document image input by the input module 18. The recognition module 22 may recognize a character string from the document image divided by the dividing module 20, or may be input by the input module 18. After the character string is recognized from the document image, the document image may be divided by the dividing module 20.

  The determination module 24 is connected to the recognition module 22 and determines whether or not the character string recognized by the recognition module 22 is included in a character string determined for each prestored area.

  The conversion module 26 is connected to the recognition module 22 and the change module 28, and the determination module 24 includes the character string recognized by the recognition module 22 in a character string determined for each pre-stored area. When it is determined, the module converts the character string recognized by the recognition module 22 into a unique attribute name (specific character string) corresponding to a character string determined for each prestored area. The change module 28 is connected to the change module 28 and the output module 30, and in order for the output module 30 to output an output document image that is a document image including the content indicated by the unique attribute name, the unique attribute name is output to the output document. This module changes according to the image. The change module 28 changes the value recognized by the recognition module 22 according to the output document image.

  The output module 30 is connected to the change module 28 and outputs an output document image including the unique attribute name and value changed by the change module 28.

  Next, a hardware configuration example of the image processing system according to the embodiment will be described with reference to FIG. The configuration illustrated in FIG. 3 is an image processing system configured by, for example, a personal computer (PC), and illustrates a hardware configuration example including a data reading unit 1217 such as a scanner and a data output unit 1218 such as a printer. . This hardware configuration is also applied to other embodiments.

  As shown in FIG. 3, an image processing system according to this embodiment includes a CPU (Central Processing Unit) 1201, a ROM (Read Only Memory) 1202, a RAM (Random Access Memory) 1203, a host bus 1204, a bridge 1205, and an external bus. 1206, interface 1207, keyboard 1208, pointing device 1209, display 1210, HDD (Hard Disk Drive) 1211, drive 1212, connection port 1214, external connection device 1215, communication unit 1216, data reading unit 1217, and data output unit 1218 Have.

  The CPU 1201 is a control unit that executes processing according to the various modules described in the above-described embodiments, that is, the computer program describing the execution sequence of each module.

  The ROM 1202 stores programs used by the CPU 1201, calculation parameters, and the like. The RAM 1203 stores programs used in the execution of the CPU 1201, parameters that change as appropriate during the execution, and the like. These are connected to each other by a host bus 1204 including a CPU bus.

  The host bus 1204 is connected to an external bus 1206 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 1205.

  A keyboard 1208 and a pointing device 1209 such as a mouse are input devices operated by an operator. The display 1210 includes a liquid crystal display device or a CRT (Cathode Ray Tube), and displays various information as text and image information.

  The HDD 1211 has a built-in hard disk, drives the hard disk, and records or reproduces a program executed by the CPU 1201 and information. Various computer programs such as various data processing programs are stored in the hard disk.

  The drive 1212 reads data or a program recorded on a mounted removable recording medium 1213 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and the data or program is read from the interface 1207 and the external bus 1206. , The bridge 1205, and the RAM 1203 connected via the host bus 1204. The removable recording medium 1213 can also be used as a data recording area similar to a hard disk.

  The connection port 1214 is a port for connecting the external connection device 1215 and has a connection unit such as USB, IEEE1394. The connection port 1214 is connected to the CPU 1201 and the like via the interface 1207, the external bus 1206, the bridge 1205, the host bus 1204, and the like. The communication unit 1216 is connected to a network and executes data communication processing with the outside. The data reading unit 1217 is a scanner, for example, and executes document reading processing. The data output unit 1218 is a printer, for example, and executes document data output processing.

  Note that the hardware configuration of the image processing system shown in FIG. 3 shows one configuration example, and the image processing system of the present embodiment is not limited to the configuration shown in FIG. 2 and will be described in the present embodiment. Any configuration can be used as long as the module can be executed. For example, some modules may be configured by dedicated hardware (for example, ASIC), and some modules may be in an external system and connected via a communication line. A plurality of systems shown in FIG. 5 may be connected to each other via communication lines so as to cooperate with each other. Further, it may be incorporated in a copying machine, a fax machine, a scanner, a printer, a multifunction machine (also called a multi-function copying machine, which has functions of a scanner, a printer, a copying machine, a fax machine, etc.).

  The described program can be stored in a recording medium, and the program can be provided by communication means. In that case, for example, the above-described program can also be regarded as an invention of a “computer-readable recording medium recording the program”.

  The “computer-readable recording medium on which a program is recorded” refers to a computer-readable recording medium on which a program is recorded, which is used for program installation, execution, program distribution, and the like.

  The recording medium is, for example, a digital versatile disc (DVD), which is a standard established by the DVD Forum, such as “DVD-R, DVD-RW, DVD-RAM,” and DVD + RW. Standards such as “DVD + R, DVD + RW, etc.”, compact discs (CDs), read-only memory (CD-ROM), CD recordable (CD-R), CD rewritable (CD-RW), etc. MO), flexible disk (FD), magnetic tape, hard disk, read only memory (ROM), electrically erasable and rewritable read only memory (EEPROM), flash memory, random access memory (RAM), etc. It is.

  The program or a part of the program can be recorded on the recording medium and stored or distributed. Also, by communication, for example, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a wired network used for the Internet, an intranet, an extranet, etc., or wireless communication It can be transmitted using a transmission medium such as a network or a combination of these, and can also be carried on a carrier wave.

  Furthermore, the above program may be a part of another program, or may be recorded on a recording medium together with a separate program.

  Next, conversion rule information, which is information for converting image data, will be described. The conversion rule information is a method for recognizing a format by image processing (such as a difference in the position and size of a table) or a character by OCR processing from a variety of image data A, and using the recognized result to other image data B. This is information indicating the relationship between the image data A and the image data B used for conversion.

  In this embodiment, there are attribute conversion rule information and attribute name conversion rule information as types of conversion rule information.

  Among these, the attribute conversion rule information is a character string determined for each pre-stored area, and is provided for each type. Here, the type refers to the type of content represented by the character string shown in the image data area (area where the table structure is considered to be formed, and this area is obtained by the division module 20 of FIG. 2). It is expressed by classification.

  In this area, for example, when the document image is a form, a character string related to product information is described in the product name, price, etc., or if the document image is an in-house document, a character string related to organization information is described. . As described above, the character strings described in the same area have associated features, and the type represents a type obtained by classifying such related character strings for each field. Specific types include, for example, the product number and organization name described above.

  Also, the attribute conversion rule information has a role to absorb word notation fluctuation and convert it into one character string. Here, the notation fluctuation indicates that there are a plurality of different notations (character strings) for the same object. Further, in the present embodiment, as will be described later, since the character string is read by executing the OCR process on the image data, the attribute conversion rule information includes the OCR error in consideration of the occurrence of erroneous character recognition due to the OCR process. It also has a role to absorb and convert it into a character string that should be recognized.

  First, the details of the attribute conversion rule information will be described with reference to FIGS. FIG. 4 shows attribute conversion rule information related to a product as an example of attribute conversion rule information. The attribute name shown in the figure indicates a character string belonging to a predetermined type. The attribute names are further classified into “notation” and “OCR”. “Notation” indicates a character string that is shaking. This notation fluctuation indicates that there are a plurality of different notations (character strings) for the same object.

  “OCR” indicates a character string that is erroneously recognized as a character string that is correctly recognized by the OCR process.

  Also, the unique attribute name shown in the figure indicates one or more attribute names by one character string, and indicates a character string that absorbs the above-mentioned notation fluctuation and OCR error.

  In the example shown in the figure, a product number, a product number, a product code, etc. are shown as notation fluctuations. A product incense number or the like is shown as an OCR error corresponding to these character strings, and a character string in which these are absorbed is a product number.

  Thus, in the attribute conversion rule information, a character string (notation: product number, product number, etc.) defined for each pre-stored area is a character having a meaning indicated by a unique attribute name (product number). It belongs to a group provided for each column, and a unique attribute name (product number) is associated with each group (a group of character strings including a product number and a product number).

  Next, attribute conversion rule information relating to organization information, which is another example of attribute conversion rule information, will be described with reference to FIG. The attribute names, notation fluctuations, OCR errors, and unique attribute names shown in the figure are as described in FIG.

  In the figure, XXXXX technical opening and the like are shown as notation shaking. As the OCR error corresponding to these character strings, XXXXX Technical Opening Department, XXXXX Technical Opening, etc. are shown, and the character string that absorbed them is the XXXXX Technical Development Department. ing.

  The attribute conversion rule information described above includes many types in addition to the above-described types related to product information and types related to organization information.

  Next, attribute name conversion rule information will be described with reference to FIG. The attribute name conversion rule information is information in which a unique attribute name is associated with an attribute character string corresponding to the output. Using this attribute name conversion rule information, the change module 28 (see FIG. 2) converts the unique attribute name converted using the attribute conversion rule information by the conversion module 26 (see FIG. 2) into an attribute corresponding to the output. Convert to string.

  The output attribute name is a character string corresponding to output, for example, an attribute name corresponding to an output destination such as Company A. Specifically, it is shown that the product number shown in the figure is changed to a product code when company A is the output destination.

  Based on the attribute conversion rule information and attribute name conversion rule information described above, the character string read by the OCR process is first converted into a unique attribute name using the attribute conversion rule information, and the unique attribute name is the attribute name conversion DB. Is used to change the output destination attribute name.

  Note that by setting conversion rule information in advance according to the type, it is possible to convert notation of the corresponding attribute value without changing the attribute name. For example, if it is an amount, the currency unit is changed, and if it is a date field, the Western calendar and the Japanese calendar are converted. Furthermore, even if the date field is the same, more flexible operation is possible by setting the case where the conversion should be performed and the case where the conversion should not be performed according to the type of area.

  Each DB described above can also be created using XML. In this case, an attribute name that is a character string read by OCR processing is set as an XML tag, and an output destination attribute name is also set as an XML tag, and is defined using an XML schema so that these tags can be converted. To do.

  Hereinafter, processing according to the present embodiment will be described with reference to each flowchart. Each flowchart shown below shows processing executed by the CPU 1201 on image data indicating a form read by a scanner or the like.

  First, the region division process will be described with reference to the flowchart of FIG. The details of this area division processing are disclosed in Japanese Patent Laid-Open No. 2002-203249.

  Based on the above, the flowchart of FIG. 7 will be described. First, in step 101, the dividing module 20 creates a circumscribed rectangle for the connected portion. A connection part is an area | region connected like the ruled line of a table | surface, for example. A rectangle surrounding the connected portion is a circumscribed rectangle.

  In the next step 102, the dividing module 20 determines the length H and the length W of the circumscribed rectangle of the created circumscribed rectangle, and the circumscribed rectangle sides for determining a predetermined table candidate. Compare the length threshold. When the length H of the vertical side and the length W of the horizontal side of the circumscribed rectangle are less than the threshold value, the dividing module 20 determines a set of the connected portion and circumscribed rectangle as a table candidate in step 106. Exclude from

  On the other hand, if at least one of the vertical side length H or the horizontal side length W of the circumscribed rectangle is longer than the threshold value for determining the table candidate in step S102, the division module 20 In Step 103, the pixel density in the circumscribed rectangle is calculated for the set of the connected portion and circumscribed rectangle.

  The pixel density in the circumscribed rectangle is obtained as a ratio between the area obtained by the product of the vertical side length H and the horizontal side length W of the circumscribed rectangle and the number of all pixels constituting the connected portion. Can do.

  In the next step 104, the division module 20 determines whether or not the calculated pixel density is equal to or less than a threshold value relating to the pixel density for determining a predetermined table candidate. If it is determined, in step 105, the dividing module 20 determines that the combination of the connected portion and the circumscribed rectangle is a table candidate. On the other hand, if the calculated pixel density exceeds the threshold value, the dividing module 20 excludes the combination of the connected portion and the circumscribed rectangle from the table candidates in step 106 described above.

  Table candidates are obtained by the above processing. Processing for determining whether or not the table candidate is a table will be described with reference to the graphs (a) and (b) of FIG. 9 and the flowchart of FIG.

The division module 20 pays attention to the circumscribed rectangle of the connected pixel component that is a table candidate, and calculates the center coordinates (X _CENTER , Y _CENTER ) in _S1111 . In the next step S1112, the segmentation module 20 is focused on the projection distribution in the horizontal direction, a portion having a frequency greater than the frequency threshold value TH _H in the projection distribution, the possibility of vertical ruled line exists extending in a vertical direction It is determined that there is a vertical ruled line candidate and is detected. When no vertical ruled line candidate is detected, the dividing module 20 determines in step S1113 that the table candidate is not a table.

When the vertical ruled line candidates are detected at one or more places in step S1112, the dividing module 20 pays attention to the vertical projection distribution. In step 1114, the dividing module 20 determines that there is a possibility that a horizontal ruled line extending in the horizontal direction exists in the projection distribution having a frequency exceeding the frequency threshold value TH _V and detects it as a horizontal ruled line candidate. To do. If no horizontal ruled line candidate is detected, the dividing module 20 determines in step S1113 that the table candidate is not a table.

When both the horizontal ruled line candidate and the vertical ruled line candidate are detected in the above processing, the dividing module 20 determines whether or not there are two or more horizontal ruled line candidates. Here, if there is only one horizontal ruled line candidate, the dividing module 20 determines whether or not there are two or more vertical ruled line candidates in step S1116. On the other hand, when there are two or more horizontal ruled line candidates, the division module 20 uses the circumscribed rectangle center coordinates Y _CENTER calculated in step S1111 as a boundary, and uses the vertical projection distribution shown in FIG. in step S1117 whether a horizontal ruled line candidate exists that exceeds the frequency threshold TH _V in two projection interval Y _START to Y _CENTER and Y _CENTER to Y _END obtained by dividing in the boundary in at least one or more places, respectively judge.

If this condition is satisfied, the dividing module 20 determines in step S1119 that the closed region surrounded by the circumscribed rectangle of the connected pixel component that is the table candidate is a table region. On the other hand, if this condition is not satisfied and there are two or more vertical ruled line candidates in step S1116, the horizontal coordinate shown in FIG. 9A is set with the circumscribed rectangle center coordinate X _CENTER calculated in step S1111 as a boundary. or vertical ruled line candidate exceeding the frequency threshold TH _H in at least one location or more, respectively present a projection distribution of directions in the two projection interval X _START to X _CENTER and X _CENTER to X _END obtained by dividing in the boundary In step S1118, the dividing module 20 determines. When this condition is satisfied, the division module 20 determines in step S1119 that the closed region surrounded by the circumscribed rectangle of the connected pixel component that is the table candidate is a table region. Conversely, if this condition is not satisfied, the dividing module 20 determines in step 1113 that the table candidate is not a table.

  Through the above processing, the dividing module 20 divides the document image for each area. In addition to the processing described in FIG. 8, the region may be divided by the processing shown in FIG. First, in step 201, the division module 20 performs labeling. This labeling is a process of adding the same label value to all the pixels constituting the connected component. In the next step 202, the dividing module 20 creates a circumscribed rectangle of the connected component, and in step 203 calculates the lengths of the vertical and horizontal sides of the circumscribed rectangle. In the next step 204, the dividing module 20 classifies the circumscribed rectangles into character / figure / field separator / noise candidates. Details of the processing relating to this classification are disclosed in paragraphs 0032 to 0037 of Japanese Patent Laid-Open No. 2000-90194, and thus further explanation here is omitted.

  Next, the type specifying process will be described with reference to the flowchart of FIG. First, in step 301, the recognition module 22 extracts a table cell existing in the table area obtained by the area dividing process. In the next step 302, the recognition module 22 recognizes the character string in each cell existing in the table area by OCR processing.

  Next, in step 303, the determination module 24 determines whether the character string recognized by the recognition module 22 is included in the attribute conversion rule information. If the determination module 24 makes an affirmative determination in step 303, the recognition module 22 holds the character string in step 304. On the other hand, if the determination module 24 makes a negative determination in step 303, the process proceeds to step 305.

  In step 305, the recognition module 22 determines whether all character strings existing in the table area have been extracted. If the recognition module 22 makes a negative determination, the process proceeds to step 301 again. If the recognition module 22 makes an affirmative determination, in step 306, the recognition module 22 sets the type indicated by the attribute DB including the largest number of stored character strings as the type of the table area.

  The processing in step 306 will be described in detail. As described above, the attribute conversion rule information includes many types such as those related to product information, and the processing in step 306 is performed for each character string existing in one table. Is a process for determining that the type indicated by the attribute conversion rule information including the most is the type of the table area. Further, as described above, since the attribute conversion rule information includes a character string that has been shaken and OCR, the character string is also considered in step 306.

  The attribute value extraction process shown in FIG. 12 is executed on the table area whose type is determined in this way. This process is a process of extracting the price and 500 yen as one set when, for example, the price is written in one cell of the table and 500 yen is written on the right side of the cell. In this example, 500 yen is the attribute value.

  First, in step 401, the recognition module 22 determines whether or not the character string in the cell is an attribute name. If the recognition module 22 makes a negative determination, the process proceeds to step 406, and if an affirmative determination is made, an adjacent cell is specified in step 402. As an example of this adjacent cell, there is a cell on the right side or on the lower side where an attribute value is often described in a normal document.

  In the next step 403, the recognition module 22 determines whether or not the character string in the identified cell is an attribute name. If the character string is an attribute name, the attribute value cannot be obtained, and thus the process proceeds to step 406. If the character string is not an attribute name, the recognition module 22 determines in step 404 the character string (attribute) It is determined whether or not (value) corresponds to the attribute name. For example, when the attribute name is a price, it is a process of determining whether or not the character string is not a price such as 2006. This processing is executed using, for example, natural language analysis. In this case, for example, it is possible to determine whether or not the character string in the cell corresponds to the character string shown in the specific expression example shown in FIG. it can.

  If the recognition module 22 makes a negative determination in step 404 above, the process proceeds to step 406. If an affirmative determination is made, the recognition module 22 extracts a pair of attribute name and attribute value as an attribute in step 405. In the next step 406, the recognition module 22 determines whether or not the processing from step 401 to step 405 has been executed for all the cells. If the determination is affirmative, the processing ends. If the determination is negative, the recognition module 22 The processing after step 401 is executed again for the cells that have not been processed.

  Next, output document image generation processing will be described with reference to FIG. This image data generation process is a process of converting a character string into a corresponding unique character string and outputting an output document image that is a document image including the content indicated by the unique character string. In order to output the output document image, the unique character string and value (attribute value) are changed according to the output document image.

  First, in step 501, the conversion module 26 converts an attribute name of attributes extracted by the attribute value extraction process (see FIG. 12) into a unique attribute name based on the attribute conversion rule information. In the next step 502, the change module 28 converts the unique attribute name into an output attribute name based on the attribute name conversion DB. For example, if the output destination is company A, the product number is converted into a product code. Further, in step 503, the change module 28 changes the attribute value according to the output document image.

  In the next step 504, the output module 30 generates an output document image in which the changed attribute name and attribute value are arranged based on the format of the output destination, and ends the processing.

  An example of the region converted by the above processing will be described with reference to FIG. FIG. 15 shows a region before conversion and a region after conversion. As shown in the figure, “product code” is converted to “product number”, “maker” is converted to “manufacturer”, and “price” is converted to output attribute names “price”. The attribute value “¥ 500” is converted to “$ 5.00”. The output module 30 generates an output document image by combining the regions thus converted.

  Note that the image processing system according to the present embodiment has specialized servers A, B, C, and D for each type, for example, as shown in FIG. It is also possible to output a document image by integrating with an integrated server. In this case, the type identification process (see FIG. 11) is executed by one computer, the table area information corresponding to the servers A, B, C, and D for executing the process for the type is transferred. The attribute value extraction process (see FIG. 12) is executed, and the integrated server executes the image data generation process (FIG. 14) for the obtained information.

  The flow of the processing of each flowchart described above is an example, and it is possible to change the processing order, add a new step, or delete an unnecessary step without departing from the gist of the present invention. Needless to say.

It is a schematic diagram which shows the outline | summary of the process which concerns on this Embodiment. It is a figure which shows the module structure which concerns on this Embodiment. It is a figure which shows the structure of the personal computer which concerns on this Embodiment. It is a figure which shows the example of attribute conversion rule information regarding the merchandise information which concerns on this Embodiment. It is a figure which shows the example of attribute conversion rule information regarding the organization information which concerns on this Embodiment. It is a figure which shows the example of attribute name conversion rule information which concerns on this Embodiment. It is a flowchart (the 1) which shows the area division | segmentation process which concerns on this Embodiment. It is a flowchart which shows the table candidate determination process which concerns on this Embodiment. It is a figure which shows the projection distribution which concerns on this Embodiment. It is a flowchart (the 2) which shows the area division | segmentation process which concerns on this Embodiment. It is a flowchart which shows the classification specific process which concerns on this Embodiment. It is a flowchart which shows the attribute value extraction process which concerns on this Embodiment. It is a figure which shows the example of the specific expression which concerns on this Embodiment. It is a flowchart which shows the image data generation process which concerns on this Embodiment. It is a figure which shows the example of area | region conversion which concerns on this Embodiment. It is a figure which shows the example of a structure of the image processing system which concerns on this Embodiment.

Explanation of symbols

18 Input Module 20 Division Module 22 Recognition Module 24 Judgment Module 26 Conversion Module 28 Change Module 30 Output Module 1201 CPU
1211 HDD

Claims (6)

Input means for inputting a document image having a plurality of regions including character strings;
Dividing means for dividing the document image input by the input means for each area;
Recognition means for recognizing a character string from the document image input by the input means;
Determining means for determining whether or not the character string recognized by the recognizing means for each area divided by the dividing means is included in a character string determined for each prestored area;
If the determination means determines that the character string recognized by the recognition means is included in a character string determined for each pre-stored area, the character string recognized by the recognition means Conversion means for converting the character string defined for each of the previously stored areas into a specific character string corresponding to the character string.   The character string determined for each pre-stored area belongs to a group provided for each character string having the meaning indicated by the specific character string, and the specific character string is associated with each group. The image processing system according to claim 1.   The image processing system according to claim 1, wherein the character string determined for each of the previously stored areas includes a character string when the recognition unit erroneously recognizes the character string.   4. The apparatus according to claim 1, further comprising a changing unit configured to change the specific character string in accordance with the output document image in order to output an output document image that is a document image including content indicated by the specific character string. The image processing system according to claim 1. The area further includes value information corresponding to the character string and indicating a predetermined value,
The recognizing means recognizes a value from the value information;
The image processing system according to claim 4, wherein the changing unit changes a value recognized by the recognizing unit according to the output document image. An input step for inputting a document image having a plurality of regions including character strings;
A division step of dividing the document image input by the input step into regions;
A recognition step of recognizing a character string from the document image input by the input step;
A determination step of determining whether or not the character string recognized by the recognition step for each region divided by the division step is included in a character string determined for each pre-stored region;
When it is determined by the determination step that the character string recognized by the recognition step is included in a character string determined for each pre-stored area, the character string recognized by the recognition step An image processing program for causing a computer to execute a process including: a conversion step of converting a predetermined character string corresponding to each of the previously stored areas and a specific character string corresponding thereto.

Images