JPH09128400A - Structured document preparing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the diversion property of document data and the possibility of an automatic database register processing by processing an attribute in a document tag based on a designated document type and generating a structured document. SOLUTION: The structured document preparing device is provided with a document picture data input means 1, a document picture area and logical structure correspondence definition storage part 2, a document picture area segmenting means 3, a pattern recognizing means 4, a structured document preparing means 5, a structured document output means 6, a control means 7 and a document-type definition storage part 35. The document picture area segmenting means 3 segments a prescribed part against a pattern-recognized input through the use of information concerning a document picture area in the correspondence definition storage part 2. When a stipulated output document is designated, the structured document preparing means 5 refers to a definition name and logical structure corresponding to designation in the document-type definition storage part 35, processes the area which is segmented by the document picture area segmenting means 3 by a processing attribute in the correspondence definition storage part 2 and prepares the structured document in accordance with logical structure.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to document image information in document management / processing such as document management database, document creation / approval, document storage / search / browsing, data aggregation, document conversion, and publication printing. The present invention relates to a structured document generation device that extracts a document logical structure from a document, converts the document information into a format suitable for computer processing, and refers to the information of the document type definition.

[0002]

2. Description of the Related Art FIG. 59 shows, for example, Japanese Unexamined Patent Publication No. 6-2.
This is the configuration of the conventional structured document generation device disclosed in 14983. However, in order to facilitate comparison, part of the device name has been changed. The purpose of this invention is to
It is to convert a plurality of document images represented by value images into a document having a logical structure divided into chapters, sections, document paragraphs, figures and the like. In the figure, reference numeral 41 denotes a rectangle / temporary block generation unit, which extracts regions adjacent to each other from a document image and analyzes a rough layout. A character recognition unit 42 recognizes characters. A header / footer analysis unit 43 extracts a header / footer area of the document image and identifies a page number and common content. 44
Is a column determination / block construction unit, which determines the column area of each page and constructs a block in the page. 45
Is a section number / display attribute analysis unit, which analyzes the display attributes in each section and divides the document into paragraphs. A logical structure generation unit 46 pays attention to a document paragraph at the end of a column / page and a document paragraph at the beginning of a subsequent column / page, and if the display attributes are almost equal, merges to generate a logical structure of the document.

Next, the operation will be described. In the figure,
The rectangle / temporary block generation unit 41 extracts a basic rectangle called a rectangle from the document image, and the character recognition unit 42,
It is sent to the header / footer analysis unit 43 and the column determination / block determination unit 44. The character recognition unit 42 character-recognizes the rectangle data and sends the result back to the rectangle / temporary block generation unit 41. The header / footer analysis unit 43 analyzes the header / footer of the rectangle data and outputs the result. The column confirmation / block confirmation unit 44 constructs the header / footer block of the rectangle data and sends the constructed block data to the node number / display attribute analysis unit 45. The node number / display attribute analysis unit 45 sends final block data generated by checking the rectangular attributes (position, size, offset value, etc.) of the block data to the logical structure generation unit 46. The logical structure generation unit 46 generates a logical structured document having a structure such as a section / paragraph from the final block data.

[0004]

The above-mentioned prior art has the following problems (1) to (3). (1) The format of the structured document to be output cannot be freely selected. That is, in the document management work, it is necessary to output a structured document having different structure information with respect to the input document image data. A structured document is generated by applying rules regarding what order and processing is performed for each document area. However, in the conventional technology, since only the information about the display data of the document image is used in the process of generating the logical structure, the required structured document format of the output is:
I cannot apply the rules to output freely. (2) Information on standard formats cannot be used effectively. That is, in the standard format of a paper document,
In many cases, it is predetermined that what is written in which place. In the prior art, since information related to standard formats other than images is not managed, it is not possible to freely specify and process these input information and effectively utilize them. (3) Information to be added cannot be specified for each document area. In other words, in order to freely select the format of the structured document to be output, a means for designating the information regarding what document structure should be added to what document area is required. . However, in the prior art, since there is no means for managing the information to be added, it is impossible to specify what information should be added to each document area.

The present invention has been made to solve the above problems, and has a user interface capable of easily generating and storing a rule for associating information relating to a document structure with each partial area in a document image. Another object of the present invention is to obtain an apparatus for generating / using a structured document according to a document type definition set according to an application from input and document image data with reference to this association rule.

[0006]

In the structured document generation apparatus according to the present invention, it is indicated that a document image area which is a predetermined portion of input document image data is a specific logical structure portion of a document. A document image area / logical structure correspondence definition storage means (hereinafter referred to as correspondence definition storage means) for storing a definition by adding a document tag to the output document and a necessary processing attribute to the output document, and pattern recognition is performed. A document image area cutout means for cutting out a predetermined portion by using information about the document image area in the correspondence definition storage means for an input, and a logic describing a collection of the document image area which is the above-mentioned portion necessary for various output documents. Document type definition storing means for storing the structure as a definition by giving a name to each document, and when a specified output document is designated, corresponds to the designation in the stored document type definition storing means. And a structured document generating means for generating in accordance with the process described above logical structure by processing attribute in the corresponding definition storage means an area cut out the document image area extracting means with reference to the definition name and logical structure.

In addition to the basic structure, a predetermined portion of the input is further cut out and compared with the corresponding document image area in the correspondence definition storage means, and if it is judged that they are similar to each other with a certain evaluation value or more, the correspondence definition storage means. A document image area estimation means for replacing the inside document image area or judging as equivalent to any of the stored document image areas is added.

Further, another detailed processing table is designated as a processing attribute, and a partial area of the input document image data is processed based on the detailed processing table to generate an output document.

Furthermore, a predetermined image portion of the input is cut out to form an image file, which is combined with the document tag and output.

In addition to the above, the output document is generated by referring to a table prepared as another detailed processing table for a predetermined part of the input or applying an algorithm to convert it into another character string. I did it.

Further, when a document tag indicating a portion of a logical structure forming a document is designated for a document image area which is a predetermined portion of the cut out input, the document image area and the document tag are linked. A document image area / logical structure associating unit to be registered / stored in the correspondence definition storing unit is added.

In addition to the above, in addition to the document tags, an attribute for performing a predetermined process on the output document is added and registered / stored in the correspondence definition storage means.

Further, when a display means for inputting document image data is added and a document tag is designated, the document image area corresponding to the document image data is identified and displayed.

Furthermore, when an arbitrary document type definition stored in the document type definition storage means is read out and displayed in a tree structure, or when the document type definition of the tree structure displayed after being corrected is instructed to be stored, the document type is stored. A document logical structure display / correction means to be stored in the definition storage means is added.

In addition to the above, the document logical structure display / correction means is added to display the modified arbitrary document type definition in a tree structure, and the document image area / logical structure associating means is displayed in a tree structure. Correspondence between each document tag and the document type definition is also performed.

In addition to the above, a display means for input document image data is added so that when a document tag is designated, the document image area corresponding to the document image data is identified and displayed.

In addition to the above, the document image area / logical structure associating unit also associates each document tag with the document type definition displayed in a tree structure, and if there is an independent tree structure with no correspondence, The document logical structure display / correction means identifies and displays a tree structure that does not correspond.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1 FIG. First, in the first embodiment, a configuration and an operation of generating and outputting a structured document of a designated format using a pre-stored association rule will be described. Embodiment 1 of the present invention will be described below with reference to the drawings. 1 is a block diagram of a structured document generation device according to a first embodiment of the present invention. In the figure, reference numeral 1 is a document image data inputting means which comprises an optical image reading device having a seed feeder, a handwriting reading device using a display-integrated tablet, or the like, and which converts the read image data into image data that can be processed by a computer. Is. For example, the image data shown in FIG. 5, which will be described later, is input and recognized to recognize or code an image. Here, a structured document refers to a document by adding a document tag that represents the role of a document component such as a title, an author name, an introductory text, and a text that are distinguished by the description content of the document. It is a document that has information describing the logical structure of the document by describing the relationship in document tag units.
In addition, the document type definition is a structural element (document tag unit) in the document identified by the document tag in the structured document.
Is a description that defines the appearance order of.

In the structured document, since the constituent elements in the document can be determined by the document tag expressing the role of the constituent elements of the document in the document, and the attribute for the subsequent processing is added, the document management data The computer can automatically extract information required by applications such as base, document creation / approval, document storage / search / browsing, data aggregation, document conversion, and publication printing. Therefore, further automation of the computer system using documents can be promoted. In addition, the flexibility of the system can be realized by setting the document type definition according to the application to be linked. On the other hand, in offices, there are many paper documents with formats such as forms that represent the roles of document constituent elements in a document by enclosing them in a rectangular frame. Therefore, if a structured document that complies with the document type definition set according to the application can be automatically generated from the existing formatted paper document, the data on the existing paper document can be automatically converted into a computer application. By incorporating the above, it becomes possible to effectively use office paper documents.

Reference numeral 2 denotes a document image area / logical structure correspondence definition storage unit whose storage state is shown in FIG. This is a main element of the present invention, and describes the correspondence relationship between the “partial area of the document image data” and the “markup tag expressing a certain part of the logical structure of the document” and the information related thereto. The “document image area / logical structure correspondence definition” is stored. Reference numeral 35 denotes a document type definition storage means for storing a definition whose storage state is shown in FIG. It stores a document type definition that defines the logical structure of the document. Reference numeral 3 denotes a document image area cutout unit, which is a main element along with the document image area / logical structure correspondence definition storage unit 2. This is a document image portion obtained by applying the document image data input from the document image data input means 1 to the document image area / logical structure correspondence definition stored in the document image area / logical structure correspondence definition storage unit. The area is cut out, and the cut out document image data is sent to the pattern recognition means 4. Reference numeral 4 is a normal pattern recognition means for generating coded data by recognizing the pattern of the document image data cut out by the document image area cutting means 3 and sending the coded data to the structured document generation means 5. Since the technique of pattern recognition is known, the detailed configuration and description of the operation will be omitted.

Reference numeral 5 denotes a structured document generation means, which is a main element along with the document image area / logical structure correspondence definition storage section 2 and the document image area cutout means 3. This is a pattern using the document image area / logical structure correspondence definition stored in the document image area / logical structure correspondence definition storage unit 2 while referring to the document type definition stored in the document type definition storage means 35. A structured document is generated by adding tags to the data encoded by the recognition means 4 for each partial data of the document, and sent to the structured document output means 6. Reference numeral 6 is a structured document output means for outputting the structured document generated by the structured document generation means 5 to a display device such as a monitor or a storage device such as a desk. Reference numeral 7 is a control means which is composed of a CPU (central processing unit) and controls the operation of the entire system.

FIG. 5 is a diagram showing an example of document image data, which is a document including character strings, bullets, images and tables.
Reference numeral 51 is the entire document image, 52 to 59 are areas surrounded by a closed curve on the document screen data, and 60 is an area in which the date is described.

FIG. 6 is a diagram showing an example of the document image area / logical structure correspondence definition. For the input data shown in FIG.
"Partial area of document image data" and "Markup tag that represents a specific part of the logical structure of the document"
It is a description of the correspondence relationship with and information related to it. Here, 61 is a partial area definition. Reference numeral 62 denotes an area ID, and in this document image area / logical structure correspondence definition, five partial areas are defined. In 63, the position and shape of the partial area are described. In 64, a sequence of document tags associated with each partial area is described. Another important point of the present invention is that
This is because, in the correspondence definition, an attribute designating a process to be performed later by the structured document generation means 5 is added for each ID. FIG.
64-68 is that part. In 65, the setting value of the “cutout mode” given for each partial area is described, and is referred to when determining the branch destination in step 8 of FIG. In 66, "character replacement guideline" given to each partial area
Is described and referred to when determining the procedure for replacing the character string in step 15 of FIG. In 67, the “text analysis guideline” given to each partial area is described.
In step 17 of FIG. 3, it is referred to when determining a procedure for analyzing a text. In 68, the setting value regarding the “recognition attribute” given to each partial area is described.
In step 13 of, the process is referred to when selecting a pattern recognition procedure according to the recognition attribute.

Reference numeral 69 is information about "repeating structure extraction 1" which is a set value of 67, and is given when "a rule defining a repeating structure appearing in the text" and "when converting the repeating structure into a structured document. Information about document tags to be written "is described. In step 17 of FIG. 3, it is referred to by the procedure for analyzing the text. Reference numeral 70 is information regarding “repeating structure cutout 1” which is the setting value of 65, and includes “shape of repeating structure”, “rule of arrangement”,
Then, “information regarding a document tag to be added when converting the repeating structure into a structured document” is described. 71
Is the "shape" for the repeating region ID introduced in 70
And "information about document tags to be added when converting to structured document" are described. 70 and 71 are the steps in FIG.
In p10, it is referred to by the procedure for recognizing the repeating structure in the document image data. Reference numeral 72 describes the setting value of the “document image area estimation mode”, which is referred to when the branch destination in step 5 of FIG. 2 is determined. A document type definition name 73 is described, and is referred to when the document type definition to be called is determined in step 19 of FIG. A set value of “external procedure calling mode” is described in 74, which is referred to when the branch destination in step 24 of FIG. 4 is determined. FIG. 11 is a diagram showing an example of another correspondence definition.

FIG. 7 is an example diagram showing an example of the document type definition, which defines the order of the document tags appearing in the document, and its meaning conforms to the document type definition in SGML of ISO standard 8879. For example, the first line indicates that the tags <name>, <date>, and <body> appear in this order in the area specified by the document tag <catalog>. FIG. 12 is a diagram showing an example of another document type definition.

FIG. 8 is an example diagram showing an example of the document tag / coded data correspondence table, and describes the relationship regarding the structure between the coded data in the pattern recognition means 4. 91 to 95 are (document tag string, encoded data,
3), which is created in step 18 of FIG. 3 and is used when generating a structured document in step 21 of FIG. FIG. 9 is a diagram showing a tree structure so that a person can easily understand the logical connection relationship of FIG.
The same relationship as in FIG. 8 is expressed.

FIG. 10 is an example diagram showing an example of a structured document, which is created in step 23 of FIG.
It is output at ep25. Reference numerals 101 to 109 are partial logical structures having a subordinate relationship to the entire structured document. FIG. 13 is a diagram showing an output example of another structured document.

The overall operation of the structured document generation apparatus configured as described above will be described with reference to the flowcharts of FIGS. In the first embodiment, the document image area / logical structure correspondence definition estimation mode, the document image area estimation mode, the character replacement mode, the text analysis mode, and the external procedure calling mode are all off in the flowcharts of FIGS. This is the case where the cutout mode is the normal mode.

First, the document image data input means 1 reads the document image data (step 1). Next, the control means 7 checks the document image area / logical structure correspondence definition estimation mode, and proceeds to step 3 if it is on and proceeds to step 4 if it is off (step 2). Document image area
In the case where the logical structure correspondence definition is the example shown in FIG. 4, since it is off as indicated by 75 in FIG. 6, the specified document image region / partial region definition of the logical structure correspondence definition is read (step 4). . The case where the document image area / logical structure correspondence definition estimation mode is ON is described in the eighth embodiment. The partial region definition is an important part in this embodiment. This is a part of the document image area / logical structure correspondence definition, and for each partial area in each document image determined by the area ID, "start point", "end point", "document tag string", "cutout mode" , "Character replacement guideline", "text analysis guideline", "recognition attribute", and the like. By storing this definition (processing rule), the structured document generation means 5 can output the designated structured document through the intermediate table as described in detail below.

The partial area definition corresponds to the area 61 in the document image area / logical structure correspondence definition shown in FIG.
D is from 1 to 5 as shown in part 62, 5
Information about one partial area is described. For example, the partial area determined by "area ID" 2 is
The "start point" is (30,50) and the "end point" is (160,8)
0), “Document tag string” is <Catalog><Date>, “Cutout mode” is usually “Character replacement guideline”, “Text analysis guideline” is not shown, and “Recognition attribute” is print / kanji. . Next, the control means 7 checks the document image estimation mode, and if it is on, proceeds to step 6, and if it is off, proceeds to step 7 (step 5). When the document image area / logical structure correspondence definition is an example shown in FIG. 6,
Since it is off as indicated by 72, the process proceeds to step 7.
If the document image estimation mode is on, the second embodiment
It describes in. In step 7, the partial areas defined by the area ID are sequentially extracted from the partial area definition, and if extracted, proceed to step 8, and if all have been extracted, proceed to step 19. When the partial area definition is 61 of the document image area / logical structure correspondence definition shown in FIG. 6, the partial areas corresponding to the area ID1 to the area ID5 are sequentially extracted.

When the partial area can be taken out in step 7, the control means 7 checks the cutout mode,
If it is the normal mode, proceed to step 9, if it is the repeated structure recognition mode, proceed to step 10, if it is the nested structure recognition mode, proceed to step 11, and if it is the image cutout mode, proceed to step 12 (step).
8). For example, in the case of the partial area determined by the “area ID” 2, it is the normal mode, so the process proceeds to step 9.
When the cutout mode is the repeated structure recognition mode, the third embodiment is described. When the cutout mode is the nested structure recognition mode, the fourth embodiment is described. When the cutout mode is the image cutout mode, the fifth embodiment is described. Enter. In step 9, the document tag string in the partial area definition extracted in step 7 is extracted and "start point"
By obtaining the image data by cutting out the area designated by "end point", a (document tag string, image data) pair is created, and the process proceeds to step 13. For example, "area ID"
When the partial area determined by 2 is taken out, the document tag string is <catalog><date and time>. Further, the image data has (30, 50) as the upper left vertex and (16
0,80) is the rectangular area with the lower right vertex as step1
This is data obtained as a result of cutting out from the document image data read in. However, here, (30, 50) indicates a point on the document image data whose X coordinate is 30 and Y coordinate is 50, and in the example of the document image data of FIG.
Corresponding to

In step 13, step 9, step 9
10, by replacing the image data in the data of the (document tag string, image data) pair generated in step 11 with the coded data using a known pattern recognition technique (document tag string, Coded data) pairs are generated. For example, in the case of the partial area determined by the "area ID" 2, the coded data "rice cooker" can be obtained from the image data obtained in step 12, so it is called (<catalog><date> rice cooker). A (document tag string, coded data) pair is generated. Next, the control means 7 checks the character replacement mode, and if it is on, s
If it is off, go to step 15 and go to step 16 (st
ep14). For example, in the case of the partial area determined by the “area ID” 2, since it is off, the process proceeds to step 16. Further, when the character replacement mode is ON, it will be described in the seventh embodiment. In step 16, the text analysis mode is checked, and if it is on, the process proceeds to step 17, and if it is off, the process proceeds to step 18. For example, in the case of the partial area determined by the “area ID” 2, since it is off, the text analysis mode is off, and the process proceeds to step 18. If the text analysis mode is on, the sixth embodiment
It describes in.

In step 18, step 13, step
The pair (document tag string, coded data) obtained in p15 and step 17 is added to the end of the document tag / coded data correspondence table, and the process returns to step 7. The document tag / coded data correspondence table is a so-called intermediate table, which is temporarily generated in the main memory not described in the text or the memory of the control means. This is the "document tag sequence" as shown in FIG.
It is a correspondence table of and "coded data". For example, ste
When the partial area determined by the “area ID” 2 is selected in p7, as shown at 92 in FIG.
obtained by step 3 (<catalog><date>,
A pair (document tag string, coded data) called rice cooker is registered. In addition, the document image of FIG.
When the processing of step 18 for all the five partial areas shown in FIG. 8 is completed, the document tag / coded data correspondence table shown in FIG. 8 is obtained.

On the other hand, in step 7, if the partial area cannot be extracted or if all the partial areas have been extracted, the structured document generation means 5 causes the document type definition storage means 35.
Read the document type definition specified by the document type definition name of the document image area / logical structure correspondence definition (step 19),
The document tag sequence is taken out in order from the document type definition (step 2
0), if the document tags can be extracted, go to step 21, and if all the document tag strings have been taken out, step 24.
Proceed to. In the example of the document image area / logical structure correspondence definition shown in FIG. 6, “/usr/local/dtd/catalog.dt” shown at 73 is displayed.
“D” is the document type definition name. FIG. 7 shows an example of the document type definition designated by the document type definition name. The document type definition defines the order of the document tags that appear in the document. The first line in the example of FIG. 7 is <name>, <date> in the area specified by the document tag <catalog>. , <
It indicates that the document tags "text>" appear in this order.
Similarly, the 4th and 5th lines in FIG. 7 indicate that the document tags in parentheses appear in order within the area specified by the document tag that appears next to <! ELEMENT. The second line in FIG. 7 indicates that a character string appears after the document tag <name>. Similarly, the 3rd, 7th, 8th, 11th, and 12th lines in FIG. 7 also indicate that a character string appears next to the document tag that appears next to <! ELEMENT in each line. The sixth line in FIG. 7 shows that 0 or more document tags <item> appear repeatedly after the document tag <list>. The 10th line of FIG.
After the document tag of <table>, a pair of document tags of <attribute> and <value> appears repeatedly. When the parent-child relationship of the document tags in the definition of FIG. 7 is represented by a tree structure, FIG.
become that way.

Next, a column including the retrieved document tag sequence is retrieved from the document tag / coded data correspondence table (s
step 21). If the same document tag string is consecutive in the document tag / coded data correspondence table, all of them are taken out. For example, when <catalog><date and time> is taken out at step 20, the portion 92 of the document tag / coded data correspondence table of FIG. 8 is taken out. Also, step2
0 in <Catalog><Body><Features><List><Item>
8 is taken out, the part 93 of the document tag / coded data correspondence table of FIG. 8 is taken out. Next, a document tag is added to the coded data of the extracted column (step 2)
2). The document tag to be added is a part obtained by removing the common part from the document tag string extracted immediately before. For example, s
In step 20, when the document tag string <catalog><text><appearance> is taken out, it is a common part with the previous document tag string <catalog><text><features><list><item>.<Catalog><Body> removed
Appearance> document tag is added. After all, step2
When the document tag string <catalog><text><appearance> is extracted at 0, the portion of the document tag / coded data correspondence table 94 of FIG. 8 is extracted at step 21, so the coded data “picture. Append the document tag <appearance> to "img" and add <appearance> picture.img <
Get the character string \ appearance>. The document tag <appearance> is the document tag indicating the end of the document tag <appearance>. Next, the document-tagged coded data generated in step 22 is added to the end of the structured document being created (step 23).

For example, in step 20, after the end of step 23 before the document tag string <catalog><text><appearance> is taken out, the structured document shown in FIG. 14A is intermediately generated. . Also, step2
At the end of step 23 after the document tag string <catalog><text><appearance> is retrieved at 0,
By adding the character string <appearance> picture.img <\ appearance> obtained in step 22 to the end,
The structured document shown in (b) is intermediately generated. And
In the case of the document tag / coded data correspondence table shown in FIG. 8, the structured document shown in FIG. 10 is obtained after all the tags have been extracted in step 20.

When the document image area / logical structure correspondence definition shown in FIG. 11 and the document type definition shown in FIG. 12 are used for the document image data shown in FIG. 5, the document type definitions shown in FIGS. By following the flowchart, the structured document shown in FIG. 13 is obtained. The structured document shown in FIG. 13 differs from the structured document shown in FIG. 10 in terms of the data representation format such as the document tag name, the document structure, and the date and time. For example, in the document tag name, for <name> in FIGS. 2 to 4,
In FIG. 13, <Product Name> is set, and the expression format of data such as date and time is September 25, 1995 in FIG. 10, whereas it is 1995/9/25 in FIG. 13. The difference in the format of the date and time is that in the document image area / logical structure correspondence definition, the value of "character replacement guideline" of "area ID" 1 is the conversion of the Japanese era in the case of FIG. 6, but in the case of FIG. It depends on the story. Therefore, by exchanging the document image area / logical structure correspondence definition and the document type definition, it is possible to freely select the format of the structured document to be output, that is, the document tag name, the document structure, the data representation format, etc. It has been shown.

At step 24, by the control means 7,
The external procedure calling mode is checked, and if it is on, proceed to step 25, and if it is off, proceed to step 25. If the document image area / logical structure correspondence definition is an example shown in FIG. 6, the external procedure calling mode is off, so the structured document generated in step 23 is a display device such as a monitor or a storage device such as a desk. Output to (ste
p25). Further, when the external procedure calling mode is ON, it will be described in the ninth embodiment.

Therefore, according to the above configuration,
In the processing from step 7 to step 18, since the coded data correspondence table is created from the document image cut out by using the information about the position of the partial area described in the document image area / logical structure correspondence definition, the fixed format in the document image data is created. Can effectively utilize the information on the specific format. In addition, based on the document tag / coded data correspondence table that describes the information about the partial areas obtained by steps 7 to 18, the document type definition is referenced with the document tag string in the document tag / coded data correspondence table as a key. However, since the structured document is configured, the format of the structured document to be output can be freely selected.

Embodiment 2 In the second embodiment, an error in the size of a region between a plurality of pieces of document image data is absorbed, so to speak, a standard partial region and a partial region of new input document image data, which are stored in the correspondence definition storage unit. Fluctuation,
Alternatively, an example in which processing is possible even if there is a difference in format will be described. This can deal with a reading error, and can reduce the number of registrations in the correspondence definition and the document type definition. FIG. 15 is a configuration diagram of the structured document generation device according to the second embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 8 receives the document image data input by the document image data input means 1, and stores an area surrounded by a closed curve in the image and a document image area / logical structure correspondence definition storage unit 2
The document image area is estimated by comparing it with the "partial area of the document image data" stored in, and the estimated partial area of the document image is cut out. It is a document image area estimation means for sending to.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 16 showing the details of step 6 thereof. The second embodiment is a case where the image area estimation mode is ON in the flowchart of FIG. 2, and the control means 7 checks the document image area estimation mode (step 5) and step 6.
Proceed to. The operation of step 6 will be described with reference to the flowchart of FIG. In the document image data (hereinafter referred to as X) input by the document image data input means 1, the area surrounded by the closed curve is recognized (step 31). For example, in the case of the document image data shown in FIG. 5, a region surrounded by a rectangle such as a name, a feature, and an external view is recognized. Then, the partial areas Y _i in the document image area / logical structure correspondence definition are sequentially taken out one by one (step 32), and if they are all taken out, the processing goes to step 33, and if all are taken out, the processing goes to step 7 in FIG. 2. However, i is the document image area
It is a natural number equal to or less than the number n of partial areas in the logical structure correspondence definition. In the case of the document image area / logical structure correspondence definition in FIG. 6, the partial area is a rectangular area designated by two points on the two-dimensional plane, and n is 5.

Next, the area having an overlapping portion with Y _i is taken out from the area in X recognized in step 31 (step 33). However, the extracted area is called Z _ij, and the area of the overlapping portion of Z _ij and Y _i is called A _ij . However, j is the number m of partial areas of X having an overlapping portion with Y _i.
It is the following natural number. Then, when you move the fixed and i j, (A area _ij) / ((the area of the Y _i) + (Z area _ij)) is seeking Z _ij as a maximum (step 34), the document image area Replace the partial area Y _i in the logical structure correspondence definition with Z _ij and return to step 32 (step 35). According to the above configuration, the format of the document image data input by the document image data input means 1 is defined as the document image area / logical structure correspondence definition stored in the document image area / logical structure correspondence definition storage unit 2. Even if the format is slightly different from the format defined by, the correspondence between the document image area and the logical structure can be correctly recognized. This function is effective when the data input by the document image data input means 1 is deviated or when the document image data having a slightly different format is handled.

In the above example, the document image area estimating means 8 uses the closed curve to determine the partial area in the document image data. , Or “closed curve obtained by interpolating a dotted line” may be used to determine the partial region.

Embodiment 3. Another example of reducing the number of registrations in the definition will be described. FIG. 17 is a configuration diagram of the structured document generation device according to the third embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 9 receives the document image data input by the document image data input means 1, recognizes an area surrounded by a closed curve that appears repeatedly in the image data, cuts out the area, sends it to the pattern recognition means 4, and repeats it. It is a repetitive structure recognizing means for sending information about the structure to the structured document generating means 5.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 18, which is a detail of step 10 thereof. The third embodiment is a case where the cutout mode is the repetitive structure recognition mode in the flowchart of FIG.
Area I in the document image area / logical structure correspondence definition in FIG.
This corresponds to the case of processing the partial area of D5. Control means 7
Checks the cutout mode (step 8), s
Proceed to step 10. The operation of step 10 will be described with reference to the flowchart of FIG. An area X _i surrounded by a closed curve is recognized in the document image data (hereinafter referred to as X) input by the document image data input means 1 (s
step 41), where i is an integer from 1 to the number n of recognized regions.

Next, since the text analysis guideline 67 in the area ID5 of the document image area / logical structure correspondence definition of FIG. 6 is "repeated structure cutout 1", complicated attribute processing that cannot be accommodated in this section is separately described. It is a detailed processing table prepared. The text analysis guideline 70 is read.
As a result, the number m (2 in this case) of repeatedly appearing areas and the shape Y _j are obtained from the document image area / logical structure correspondence definition using the repeated area ID as a key (step 4).
2). Next, by sequentially associating the shapes obtained in step 42, a candidate of Y _j corresponding to X _i is obtained (st
ep43). It is confirmed whether the repeating area relation described in the document image area / logical structure correspondence definition is satisfied (s
(Step 44). St if the repeating structure is satisfied
Proceed to ep45, and if not, return to step43. In the case of the document image area / logical structure correspondence definition in FIG. 6, the repeating area relationship is described in the portion 70,
In this case, it means that the rectangles R1 and R2 defined by 71 are arranged in this order in the right lateral direction, and the arrangement is repeated vertically downward. R1 and R2 correspond to the regions 58 and 59 of the document image image of FIG. Therefore, in step 44, “R obtained as a candidate of Yj is obtained.
Make sure that the shapes 1 and R2 are lined up on the right side, and that the pair of R1 and R2 is repeatedly lined up vertically.

Next, with respect to the pair of the area X _i and the shape Y _j obtained in step 43, the number of appearances of the shape Y _j is set to A _i, and then the area X _i is cut out from the document image data and cut out. The acquired image data is set to Z _i (step 4
6), (Document tag string, Z _i , A _i ) triplets step
13 (step 47). Note that the document tag string specifies the document tag string of the partial area being processed with the document tag specified by the repeating area relationship added. For example, when it is desired to process the portion 59 of FIG. 5 which is an example of the document image data, it corresponds to the portion R1 of 71 of FIG. 6, and therefore the document tag of “area ID” 5 of 61 of FIG. By adding 70 parent tags <table> and 71 document tags <attribute> to <text><specification>,<catalog><text>
<Specification><table><attribute> is obtained as a document tag string.
Further, the data obtained by encoding the image data of the portion 59 of FIG. 5 is the size, and the portion of the portion 59 of FIG.
Since it corresponds to the second column of the table of No. 56, A _i becomes 2.
(<Catalog><text><specification><table><attribute>, size, 2) (document tag string, Z _i , A _i ) triplet is generated. This triplet corresponds to the third row in 95 of the document tag string / coded data correspondence table in FIG. Further, since the part 58 in FIG. 5 corresponds to the part R2 in 71 in FIG. 6, the document tag <value> is used instead of the document tag <attribute> 71.
By the same procedure as the processing of the 59 part (<catalog><text><specification><table><
Value>, 50 × 30 × 30, 2) (document tag string, Z
_i , A _i ) triplets are generated. This triplet is shown in Figure 8.
This corresponds to the fourth line in 95 of the document tag string / coded data correspondence table of. In step 7, if the partial area of “area ID” 5 of 61 in FIG. 6 is extracted, st
In ep18, the portion 95 in FIG. 8 is added to the document tag string / coded data correspondence table. This part corresponds to the logical structure element 105 shown in the structured document FIG. The document image data is as shown in FIG.
When the logical structure correspondence definition is shown in FIG. 6, the area 56 in FIG.
Generate a logical structure corresponding to the table shown in FIG.
p23).

According to the above-mentioned structure, the structured document can be generated even when the document image input by the document image data inputting means 1 includes the repeating structure surrounded by the closed curve. In the case of the document image data shown in FIG. 5, the structural information of the table in the area 56 can be recognized. In the above embodiment, the repeated structure recognition means 9 uses the closed curve to determine the partial area in the document image data. However, the "character attribute such as font or size" or The partial area may be determined by using a “closed curve obtained by interpolating a dotted line or the like”.

Embodiment 4 When the partial areas of the input document image data have an inclusive relation, that is, when there is a nested structure between the data, the number of registrations in the definition can be similarly reduced. This will be described below. FIG. 19 is a configuration diagram of the structured document generation device according to the fourth embodiment.
In the figure, the same or corresponding portions as those of the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 10 receives the document image data input by the document image data input means 1, recognizes the inclusion relation between the areas surrounded by the closed curves in the document image data, and outputs information on the inclusion relation to the structured document generation means. It is a nested structure recognition means to be sent to 5.

The operation of the structured document generation apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 20 showing the details of step 11 thereof, and the correspondence definition of FIG. FIG. 21 is an example diagram showing an example of the document image area / logical structure correspondence definition. The part 141 is different from the document image area / logical structure correspondence definition 64 of FIG. For example, the third line of 64 is <catalog><text><feature>, and the fourth line of 141 is <feature>. In this way, by recognizing the inclusion relationship in the document image, the nesting relationship in the structured document can be automatically extracted, so that the definition in the document image area / logical structure correspondence definition can be simplified. 142 indicates that the cutout mode is the nested structure mode. This embodiment is a case where the cutout mode is the nested structure recognition mode in the flowchart of FIG. 3, and corresponds to, for example, the case of processing the partial area of the area ID3 in the document image area / logical structure correspondence definition of FIG. Reference numeral 142 in FIG. 21 indicates that the cutout mode is the nested structure mode. The control means 7 checks the cutout mode (step 8), and proceeds to step 11. The operation of step 11 will be described with reference to the flowchart of FIG.

An area X _i surrounded by a closed curve is recognized in the document image data (hereinafter referred to as X) input by the document image data input means 1 (step 51), where i is an area recognized from 1 It is an integer up to the number n. Next, for each i, recognizes the inclusion relation between the retrieved area X _i (the step 52), retrieves the area X _j including the area X _i, for the collection and Y _i (step
53). In the case of the document image data of FIG. 5, the area including the area 54 is the area 53, and the area including the area 55 is 53. Next, a region X _i is cut out from the document image data, and the cut out image is set as Z _i (step 5
4), (Document tag string, Z _i , Y _i ) triplets step
Generate at 55 and move to step 13. The document tag string generated at this time is obtained by concatenating the document tag string of the area including this area before the document tag string of the area cut out in the document image area / logical structure correspondence definition. To be For example, in the document image area / logical structure correspondence definition of FIG. 21, the area ID corresponding to the area 53 is 3, that is, the document tag string is <catalog><body>, and the area ID corresponding to the area 54 is 4, that is, the document tag string is <feature>. Therefore, when the area 54 is cut out, the document tag string sent to step 13 is <catalog><text><feature>. The parts other than the generation of the document tag string are the same as those in the first embodiment.

With the above arrangement, the inclusion relation between the areas in the document image input by the document image data input means 1 can be recognized as the nesting relation in the structured document. In the above embodiment, the nested structure recognition means 10 uses the closed curve in order to determine the partial area in the document image data. However, the "character attribute such as font or size" or The partial area may be determined by using a “closed curve obtained by interpolating a dotted line or the like”.

Embodiment 5 The case where the input document has an image as an image and is used as it is will be described. That is, the image area may be cut out, and the cut out image file may be named and used as the attribute of the document tag. FIG. 22 is a configuration diagram of the structured document generation device according to the fifth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted.
Reference numeral 11 denotes an image cutout unit that receives the document image data, cuts out the “partial area of the document image data” and outputs it as an external file, and at the same time sends the name of the external file to the structured document generation unit 5.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 23 showing the details of step 12 thereof. The fifth embodiment is a case where the cutout mode is the image cutout mode in the flowchart of FIG. 3, and corresponds to, for example, the case of processing the partial area of the area ID4 in the document image area / logical structure correspondence definition of FIG. The control means 7 checks the cutout mode (step 8), s
Proceed to step 12. The operation of step 12 will be described with reference to the flowchart of FIG.

An image area is cut out from the document image data input by the document image data input means 1 by referring to the information about the area shown in 63 of the document image area / logical structure correspondence definition in FIG. 6 (step 6).
1). Next, image the cut out document image data.
It is saved as a file (step 62), a name is given (step 63), and a pair of (area ID, image file name) is sent to step 13 (step 64). For example, in step 7, “area ID” 4 61 in FIG.
When the partial area of is extracted, the document tag string / coded data correspondence table 94 in FIG.
Is added. This portion corresponds to the logical structure element shown at 104 in the structured document FIG.

If the document image data is as shown in FIG. 5 and the document image area / logical structure correspondence definition is as shown in FIG. 6, the area 55 in FIG. 5 is recognized as an image area and is saved as an external image file. , Structured document Figure 10
A logical structure to be referred to like a file name picture.img indicated by 104 is generated (step 23). According to the above configuration, even if the document image data includes data that is difficult to convert into coded data such as diagrams and photographs, a part of the read document image is referred to as an external file. A structured document that can be generated can be generated.

Embodiment 6 FIG. The input document is typically
It is composed of chapters, sections, paragraphs, etc., and each has a logical meaning in the document structure. By identifying and managing these, more precise and highly-useful document management can be performed, and an output document rich in change can be obtained. Hereinafter, an example including means for analyzing the division of chapters, sections, etc. will be described. FIG. 24 is a block diagram of the structured document generation device according to the sixth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. A text analysis unit 12 receives the generated coded data, recognizes the document structure such as chapters, sections, and bullets, and sends the recognized result to the structured document generation unit 5.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 25 showing the details of step 17 thereof. The present embodiment is a case where the text analysis mode is ON in the flowchart of FIG. 3, and corresponds to, for example, a case where the partial area of area ID3 in the document image area / logical structure correspondence definition of FIG. 6 is processed. The control means 7 checks the text analysis mode (step 16), and proceeds to step 17. The operation of step 17 will be described with reference to the flowchart of FIG. Since the text analysis guideline in the area ID3 of the document image area / logical structure correspondence definition in FIG. 6 is "repeated structure extraction 1", the text analysis guideline indicated by 69 is read (step 71). With respect to the repeating structure, this guideline specifies the start of the repeating structure by the arrangement of “numbers” and “.” As in 54 of the document image in FIG. Express what you are doing. Regarding the correspondence with document tags, this guideline shows that a document tag called <list> is used for the entire repeating structure, and <list> is used for each repeating element.
Indicates that a document tag of item> is added.

A document structure is extracted according to the analysis guideline obtained in step 71, and "coded data", "document tag string", and "order" are recognized for each document structure (step).
72). For example, in the case of the second line of the document image 54 in FIG. 5, the coded data is “early”, the document tag sequence is “<catalog><text><feature><list><item>”, and the order is “ 2 ". However, here, the <list><item> specified by the document tag string 69 is linked to the <catalog><text><feature> specified in the third line of 64. Next, the triplet of (coded data, document tag sequence, order) obtained in step 72 is sent to step 13 (step
73). For example, in step 7, when the partial area of “area ID” 3 of 61 in FIG. 6 is extracted, step 7
18, the part 93 of FIG. 8 is added to the correspondence table of the document tag string / coded data. This portion corresponds to the logical structure element shown in 103 of the structured document FIG. The document image data is as shown in FIG.
When the logical structure correspondence definition is shown in FIG. 6, the logical structure having the repeating structure shown by 105 in the structured document FIG. 10 is generated by recognizing the region 54 in FIG. 5 as the repeating structure (step 23).

According to the above-mentioned configuration, it is difficult to recognize from the document image input by the document image data input means 1, but the information about the document structure that can be recognized from the coded information is recognized. By using it, a structured document with a finer grain can be generated.

Embodiment 7 FIG. As in the third and fourth embodiments, since the capacity of processing attributes is limited, a detailed processing table can be prepared separately and designated as an attribute to perform complicated attribute processing. For example, it is often the case that a character string in a specific area is desired to be converted into another character string for use in an output document. In the example below, the simplest conversion between era and AD is explained, but if you expand it, you can replace the input of the partial area with other information and output it. By doing so, the output can be performed in accordance with the software interface on the output side, and the diversion of the document data is enhanced. FIG. 26 is a configuration diagram of the structured document generation device according to the seventh embodiment. In the figure, the same or corresponding portions as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 13 is a character string conversion means for receiving the generated encoded data and converting a certain character string appearing in the data into another character string and sending the result to the structured document generation means 5.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 25 which is the detail of step 15 thereof. The present embodiment is a case where the character replacement mode is ON in the flowchart of FIG. 3, and corresponds to, for example, a case where the partial area of area ID1 in the document image area / logical structure correspondence definition of FIG. 6 is processed. The control means 7 checks the character replacement mode (step 14) and proceeds to step 15. s
The operation of step 15 will be described with reference to the flowchart of FIG. 67 of the document image area / logical structure correspondence definition in FIG.
The text analysis guideline shown in is read, and the part corresponding to the area ID1 is "AD conversion", so that the name of the filter "AD conversion" used as a conversion algorithm at the time of conversion is obtained (step 81).

Then, by replacing the coded data in the (document tag string, coded data) pair obtained by step 13 with the result passed through the filter "AD conversion", a new (document tag) (Column, coded data) is generated (step 82). For example, ste
When the partial area of "area ID" 1 of 61 in FIG. 6 is taken out in p7, the document image data of the corresponding partial area becomes the portion of the document image 60 of FIG. 5, so the character code "1995/9 "/ 25" is converted to "September 25, 1995", and (<catalog><date and time>, September 25, 1995)
Day) is generated. As a result, in step 18,
It is written as 91 in FIG. 8 in the correspondence table of the document tag string / coded data. This part corresponds to the structured document 10 in FIG.
2 corresponds to the logical structure element shown in FIG. According to the above configuration, it is possible to replace a certain character string existing in the document image data with the character string required by the structured document to be output.

Embodiment 8 FIG. In the eighth embodiment, in the second embodiment, the fluctuation with the standard partial area stored in the correspondence definition storage unit is absorbed when the document image data is input, whereas the plurality of stored standard partial areas are stored. An example of finding a matching partial area from the inside and performing the corresponding processing will be described. Even in this way, the number of registrations in the definition can be reduced. FIG. 28 is a configuration diagram of the structured document generation device according to the eighth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. 14
Receives the document image data input by the document image data input unit 1 and the coded data generated by the pattern recognition unit 4, and selects one document image region / logical structure correspondence from a plurality of document image region / logical structure correspondence definitions. This is an area / logical structure correspondence definition estimating means for selecting a definition and sending the image area / logical structure correspondence definition to the document image area cutout means 3.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 29 showing the details of step 3 thereof. The eighth embodiment is a case where the document image area / logical structure correspondence definition estimation mode is ON in the flowchart of FIG.
The control means 7 checks the document image area estimation mode (step 2) and proceeds to step 3. The operation of step 3 will be described with reference to the flowchart of FIG. step
After executing 31, the document image area / logical structure correspondence definition is sequentially fetched.
91). Further, the extracted document image area / logical structure correspondence definition will be referred to as R _k hereinafter.

Step32, step33, step3
4 is executed similarly to the flowchart of FIG. However, in step 32, when all the partial areas in the document image area / logical structure correspondence definition have been extracted, the process proceeds to step 92. The value (area of A _ij ) / ((area of Y _i ) + (area of Z _ij )) obtained in step 34 will be referred to as B _i hereinafter. In step 92, Σ {i
= j} {n} B _i is obtained, and this value is set as C _k . However, n is s
This is the number of partial areas in the document image area / logical structure correspondence definition specified in step 32. C _k is the maximum R _k
Is the estimated document image area / logical structure correspondence definition (step 93). That is, it is estimated that the cutout partial area is compared with so-called some standard document image areas that are registered and stored, and the one that maximizes the similarity C _k is the definition. In step 4, the estimated document image area / logical structure correspondence definition is read. According to the above configuration, it is possible to save the trouble of designating the image area / logical structure correspondence definition required when converting the structured document from the document image data.

Embodiment 9 In order to register in the database, the items, formats, codes, etc. specified for each database are prescribed, and if a data group is prepared in a form conforming to these, the database can be registered by an automatic registration program or the like. In the ninth embodiment, an example of obtaining an output document adapted to a desired database for such purpose will be described. FIG. 30 is a configuration diagram of the structured document generation device according to the ninth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 21 receives the structured document generated by the structured document generating means 5 and the script / logical structure correspondence definition stored in the script / logical structure correspondence definition storage unit 22, and the script / logical structure correspondence is made for each part of the structured document. It is an external procedure calling means for executing the instruction specified by the definition.

The operation of the structured document generating apparatus configured as described above will be described with reference to FIGS. 2 to 4 and the flowchart of FIG. 29 showing the details of step 26 thereof. In the present embodiment, the document image area / logical structure correspondence definition external call mode is ON in the flowchart of FIG. 4, and the control means 7 checks the document image area estimation mode (step 24) and proceeds to step 26. . ste
The operation of p26 will be described with reference to the flowchart of FIG.

FIG. 32 is an example diagram showing an example of a document image area / logical structure correspondence definition, and describes a script that defines automatic registration of a structured document in a database. 251 defines that the script type is “DB registration”. 252 and 253 are step 10
2. In step 103, it is referred to when determining which part of the structured document should be extracted as the value of the DB attribute.

The document image area / logical structure correspondence definition is read, and the external procedure to be called is determined (step 10).
1). For example, the items and formats of the data requested by the external database side that wants to register the data are shown in FIG.
252 attributes. When the document image area / logical structure correspondence definition in FIG. 33 is read, the external procedure is confirmed as “DB registration (register in database)” by referring to 251. Next, the argument values required by the external procedure are extracted from the structured document (step 1).
02). In the case of the document image area / logical structure correspondence definition in FIG. 33, the portion indicated by the document tag having the same area ID as the value of the DB attribute designated by 252 is extracted from the structured document. That is, in the case of the structured document of FIG. 10, for example, as the value of the DB attribute “figure” indicated by 253, “picture.img” which is the part indicated by the document tag <catalog><text><appearance> is extracted. .

Next, an external procedure is executed based on the value obtained in step 102 (step 103). FIG.
In the case of the document image area / logical structure correspondence definition of 2 and the structured document of FIG. 10, the value of the DB attribute “title” is “rice cooker” and the value of the DB attribute “feature” is “(delicious, fast, bargain. ) ”, The value of DB attribute“ figure ”is“ (delicious, fast, bargain) ”, and the value of“ figure ”is“ (pict
ure.img) ”is registered in the database as a document.

According to the above-mentioned structure, it becomes possible to automatically register the document in the database by using the structure information of the structured document. Although the above embodiment relates to automatic registration of documents in the database, the present invention is not limited to this, and automatic generation / automatic shipment of business documents such as automatic shipment of order slips and automatic calculation of spreadsheet data. It can be applied to linkage with various application software such as linkage with spreadsheet software such as generation.

Embodiment 10 FIG. In the first embodiment, the operation has been described assuming that the document tag as the logical structure is already added to the input document image data for each document image area as the predetermined portion. In this embodiment, the registration of the correspondence between the document image area and the logical structure, that is, the range specifying operation of the area of the first document tag will be described. FIG. 33 is a configuration diagram of the structured document generation device according to the tenth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. 31
Accepts the correspondence between the “partial area of the document image data” specified by the user and the “markup tag (document tag) that represents a specific part of the logical structure of the document”, A document image area / logical structure associating unit that generates a logical structure correspondence definition and stores it in the document image area / logical structure correspondence definition storage unit.

FIG. 35 shows an interface for controlling the association between the "partial area of the document image data" and the "markup tag (document tag) expressing a specific portion of the logical structure of the document". This is an example. "Partial area of document image data" is automatically created "area ID" 2
The "markup tag that represents a certain portion of the logical structure of the document" specified by the registration unit 71 is a "document tag string" that is a sequence of document tags from the beginning to the end. 273. For the position and size of the "partial area of the document image data" in the document image data, the "partial area coordinates" 272 specify the X and Y coordinates of the start point and the end point. When the association is completed, it is confirmed by pressing the button 274. If the association operation is incorrect, the button 275 cancels the association operation.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, display the partial area definition menu (step
p261). The partial area definition menu is an interface that allows the user to specify the definition part related to the partial area definition in the document image area / logical structure correspondence definition. FIG.
Is an example of a partial area definition menu, and includes an “area ID”, a “start point and an end point of a partial area” specified for each partial area
There is an item "document tag", and the user fills in this item to complete the partial area definition. Next, if the designation of the partial area is continued, it is regarded as step 263. If it is not continued, it is considered that the partial area definition is completed, and the partial area definition menu screen is closed to end (step 2).
62). In step 263, the area ID is given to the document area to be designated. As the area ID, for example,
It suffices to use a number indicating which order the user has designated the partial area. This area ID is used to identify the partial area designated by the user.

Next, the user specifies the position information of the partial area to be specified by filling in the menu item (step 264). For example, in the partial area definition menu shown in FIG. 35, a rectangular area having the start point coordinates and end point coordinates 272 as the upper left vertex and the lower right vertex is designated. Next, the user specifies the document tag string to be added to the partial area specified in step 264 (step 26).
5). For example, in the partial area definition menu shown in FIG. 35, the document tag string <catalog><name> shown at 273 is designated. Finally, the "region ID given in step 263", "position information of the partial region designated in step 264", and "the document tag string designated in step 265" are added to the document image region / logical structure correspondence definition (step 266). ). For example, in the case of designation as in the partial area definition menu shown in FIG. 35, a part having an area ID of 2 in the document image area / logical structure correspondence definition in FIG. 6 is added. According to the above configuration, the user can specify the document image area / logical structure correspondence definition.

Embodiment 11 FIG. In the present embodiment, following the range designation of the partial area described in the previous embodiment, the registration of the attribute that defines the processing when outputting the document tag will be described. FIG. 36 is a configuration diagram of the structured document generation device according to the eleventh embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. 32 receives attribute information relating to a pattern to be recognized designated by the user, adds the information to the document image area / logical structure correspondence definition, and the resulting document image area / logical structure correspondence definition is the document image area / logic. It is a recognition attribute designating unit that is stored in the structure correspondence definition storage unit.

FIG. 38 shows an example in which a setting item for character recognition of the "partial area of document image data" is added to the interface of FIG. "Cutout mode" 297
Is used to instruct whether to cut out a partial area of the document image data as a simple image area without applying pattern recognition processing. "Text Analysis Guidelines" 296
Is used to instruct what text analysis is to be performed as a result of pattern recognition processing performed on a partial area of document image data. The “character replacement guideline” 295 is used to instruct what character replacement processing is to be performed on the result of pattern recognition processing performed on the partial area of the document image data. The replacement process includes, for example, replacement with a normalized format such as "American era / quantity unit / address / name".
"Recognition attribute" 294 is a pattern such as "character / image", "handwriting / printing", "alphabet / number / symbol / katakana / hiragana / kanji / mixture thereof" in the partial area of the document image data. It is used to indicate which belongs to. In this way, by further designating and storing the type of processing, it is possible to cope with changes in the output document.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, display the partial area definition menu (step
p281). In the partial area definition menu of FIG. 38,
It has the items "area ID" specified for each partial area, "start and end points of partial area", "document tag", "cutout mode", "character replacement guideline", "text analysis guideline", and "recognition attribute". The user completes this item to complete the partial area definition. Next, if the designation of the partial area is continued, it is considered to be step 263. If it is not continued, it is considered that the partial area definition is completed, and the partial area definition menu screen is closed to end (step 262). st
At ep263 and step 264, the same processing as that of the tenth embodiment is performed.

Next, the user specifies the document tag, the character replacement guideline, the text analysis guideline, and the recognition attribute for the partial area that is currently specified (step 282). Figure 38
In the partial area definition menu shown in, the document tag string is 2
<Catalog><name> shown in 73, “none” shown in 297 as a cutout mode, and 29 as a character replacement guideline.
5, “None” shown as a text analysis guideline, “None” shown as a text analysis guideline 296, and “Type /
"Kanji" is specified for each. step263, step
H.264, "area ID" specified in step 282,
"Position information of partial area", "Document tag string", "Cutout mode", "Character replacement guideline", "Text analysis guideline",
A "recognition attribute" is added (step 283). For example,
In the case of designation as in the partial area definition menu shown in FIG. 38, a part having an area ID of 2 in the document image area / logical structure correspondence definition in FIG. 6 is added. According to the above configuration, the pattern information of the document image data is recognized, and the attribute information designating the processing to be performed when outputting the recognized partial area is transmitted to the structured document generation means, and each is required by the software in the subsequent process. Can be converted into a form suitable for the output.

Embodiment 12 FIG. In the tenth embodiment, the example in which the partial area designation of the document image data is set by the XY coordinate values has been shown. In the present embodiment, in order to further improve the operability, an example will be described in which the document image data itself is displayed on the screen, and the range of the area is directly designated on the screen with a cursor, a mouse, or the like. FIG. 39 is a configuration diagram of the structured document generation device according to the twelfth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 33 is a document image data display means for displaying the image data read from the image reading device as document image data. 31 is the document image data display means 3
When a partial area in the display data displayed by 3 is designated by the user, the designated area is recognized, information regarding the area is added to the document image area / logical structure correspondence definition, and the resultant document is obtained. The document image area / logical structure correspondence definition means stores the image area / logical structure correspondence definition in the document image area / logical structure correspondence definition storage unit 2.

In FIG. 41, "document image data" 311 is displayed as a background, and "document image data partial area" is displayed on the background.
This is an example of an interface for creating / changing 312 using a “pointing device” 313. At the same time that the document image data 311 is displayed on the screen, the correspondence definition storage unit 2
The document tag portion of the corresponding ID is called and displayed. Then, the “partial area of the document image data” 31 created / changed using the “pointing device” 313 31
The value of the “partial area coordinates” 272 of the correspondence definition storage unit 2 is changed in association with the position and size of 2 and vice versa.
When the value of the “partial area coordinates” 272 is changed, the position and size of the “partial area of the document image data” 312 are correspondingly changed.

The operation of the structured document generation apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image data read by the document image data input means 1 is displayed by the document image data display means (step 341). Next, the step of the tenth embodiment
The same processing as 261, step 262, and step 263 (adding a document tag) is performed. Next, the user specifies a partial area in the document image displayed in step 301 with the pointing device (step 302). For example, “rice cooker” in the document image data 311 shown in FIG. 41.
When designating a partial area written as, 31 in FIG.
The upper left and lower right vertices of the rectangular area shown in 2 are designated by the pointing device 313.

Next, the coordinates of the partial area designated in step 302 are confirmed and written in the partial area definition menu (step 346). For example, when the partial area 312 in FIG. 41 is designated, as shown in 272 in FIG.
After determining the upper left (start point) coordinates (30, 50) and the lower right (end point) coordinates (160, 80) of the area, write as 272 in the partial area definition menu shown in FIG. Finally, step 266 of the tenth embodiment
By performing the same processing as in step 263, s
The information specified in step 302, step 303, and step 265 is added to the document image area / logical structure correspondence definition. According to the above configuration, the user can specify the area while referring to the document image data, so that the load required by the user to specify the area can be reduced.

Thirteenth Embodiment In the above embodiment, an example of designating the range of the partial area of the document image data by displaying the actual input data on the screen has been described.
An example will be described in which the same function is provided for the designation of the repeating area in 4 and the range designation of the nested structure. FIG.
22 is a configuration diagram of the structured document generation device according to the thirteenth embodiment. In the figure, the same or corresponding portions as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. When the user designates an example of forming a repeating structure with respect to the image data displayed by the document image data display means 33, the document recognition unit 34 recognizes the specified repeating structure and provides information about the repeating structure to the document. The document image area / logical structure correspondence definition is added to the image area / logical structure correspondence definition, and the resulting document image area / logical structure correspondence definition is stored in the document image area / logical structure correspondence definition storage unit 2.
It is a repeating structure designating means for storing in.

FIG. 44 is an example of an interface for setting a repeating pattern of the “partial area of the document image data” designated as having repeating in “presence / absence of repetition” 296. In the “repeated structure extraction” 333, the parent tag associated with the “partial area of the document image data” that is a collection of the characteristics of the repeated start point and the entire “partial area of the document image data” that is repeated and the repeated “of the document image data” The repeat tag associated with each of the “partial areas” is specified. This part specifies the information referred to by the text analysis means.

"Area ID 1" 334 and "Area ID"
“2” 335 designates a document tag associated with each of the two “partial areas of the document image data”. The “relationship between areas” 337 includes “area ID 1” 334 and “area ID”.
2 ”335 specifies the positional relationship of each“ partial area of the document image data ”. “Repeat direction” 338 specifies the direction in which “region ID1” 334 and “region ID2” 335 are repeated. The “parent tag” 336 specifies to which parent tag the repeated “partial area of the document image data” belongs. This part specifies the information referred to by the repeating structure recognition means.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, similar to step 301 of the twelfth embodiment,
It is displayed by the image data display means. Next, a repeating structure definition menu for designating conditions regarding the repeating structure in the partial area in the document image data is displayed (st
ep321). By the repeating structure definition menu of FIG. 44, “shape”, “document tag to be added”, “positional relationship between constituents”, and “how to repeat constituents” of constituents that appear repeatedly are specified. Next, the user designates the shape of the repeatedly appearing component by pointing to the partial area of the image data displayed by step 301 (step 322). For example, when the partial area 331 of the image data in the upper part of FIG. 44 is designated, it is recognized as the partial area of the area ID1, and its shape is the horizontal length 2
Numerical values of 0 and vertical length 10 are entered in the repeating structure definition menu. Similarly, the partial area 332 is recognized as the partial area of the area ID2.

Next, the document tag to be given to the constituent element designated in step 322 is designated (step 3).
23). For example, if a document tag of <attribute> is to be added to the area 331 in the image data, 334
Specify as in. Similarly, when it is desired to add a document tag of <value> to the area 332 in the image data, it is designated as in 335. Also, 33
When it is desired to add a document tag <table> to a block in which rectangular areas obtained by arranging 1 and 332 horizontally appear repeatedly, it is specified as in 336. Next, the user designates the positional relationship between the constituent elements designated in step 322 (step 324). For example, if it is desired to specify that 331 and 332 in the image data are arranged side by side, the inter-region relation 337 is set horizontally. Next, the user instructs how to repeat a group of regions that satisfy the positional relationship between the constituents designated in step 324 (step 325). For example, in FIG.
If you want to instruct that the rectangular area created by arranging 331 and 332 in the horizontal direction is repeated in the vertical direction,
The repeating direction of 38 is set to vertical.

Finally, the shape of the component set by step 322, the document tag given to the component set by step 323, the positional relationship between the components set by step 324, and the repetition of the component set by step 325. The direction is written in the document image area / logical structure correspondence definition (step 326). For example, when the setting as shown in FIG. 44 is made, the portions 70 and 71 in the document image area / logical structure correspondence definition in FIG. 6 are written. In this way, it is possible to freely set a complicated process that does not fit in the normal attribute assignment section and assign it to the document tag (area ID). According to the above-mentioned configuration, the user can specify that the region includes the repeating structure by only showing a part of the repeating structure, and thus the load required for the user to specify the region can be reduced.

Fourteenth Embodiment In the above embodiment, the operation has been described on the assumption that the output format of the output document is specified, that is, the document type is defined and registered in the document type definition section. In this embodiment, the designation of the output format of this output document will be described. For this purpose, it must be possible to display the logical structure of the document in a form that is easy to understand, such as a tree structure. This is shown below. FIG.
It is a block diagram of the structured document generation device of the fourteenth embodiment,
In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. Reference numeral 36 denotes a document logical structure display / correction means for displaying the logical structure of a document as a tree structure by referring to the document type definition in the document type definition storage means 35. Reference numeral 31 denotes a node in the tree structure by the user. When specified, the document tag corresponding to the node is determined, information about the tag is added to the document image area / logical structure correspondence definition, and the resulting document image area / logical structure correspondence definition is added to the document image area / A document image area / logical structure associating unit to be stored in the logical structure correspondence definition storage unit.

FIG. 47 shows an example of an interface for displaying the “document logical structure” in the form of “tree” and designating the “specific part of the document logical structure” 351 using the “pointing device” 313. The "document tag string" 273 is changed according to the "specific part having the logical structure of the document" specified by using the "pointing device" 313, and as a result, "the specific part having the logical structure of the document" is expressed. The document tag for markup "can be obtained.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image data read by the document image data input means 1 is displayed by the document image data display means (step 341). For example, the image data of FIG. 5 is displayed as the document image data. Next, a tree structure expressing the parent-child relationship of the document tags defined by the document type definition specified by the user is displayed (step 342). For example, if the user
When the document type definition shown in is designated, the tree structure in the right part of FIG. 47 is displayed. For the parent-child relationship of document tags,
This is as described in the description of FIG. 7 of the embodiment. next,
If the user continues to specify the partial area in the document image, the process ends at step 344. If not, the process ends there.

At step 344, the user designates the partial area in the document image displayed at step 341 with the pointing device. For example, in the case of designating a partial area written as "rice cooker" in the document image data of FIG. 5, the rectangular area shown in the document screen data 312 shown on the left of FIG. 47 is pointed at the upper left and lower right vertices. Device 3
Specify by 13. Next, an area ID is given to the designated document area (step 345). As the area ID, for example, a number indicating how many times the user has designated the partial area may be used. This area ID is used to identify the partial area designated by the user. next,
The coordinates of the partial area designated in step 344 are confirmed (step 346). For example, when the partial area 312 of FIG. 47 is designated, the upper left (start point) coordinates (30, 50) and the lower right (end point) coordinates of that area are (16).
0,80) is obtained. Next, the user designates a node in the tree structure displayed in step 342 with a pointing device (step 347). For example, when the document tag “name” of the document type definition shown in FIG. 7 is designated, the node 351 in the tree structure of FIG. 47 is designated using the pointing device 353.

Next, the document tag string is decided from the node designated in step 347 (step 348). The document tag string is an array of document tags corresponding to the nodes existing on the path from the node located at the top of the displayed tree to the designated node. For example, in FIG.
When the node “name” of node 351 is designated, the document tag string is <catalog><name> as indicated by 273. When the node “attribute” of 352 in FIG. 47 is designated, the document tag string is <catalog><text><specification><table><attribute>. Next, the "area ID" obtained in step 345, the "coordinates of partial areas" obtained in step 346, and the "document tag string" obtained in step 348.
Is added to the document image area / logical structure correspondence definition, and step
Return to step 343 (step 349). For example, when the rectangular area shown in the partial area 312 of FIG. 47 is designated second in the document image and the node 351 in the tree structure of FIG. 47 is designated, the document image area / logic shown in FIG. The start point of the "area ID" 2 shown in the 3rd line of 61 in the structure correspondence definition,
Portions related to the end point and document tag are added. According to the above configuration, the user can specify the tag while referring to the logical structure of the document, so that the load required for the user to specify the tag can be reduced.

Fifteenth Embodiment In designating the output format of the output document, it is important to intuitively know and understand the correspondence between the area of the input document image data and the ID or the document tag, that is, the decomposition unit. To this end, it is important to display the correspondence between the range of the input partial area and the document tag, especially the tree structure display. This will be described in the present embodiment. 48 is a configuration diagram of the structured document generation device according to the fifteenth embodiment. In the figure, the same or corresponding parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. 37 refers to the document image area / logical structure correspondence definition to determine the node corresponding to the tag referred to in the document image / logical structure correspondence definition, and emphasizes that node in the displayed tree structure. This is a document image area / logical structure correspondence display means that is displayed by being displayed.

FIG. 49 shows the "logic of document" associated with each "specific part of the logic structure of document" of the "logic structure of document" shown in FIG. 47 in the form of "tree". This is an example in which a "specific portion having a structure" 371 and a "specific portion having a logical structure of a document" 372 that is not associated are displayed so as to be visually distinguishable. Here, the associated "specific portion having the logical structure of the document" is the terminal "specific portion having the logical structure of the document" or the lower order that is associated with the "partial area of the document image data". Of all the required "specific parts of the logical structure of the document" are associated and one of the subordinate "specific parts of the logical structure of the document" is associated "A specific part having a logical structure of a document" other than "a specific part having a logical structure of a document" that is not associated with "a specific part having a logical structure of a document" It is.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image area / logical structure correspondence definition is read (step 371). Next, the step of the fourteenth embodiment
The same processing as p342 is performed. In step 372,
The partial areas are sequentially taken out from the document image area / logical structure correspondence definition read in step 371, and if the partial areas can be taken out, the processing ends at step 373, and if the extraction is completed, the processing ends. s
In step 373, the document tag is extracted for each of the extracted partial areas (step 373). For example, in the case of the document image area / logical structure correspondence definition shown in FIG. 51, the document tag string <catalog><date / time> in the partial area with the area ID 1 and <catalog><in the partial area with the area ID 3 A document tag string "text><appearance>" is extracted.

Next, the road in the tree structure corresponding to the document tag sequence obtained in step 372 is highlighted and displayed, and
Return to p372 (step 374). For example, step
When the document tag <catalog><date / time> is obtained by 373, the portion 371 in FIG. 49 is emphasized.
After the processing of step 374 for all the partial areas of the document image area / logical structure correspondence definition in FIG. 51 is completed,
The tree structure shown in FIG. 49 is displayed. <Features> in Figure 49
The reason why the road starting from is blank is that the document tag sequence <catalog><text><feature><list><item> does not appear in the document image area / logical structure correspondence definition in FIG. 51. In addition, the body node in FIG. 49 is white, which is a document tag subordinate to <body>, <feature>,
This is because all the document tags of <Appearance> and <Specification> have not appeared. According to the above configuration, the user
Since it is possible to easily know whether or not the tag in the document structure is already referenced from the area definition file, it is possible to reduce the load required for the user to create / correct the document image area / logical structure correspondence definition.

Sixteenth Embodiment In the above embodiment, only the correspondence between the document tag and the tree structure logical display is shown, and the intuitive display of the partial area range and the tree structure display is not shown yet. In the present embodiment, this correspondence display will be described. The device configuration in this embodiment is shown in FIG. 48, which is the same as the previous embodiment.

In FIG. 52, "document image data" 311 is displayed as a background, and one of a plurality of "partial areas of document image data" displayed thereon is displayed by "pointing device" 313. When the “partial area of the document image data to be noted” 381 is set, it is associated with the “partial area of the document image data to be noted” 381 in the “logical structure of the document” displayed in the form of “tree”. This is an example of visually displaying "a specific portion having a logical structure of a document" as "a specific portion having a logical structure of a document of interest" 382 and visually distinguishing it from other "specific portions having a logical structure of a document".

The operation of the structured document generation apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image area / logical structure correspondence definition is read (step 381). Next, the step of the fourteenth embodiment
The same processing as p341 and step 342 is performed. Next, the user designates a point in the document image data displayed in step 341 with the pointing device (step 382). Next, a partial area including the position designated by step 382 is obtained from the document image area / logical structure correspondence definition (step 383). For example, s
In the document image data 311 displayed by step 341, when the pointing device 313 in FIG. 52 points to the position in the figure, the document image area / logical structure correspondence definition in FIG. 6 is defined as an area including the point. Area ID
A partial area whose value is 2 is obtained.

Next, the document tag string attached to the partial area obtained in step 383 is obtained (step 384).
For example, when a partial area having an area ID of 2 is obtained by step 343, the document image area shown in FIG.
The document tag string <catalog><name> assigned to the partial area having the area ID 2 in the logical structure correspondence definition is obtained. Finally,
The node at the end of the road in the tree structure corresponding to the document tag sequence obtained in step 384 is emphasized and displayed (step
385). For example, when the document tags <catalog><name> are obtained by step 384, step
382 in the tree structure on the right of FIG. 52 displayed by 342
The node of is highlighted and displayed. According to the above configuration, the user can easily know which tag in the document structure is associated with the area in the document image data. Therefore, the user's document image area / logical structure correspondence definition / The load required for correction can be reduced.

Embodiment 17 FIG. In the above embodiment, an example in which the corresponding portion of the tree-displayed logical structure is identified and displayed when the partial area is designated has been described. However, in the present embodiment, conversely, when the specified portion of the tree-displayed logical structure is designated. An example in which the corresponding partial area of the document image data is identified and displayed will be described. The device configuration in the present embodiment is also shown in FIG. 48, which is the same as the previous embodiment.

In FIG. 54, contrary to FIG. 52, one of the “specific parts of the logical structure of the document” in the “logical structure of the document” displayed in the form of “tree” is designated as the “pointing device”.
When the “specific portion having the logical structure of the document to be noticed” 391 is made using 313, the “partial region of the document image data” of the “document image data” 311 in the background The "partial area of the document image data" associated with the "specific part having a logical structure" 391 is visually separated from other "partial areas of the document image data" as "partial area of the document image data to be noticed" 392. This is an example of displaying separately.

The operation of the structured document generation apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image area / logical structure correspondence definition is read (step 381). Next, the step of the fourteenth embodiment
The same process as p341, step 342, and step 343 is performed. Next, from the document image area / logical structure correspondence definition, a partial area having a document tag string corresponding to the way to the node designated by step 343 is obtained (step
391). For example, when the node 391 in the tree structure on the right side of FIG. 54 displayed by step 342 is specified, the document tag string <catalog> corresponding to the way to the node
As the document area having <date and time>, the partial area having the area ID 1 in the document image area / logical structure correspondence definition shown in FIG. 6 is obtained. Finally, in the document image data, step3
The partial area obtained in step 91 is displayed with emphasis (step
p392). For example, the partial area having the area ID 1 in the document image area / logical structure correspondence definition shown in FIG.
When it is obtained by 1, the portion of the rectangular area portion 392 having the start point (20, 10) and the end point (160, 40) is emphasized and displayed. According to the above configuration, the user can easily know which tag in the document structure is associated with which region in the document image data. Therefore, the user's document image region / logical structure It is possible to reduce the load required to create / correct the correspondence definition.

Eighteenth Embodiment An example in which it is possible to easily check whether or not the output document is a structured document in which necessary and sufficient information is described will be described. FIG. 56 is a configuration diagram of the structured document generation device according to the eighteenth embodiment. In the figure, the same or corresponding portions as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. 38 determines whether the designated document area / logical structure correspondence definition satisfies the document type definition. Next, when the condition is not satisfied, the document type definition correctness confirmation means for confirming the tag that causes the tag and displaying the node in the tree structure corresponding to the tag by emphasizing it.

FIG. 58 analyzes whether or not the generated structured document is valid, and when an invalid portion is found, the document tag of the logical structure of the document corresponding to the invalid portion is displayed in an easy-to-understand manner. This is an example of a user interface that displays an unreasonable reason.

The operation of the structured document generating apparatus configured as described above will be described with reference to the flowchart of FIG.
First, the document image area / logical structure correspondence definition is read (step 371). Next, the step of the fourteenth embodiment
The same processing as p342 is performed. Next, find a document tag that is defined in the document type definition but not in the document image area / logical structure correspondence definition (step 40).
1). For example, when the document type definition shown in FIG. 7 is specified in step 342 and the document image area / logical structure correspondence definition shown in FIG. 51 is read in step 371, it exists in the document type definition in FIG. As a document tag string not defined in the document image area / logical structure correspondence definition, <
Catalog><text><features> can be found. Next, ste
After highlighting the node at the end of the road in the tree structure corresponding to the document tag string found in p401 (step 4
02), the reason for inconsistency between the document type definition and the document image area / logical structure correspondence definition is displayed (step 403). For example, if the document tag string <catalog><text><feature> is found in step 401, the document tag string 411 in FIG.
The node indicated by is highlighted, and the reason for the inconsistency indicated by 412 is displayed. According to the above configuration, the user can easily know whether the document area / logical structure correspondence definition is appropriate, and if not, which part is the problem, so that the user's document It is possible to reduce the load required for creating / correcting the image area / logical structure correspondence definition.

[0110]

As described above, according to the present invention, the correspondence definition storing means for storing the document image area as the definition in the logical structure of the document, and the document image area cutting means for cutting out the partial area of the document image, Document type definition storing means for storing a document as a logical structure, which is a collection of necessary areas, and structured document generating means for generating a structured document by performing attribute processing in a document tag based on a specified document type Therefore, it is possible to freely set and select the format of the structured document to be output that conforms to the interface required by the program in the latter stage, and it is possible to effectively use the standard format information in the image document data. There is an effect that the diversion property and the possibility of automatic processing such as database registration are improved.

Furthermore, since the document image area estimating means is added, even if the format format of the read document image data is slightly different from the prescribed format format in the correspondence definition storing means, The correspondence of the logical structure can be correctly recognized, the recognition tolerance is expanded, the stability of the device is increased, and the labor required for registering the document format is reduced.

Furthermore, since the detailed processing table for another processing attribute is provided, it is possible to perform complicated processing according to a required output document and output, and it is possible to freely change the processing content. It is effective for fine-grained output document management and simplification of the management.

Furthermore, since image cut-out means for cutting out a partial area of the document image data and outputting it as an image file is added, if there is a mixture of data that is difficult to convert into coded data such as figures and photographs. However, there is an effect that a structured document can be freely generated.

Furthermore, since the input character string is converted into another character string, a structured document having a format that can be easily processed by various kinds of application software on the output side can be obtained. It has the effect of increasing the possibility.

Furthermore, a document image area in which a partial area of the document image data is decomposed into units suitable for the output side application software and a document tag is added for registration / storage.
Since the logical structure associating means is provided, the user can freely define the correspondence between the document image area and the logical structure. By registering the format of the document image data that meets the user request, the degree of freedom in structured document management is increased.

Further, even when it is desired to perform complicated processing on a partial area of the document image data, it is possible to independently specify a free processing that meets the user's request, which has the effect of improving the utilization of the document data. .

Furthermore, since the identification display corresponding to the partial area of the document image data and the document tag can be performed, the user can specify the area while referring to the document image data, and the load of area specification can be reduced. , It is effective in reducing the load of structured document management.

Further, since the document logical structure display / correction means for displaying the logical structure of the document defined by the document type definition as a tree structure is provided, the user can specify the document tag while referring to the logical structure of the document. It is possible to reduce the load of tag specification. Thus, there is an effect that the load of structured document management can be reduced.

Furthermore, since the document type definition is shown by a tree structure and the correspondence between the partial area of the document image data and the document tag is stored, the user can already find the tag in the document structure from the area definition file. It is possible to easily know whether or not it is referenced, and it is possible to reduce the load required for the user to create / correct the document image area / logical structure correspondence definition. That is, there is an effect that the load of structured document management can be reduced.

Furthermore, since it is possible to identify and display the document type definition represented by the tree structure, the partial area of the document image data, and the document tag, the user intuitively knows the relationship between them and configures them. This has the effect of reducing the load of document management.

Furthermore, when there is no correspondence between the tree structure display, the partial area and the document tag, the fact is displayed so that the user can know the suitability or unsuitability of the document area / logical structure correspondence definition. It is possible to reduce the load of structured document management.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a structured document generation device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an example of processing operation of the structured document generation device of the present invention.

FIG. 3 is a flowchart showing an example of processing operation of the structured document generation device of the present invention.

FIG. 4 is a flowchart showing an example of processing operation of the structured document generation device of the present invention.

FIG. 5 is a diagram showing an example of document image data in the structured document generation device of the present invention.

FIG. 6 is a diagram showing an example of a document image area / logical structure correspondence definition of the structured document generation device of the present invention.

FIG. 7 is a diagram showing an example of a document type definition in the structured document generation device of the present invention.

FIG. 8 is a diagram showing an example of a document tag / coded data correspondence table in the structured document generation device of the present invention.

9 is a diagram showing a parent-child relationship of the document tags of FIG. 7 in a tree structure.

FIG. 10 is a diagram showing an example of a structured document output by the structured document generation device of the present invention.

FIG. 11 is a diagram showing another example of a document image area / logical structure correspondence definition.

FIG. 12 is a diagram showing another example of a document type definition.

FIG. 13 is a diagram showing another example of a structured document.

FIG. 14 is a diagram illustrating an example of an intermediate structured document that is intermediately generated.

FIG. 15 is a diagram showing a configuration of a structured document generation device according to a second embodiment of the present invention.

FIG. 16 is a flowchart showing a processing operation of the structured document generation device according to the second embodiment of the present invention.

FIG. 17 is a diagram showing a structure of a structured document generation device according to a third embodiment of the present invention.

FIG. 18 is a flowchart showing a processing operation of the structured document generation device according to the third embodiment of the present invention.

FIG. 19 is a diagram showing a structure of a structured document generation device according to a fourth embodiment of the present invention.

FIG. 20 is a flowchart showing a processing operation of the structured document generation device according to the fourth embodiment of the present invention.

FIG. 21 is a diagram showing an example of a document image area / logical structure correspondence definition of the structured document generation device according to the fourth embodiment of the present invention.

FIG. 22 is a diagram showing the structure of a structured document generation device according to a fifth embodiment of the present invention.

FIG. 23 is a flowchart showing a processing operation of the structured document generation device according to the fifth embodiment of the present invention.

FIG. 24 is a diagram showing a structure of a structured document generation device according to a sixth embodiment of the present invention.

FIG. 25 is a flowchart showing a processing operation of the structured document generation device according to the sixth embodiment of the present invention.

FIG. 26 is a diagram showing the structure of a structured document generation device according to a seventh embodiment of the present invention.

FIG. 27 is a flowchart showing a processing operation of the structured document generation device according to the seventh embodiment of the present invention.

FIG. 28 is a diagram showing a structure of a structured document generation device according to an eighth embodiment of the present invention.

FIG. 29 is a flowchart showing a processing operation of the structured document generation device according to the eighth embodiment of the present invention.

FIG. 30 is a diagram showing a structure of a structured document generation device according to a ninth embodiment of the present invention.

FIG. 31 is a flowchart showing a processing operation of the structured document generation device according to the ninth embodiment of the present invention.

FIG. 32 is an example diagram showing an example of a document image area / logical structure correspondence definition of the structured document generation device according to the ninth embodiment of the present invention.

FIG. 33 is a diagram showing the structure of a structured document generation device according to a tenth embodiment of the present invention.

FIG. 34 is a flowchart showing a processing operation of the structured document generation device according to the tenth embodiment of the present invention.

FIG. 35 is a diagram showing an example of a document image area / logical structure association interface of the structured document generation device in the tenth embodiment of the present invention.

FIG. 36 is a diagram showing a structure of a structured document generation device according to an eleventh embodiment of the present invention.

FIG. 37 is a flow chart diagram showing the processing operation of the structured document generation device in the eleventh embodiment of the present invention.

FIG. 38 is a diagram showing an example of a recognition attribute designation interface of the structured document generation device in the eleventh embodiment of the present invention.

FIG. 39 is a diagram showing the structure of a structured document generation device according to a twelfth embodiment of the present invention.

FIG. 40 is a flow chart diagram showing the processing operation of the structured document generation device according to the twelfth embodiment of the present invention.

FIG. 41 is an example diagram showing an example of a document image data display interface of a structured document generation device according to a twelfth embodiment of the present invention.

FIG. 42 is a diagram showing the structure of a structured document generation device according to a thirteenth embodiment of the present invention.

FIG. 43 is a flowchart showing a processing operation of the structured document generation device according to the thirteenth embodiment of the present invention.

FIG. 44 is a diagram showing an example of a repeated structure designation interface of the structured document generation device in the thirteenth embodiment of the present invention.

FIG. 45 is a diagram showing a structure of a structured document generation device according to a fourteenth embodiment of the present invention.

FIG. 46 is a flowchart showing a processing operation of the structured document generation device according to the fourteenth embodiment of the present invention.

FIG. 47 is a diagram showing an example of a document logical structure display interface of the structured document generation device in the fourteenth embodiment of the present invention.

FIG. 48 is a diagram showing the structure of the structured document generation device according to the fifteenth, sixteenth and seventeenth embodiments of the present invention.

FIG. 49 is a diagram showing an example of a document area / document logical structure correspondence display interface of the structured document generation device in the fifteenth embodiment of the present invention.

FIG. 50 is a flowchart showing a processing operation of the structured document generation device according to the fifteenth embodiment of the present invention.

FIG. 51 is a diagram showing an example of a document image area / logical structure correspondence definition of the structured document generation device in the fifteenth embodiment of the present invention.

FIG. 52 is a diagram showing an example of a document area / document logical structure correspondence display interface of the structured document generation device in the sixteenth embodiment of the present invention.

FIG. 53 is a flowchart showing a processing operation of the structured document generation device according to the seventeenth embodiment of the present invention.

FIG. 54 is a diagram showing an example of a document area / document logical structure correspondence display interface of the structured document generation device in the seventeenth embodiment of the present invention.

FIG. 55 is a flowchart showing a processing operation of the structured document generation device according to the seventeenth embodiment of the present invention.

FIG. 56 is a diagram showing the structure of the structured document generation device according to the eighteenth embodiment of the present invention.

FIG. 57 is a flow chart diagram showing the processing operation of the structured document generation device in the eighteenth embodiment of the present invention.

FIG. 58 is a diagram showing an example of a document type definition validity confirmation interface of the structured document generation device in the eighteenth embodiment of the present invention.

FIG. 59 is a configuration diagram showing a conventional structured document generation device.

[Explanation of symbols]

1 document image data input means, 2 document image area / logical structure correspondence definition storage section, 3 document image area cutout means,
4 pattern recognition means, 5 structured document generation means, 6
Structured document output means, 7 control means, 8 document image area estimation means, 9 repeated structure recognition means, 10 nested structure recognition means, 11 image cutout means, 12 text analysis means, 13 character string conversion means, 14 area / logical structure Correspondence definition estimating means, 21 external procedure calling means, 22 script / logical structure correspondence definition storage section, 31
Document image area / logical structure associating unit, 32 recognition attribute designating unit, 33 document image data displaying unit, 34 repeating structure designating unit, 35 document type definition storing unit, 36
Document logical structure display means, 37 document image area / logical structure correspondence display means, 38 document type definition correctness confirmation means.

Claims (12)

[Claims] 1. A document tag corresponding to a specific logical structure portion of a document is added to a document image area, which is a predetermined portion of input document image data, and is necessary for an output document. A document image area / logical structure correspondence definition storage means (hereinafter referred to as correspondence definition storage means) that stores a processing attribute as a definition and a document image area in the correspondence definition storage means for the pattern-recognized input. A document image area cutting-out unit that cuts out a predetermined portion using information on a document, and a logical structure that describes a collection of the document image areas, which are the above-mentioned portions necessary for various output documents, is named and stored as a definition for each document. When the document type definition storage means and the specified output document are designated, the definition name and the logical structure corresponding to the designation of the stored document type definition storage means are referred to A structured document generation device, comprising: a structured document generation unit that processes the region cut out by the document image region cutout unit according to the processing attribute in the correspondence definition storage unit and generates according to the logical structure. 2. When a predetermined part of the input is cut out and compared with the corresponding document image area in the correspondence definition storage means and it is judged that they are similar to each other with a certain evaluation value or more, the document in the correspondence definition storage means is similar. 2. The structured document generation apparatus according to claim 1, further comprising a document image area estimation unit that replaces the image area or determines that the image area is equivalent to the document image area. 3. A detailed processing table is specified as a processing attribute, and a partial area of input document image data is processed based on the detailed processing table to generate an output document. The structured document generation device according to claim 1. 4. The structured document generation device according to claim 1, wherein a predetermined image portion of the input is cut out to form an image file, which is combined with the document tag and output. 5. An output document is generated by referring to a table prepared as another detailed processing table or applying an algorithm to a predetermined part of the input, converting it into another character string. The structured document generation device according to claim 3, wherein 6. When a document tag indicating a portion of a logical structure that constitutes a document is designated for a document image area which is a predetermined portion of the cut out input, the two are linked to the correspondence definition storage means. 2. The structured document generation device according to claim 1, further comprising a document image area / logical structure associating unit to be registered / stored. 7. The structured structure according to claim 6, wherein the document tag is supported, and an attribute for performing a predetermined process is added to the output document and registered / stored in the correspondence definition storage means. Document generation device. 8. A document image area of a corresponding portion of the document image data is identified and displayed by adding display means for input document image data and designating a document tag. The structured document generation device according to claim 1. 9. An arbitrary document type definition stored in the document type definition storing means is read out and displayed in a tree structure, or when the document type definition of the tree structure which is corrected and displayed is instructed to be stored, 2. The structured document generation apparatus according to claim 1, further comprising a document logical structure display / correction unit stored in the document type definition storage unit. 10. A document logical structure display / correction means is added to display an arbitrary document type definition modified by a tree structure, and a document image area / logical structure associating means is displayed by a tree structure. 7. The structured document generation device according to claim 6, wherein correspondence is made for each document tag with. 11. A display means for input document image data is added, and when a document tag is designated, the document image area of the corresponding portion of the document image data is identified and displayed. The structured document generation device described. 12. The document image area / logical structure associating unit also associates each document tag with a document type definition displayed in a tree structure, and if there is an independent tree structure without the correspondence, the document logical structure. 11. The structured document generation apparatus according to claim 10, wherein the display / correction unit is configured to identify and display the tree structure having no correspondence.