JP2004038605A - Image processor, image processing method, and program for controlling image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize a device capable of automatically designating an important area and a non-important area in a document. <P>SOLUTION: An item extraction part 20 extracts an item area from a document image of a prescribed format based on the format, and extracts an item kind showing of which kind the item area is and an item content showing the content of the item area. An importance imparting reference holding part 300 holds importance imparting references showing the relation of importance to item kind and item content. An importance imparting part 30 compares the item kind and item content extracted by the item extraction part 20 with the importance imparting references of the part 300 and imparts an importance for each item. An output part 40 outputs document information with the importance imparted for each item. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image processing apparatus that extracts an arbitrary region of a document determined to be important based on the content of the document, and for example, hides an arbitrary region of a document determined to be important. This is used when a document is made public.
[0002]
[Prior art]
Conventionally, an arbitrary area of a document determined to be important based on the content of the document has been designated by a user on a display screen. For example, in Japanese Patent Application Laid-Open No. 6-176127, "electronic filing device", a document image to be registered is displayed on a display screen, and a user designates an arbitrary region in the document image as an important region. Here, it is also possible to specify a unique password for each important area. When a viewer displays the image, the designated important area is kept secret or the unique password is entered to input the unique password. And make it public.
However, in the electronic filing apparatus, since a document image is displayed on a display screen and an important area is manually specified, it is necessary to individually specify an important area when registering documents of the same document type. In some cases, the entire document image cannot be displayed depending on the display screen, and it is necessary to scroll the document image on the display screen to specify an important area. .
[0003]
As a solution to the above problem, there is Japanese Patent Laid-Open Publication No. Hei 7-114616 "Slip Document Information System". In this slip document information system, form data such as a table frame position and a character line position of the slip document is extracted and collated with the form data (dictionary data) registered by the user to identify and read the form of the slip document. An area or important area is specified. A plurality of access levels are set for the important area, and whether the area is open or closed can be set according to the access level of the viewer.
[0004]
[Problems to be solved by the invention]
However, in the voucher document information system, since the important area is automatically extracted, the problem of operational complexity is solved, but the important area described on the dictionary data is fixed. In the case of documents of the same type but different formats, it was necessary to individually specify important areas. Also, when it is desired to specify areas having the same contents as important areas, although the document types are different, it is necessary to individually specify important areas.
[0005]
[Means for Solving the Problems]
The present invention employs the following configuration to solve the above-described problem.
<Configuration 1>
For a document in a predetermined format, an item type that extracts an item area, which is a unit forming the content of the document based on the document format, from the image of the document and indicates the type of the item area An item extraction unit that extracts item contents indicating the contents of the item area; an importance assignment reference holding unit that holds an importance assignment criterion indicating a relationship between the item type and the item content in advance; Comparing the item type and the item content extracted by the extraction unit with the importance assignment criterion of the importance assignment criterion holding unit, and having an importance assigning unit for assigning an importance to each item. Image processing device.
[0006]
<Configuration 2>
For a document of a predetermined format, an item area which is a unit forming the content of the document based on the document format is extracted from the image of the document, and how the item area differs from other item areas in the document. A logical structure extraction unit that obtains a logical structure indicating the relationship, and extracts a logical structure type indicating the type of the logical structure, and a logical structure content indicating the content of the logical structure, An importance assigning criterion holding unit that holds an importance assigning criterion indicating a relationship between the logical structure type and the logical structure content, and an importance assigning unit that assigns the logical structure type and the logical structure content extracted by the logical structure extracting unit. An image processing apparatus comprising: an importance assigning unit that assigns an importance to each logical structure as compared with an importance assignment reference of a reference holding unit.
[0007]
<Configuration 3>
For a document of a predetermined format, from a document image, a document identification unit for identifying a document type indicating what kind of document the document is based on the document format, and for a document of a predetermined format, From the document image, an item area that is a unit forming the contents of the document based on the document format is extracted, and an item type indicating the type of the item area and an item indicating the contents of the item area An item extraction unit for extracting the contents, an importance assignment criterion holding unit for holding an importance assignment criterion indicating a relationship between the document type, the item type, and the item content in advance, and a document identification unit. The document type, the item type and the item content extracted by the item extraction unit are compared with the importance assignment criterion of the importance assignment criterion holding unit, and the importance assigning unit that assigns the importance to each item is compared. An image processing apparatus comprising:
[0008]
<Configuration 4>
For a document of a predetermined format, from a document image, a document identification unit for identifying a document type indicating what kind of document the document is based on the document format, and for a document of a predetermined format, From the image of the document, extract the item area that is the unit that forms the content of the document based on the document format, and create a logical structure that indicates how the item area is related to other item areas in the document. A logical structure type indicating the type of the logical structure obtained and the logical structure; a logical structure extracting unit for extracting logical structure contents indicating the contents of the logical structure; and a document type, a logical structure type, and a logical structure An importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between the content and the importance, a document type identified by the document identification unit, a logical structure type and a logical structure content extracted by the logical structure extraction unit And the importance assignment criteria Compared to the importance granted criterion lifting unit, an image processing apparatus characterized by comprising a severity imparting unit for imparting the importance for each logical structure.
[0009]
<Configuration 5>
For a document of a predetermined format, from a document image, a use identification unit for identifying a use type indicating a purpose of use of the document based on the document format, and Then, from the image of the document, an item area which is a unit forming the content of the document based on the format of the document is extracted, and an item type indicating the type of the item area and the content of the item area are extracted. An item extracting unit for extracting the item contents to be indicated, an importance assigning criterion holding unit for holding in advance an importance assigning criterion indicating a relationship between the use type, the item type and the item contents, and a use identifying unit. The importance assigning unit assigns an importance to each item by comparing the application type, the item type and the item content extracted by the item extracting unit with the importance assigning criterion of the importance assigning criterion holding unit. An image processing apparatus comprising:
[0010]
<Configuration 6>
For a document of a predetermined format, from a document image, a use identification unit for identifying a use type indicating a purpose of use of the document based on the document format, and And extract, from the image of the document, an item area which is a unit forming the contents of the document based on the document format, and indicate how the item area is related to other item areas in the document. A logical structure type that determines the structure and indicates the type of the logical structure, a logical structure extracting unit that extracts a logical structure content indicating the contents of the logical structure, and a use type and a logical structure type in advance. An importance assignment criterion holding unit that holds an importance assignment criterion indicating the relationship of the importance to the logical structure content, a use type identified by the use identification unit, and a logical structure type and a logic extracted by the logical structure extraction unit. Structural content and importance Compared to the importance granted criterion reference holder, the image processing apparatus characterized by comprising a severity imparting unit for imparting the importance for each logical structure.
[0011]
<Configuration 7>
For a document of a predetermined format, from the image of the document, a use type indicating a purpose of use of the document based on the format of the document and a document type indicating the type of the document are set. A use / document identification unit for identifying, and for a document in a predetermined format, extracting, from an image of the document, an item area which is a unit forming the content of the document based on the document format, and how the item area is Item extraction unit that extracts an item type indicating whether the item type is an appropriate type, and an item content indicating the content of the item area, and an important item that indicates in advance the relationship between the application type, the document type, the item type, and the importance of the item content. The importance assignment criterion holding unit for holding the importance assignment criterion, the application type and the document type identified by the use / document identification unit, and the item type and the item content extracted by the item extraction unit are held by the importance assignment criterion. Of each department The image processing apparatus characterized by comprising a severity imparting unit for imparting the importance for each eye.
[0012]
<Configuration 8>
For a document of a predetermined format, from the image of the document, a use type indicating a purpose of use of the document based on the format of the document and a document type indicating the type of the document are set. A use / document identification unit for identifying, and for a document of a predetermined format, an item area which is a unit forming the content of the document based on the document format is extracted from the image of the document, and the item area is included in the document. Finds the logical structure indicating the relationship with other item areas, and extracts the logical structure type indicating the type of the logical structure and the logical structure contents indicating the contents of the logical structure A logical structure extracting unit, an importance assigning criterion holding unit that holds in advance an importance assigning criterion indicating a relationship between importance for use type, document type, logical structure type, and logical structure content, and a use / document identifying unit Type and document identified by Separately, the logical structure type and the logical structure content extracted by the logical structure extracting unit are compared with the importance assigning criterion of the importance assigning criterion holding unit, and the importance is assigned to each logical structure. And an image processing apparatus.
[0013]
<Configuration 9>
For a document of a predetermined format, an item type indicating a type of the item region as well as extracting an item region as a unit forming the content of the document based on the document format from the document image And the item contents indicating the contents of the item area are extracted. Next, the item types and the item contents are compared in advance with an importance assignment criterion indicating the relationship between the item types and the item contents, and An image processing method characterized by assigning importance to each item.
[0014]
<Configuration 10>
A computer constituting the image processing apparatus is configured to extract, from a document image, an item area serving as a unit forming the content of the document based on the document format for a document in a predetermined format, and determine how the item area is determined. An item extraction unit that extracts an item type indicating whether the item type is an appropriate type, an item content indicating the content of the item area, and the item type and item content extracted by the item extraction unit A control program for an image processing apparatus for causing the image processing apparatus to function as an importance assigning unit that assigns an importance to each item, as compared with an importance assignment criterion indicating a relationship of importance with respect to each item.
[0015]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail using specific examples.
<< Specific Example 1 >>
In the first specific example, an area of a document that may be determined to be important based on the content of the document, a type indicating the content of the area and the content of the area are extracted, and the content of the area defined by the user is extracted. The conventional problem is solved by designating the area of the document as an important area based on the correspondence between the type and the content of the area indicating the importance and the importance in a plurality of stages.
[0016]
<Constitution>
FIG. 1 is a configuration diagram showing a specific example 1 of the image processing apparatus of the present invention.
The illustrated image processing apparatus 1 is connected to an input device 2 and a storage device 3, and includes an input unit 10, an item extraction unit 20, an importance assignment unit 30, an output unit 40, a document image storage unit 100, and an importance assignment target. An information storage unit 200 and an importance assignment reference holding unit 300 are provided. Further, the image processing apparatus 1 is composed of, for example, a personal computer.
[0017]
The input device 2 includes, for example, a scanner that optically reads an image of a document and outputs it as digital data, or a communication interface that provides image data of the document. The storage device 3 includes an external storage device, a communication interface, and the like for storing data of the document to which the importance is output, which is output from the image processing device 1. The input unit 10 is an input interface of the image processing apparatus 1 for inputting document image data from the input device 2 and transmitting the document image data to the document image storage unit 100. The item extracting unit 20 extracts, from the image of the document having a predetermined format stored in the document image storage unit 100, an item region serving as a unit for forming the content of the document based on the format, and It has a function of extracting an item type indicating the type and an item content indicating the content of the item area. The importance assigning unit 30 compares the item type and the item content extracted by the item extracting unit 20 with the importance assigning criterion of the importance assigning criterion holding unit 300, and assigns an importance to each item. It is. The output unit 40 stores the document image stored in the document image storage unit 100 and the importance assignment target information assigned importance for each item stored in the importance assignment target information storage unit 200 in the storage device 3. This is an output interface of the image processing apparatus 1 for outputting to the image processing apparatus.
[0018]
The document image storage unit 100 is a storage unit that stores a document image input by the input unit 10, and includes a semiconductor memory, a hard disk device, and the like. The importance assignment target information storage unit 200 is a storage unit that stores the data of each item extracted by the item extraction unit 20 and holds the data of the item assigned the importance by the importance assignment unit 30. It comprises a semiconductor memory, a hard disk device, and the like. The importance assignment criterion holding unit 300 is a storage unit that holds an importance assignment criterion indicating the relationship between the item type and the item content, and is composed of a semiconductor memory, a hard disk device, or the like.
In the drawing, the thin arrows indicate the flow of data, and the thick arrows indicate the flow of control. The item extracting unit 20 and the importance assigning unit 30 are configured by software corresponding to each function and hardware such as a CPU and a memory for executing the software.
[0019]
<motion>
FIG. 2 is a flowchart illustrating the operation of the first embodiment.
[Operation of Input Unit 10]
The input unit 10 acquires a document image from the input device 2 such as a scanner connected to the image processing apparatus 1, and stores the acquired document image in the document image storage unit 100 (Step S10).
[0020]
[Operation of item extraction unit 20]
Next, the item extracting unit 20 acquires the document image stored in the document image storage unit 100 by the input unit 10 and, for all or some of the items in the acquired document image, the item type and the item area (item area). (The coordinate value of the circumscribed rectangle of) and the item contents (character recognition result of the item). Here, the item is a unit forming the content of the document. Then, importance assignment target information describing the extracted item type, item area, and item content is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200 (step S20).
[0021]
FIG. 3 is an explanatory diagram illustrating an example of a document image.
FIG. 4 is an explanatory diagram illustrating an example of a component.
FIG. 5 is an explanatory diagram illustrating an example of an item.
For example, when the document image stored in the document image storage unit 100 by the input unit 10 as the document in the predetermined format is the paper image shown in FIG. 3, the item extraction unit 20 determines all of the paper images shown in FIG. Alternatively, for some items, item types such as “body”, “section title”, “figure”, and “caption”, and item areas such as “body area”, “section title area”, “figure area”, and “caption area” The coordinates (rectangular coordinate values) and item contents (character recognition results of items) such as “body content”, “section title content”, and “caption content” are extracted. However, the item contents are extracted only for the items whose item type is determined from the components whose component type is “text” shown in FIG. That is, the item types are only items such as “text”, “section title”, and “caption”. A component is a unit that forms a paper surface of a document.
[0022]
Next, the item extracting unit 20 extracts the item types such as “text”, “section title”, “figure”, and “caption” and the items such as “body area”, “section title area”, “figure area”, and “caption area”. Creates importance assignment target information describing an area and item contents such as "text content", "section title content", and "caption content", and stores the created importance assignment target information in the importance assignment target information storage unit 200. I do.
[0023]
FIG. 6 is an explanatory diagram illustrating an example of the importance assignment target information.
In FIG. 6, the item area is indicated by “(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)” of the circumscribed rectangle of the item. The item type, the item area, and the item content are arranged in order from top to bottom in the left column.
Hereinafter, item types such as "body", "section title", "figure", and "caption", and item areas such as "body area", "section title area", "figure area", and "caption area", and "body content" and "section title" Extraction of item contents such as "contents" and "caption contents" will be described.
[0024]
The item extracting unit 20 first extracts a black pixel set by examining the connectivity of black pixels in the paper image. If the height and width of the circumscribed rectangle of the extracted black pixel set is greater than an arbitrary value, the extracted black pixel set is determined to be a component whose component type is “figure”, and the extracted black pixel set The circumscribed rectangle is extracted as a “figure area”. Since the components of the figure can be considered as items that are the units forming the contents of the document, it is determined that the component with the component type of the figure is the item of the item type of the figure. , A “drawing area” which is a component area is an item area.
On the other hand, if the height and width of the circumscribed rectangle of the extracted black pixel set is smaller than an arbitrary value, the extracted black pixel set is determined to be a component having the component type “character”, and the extracted black pixel The circumscribed rectangle of the set is extracted as a “character area”. If the components of the “character” are arranged in the horizontal direction, and the horizontal distance between two adjacent components of the “character” is smaller than an arbitrary value, the configuration of the “character” arranged in the horizontal direction The element set is determined to be a component whose component type is “character string”, and a circumscribed rectangle of the component set of “character” is extracted as a “character string area”.
[0025]
Further, when the components of the “character string” are arranged in the vertical direction, and when the vertical distance between two adjacent components of the “character string” is smaller than an arbitrary value, the “character string” is arranged in the vertical direction. Is determined to be a component having a component type of “text”, and a circumscribed rectangle of the component set of “character string” is extracted as a “text region”.
Subsequently, a character set is processed for a set of components of "character", which is a component of "sentence". If the obtained character recognition result includes punctuation, the component of "sentence" is an item type. Is determined to be an item of “text”, the “text area”, which is a component area, is set to “text area”, and the character recognition result is set to “text content”.
On the other hand, in the obtained character recognition result, if the character string from the beginning is “number”, “.”, “Number”, “.”, The component of “sentence” is an item whose item type is “section title”. It is determined that there is, and the “text section”, which is a component area, is set as a “section title area”, and the character recognition result is set as “section title content”. In the case of a paper, if the character string from the beginning is “number”, “.”, “Number”, “.”, The item extraction unit 20 has in advance pattern data indicating that it is an item of “section title”. Suppose you are.
In addition, in the obtained character recognition result, when the first character is “figure”, the component of “sentence” is determined to be an item of item type “caption”, and “sentence area” which is a component area is determined. "As a" caption area "and the character recognition result as" caption content ". Also in this case, the item extracting unit 20 has such pattern data in advance.
[0026]
[Operation of importance imparting section 30]
The importance assigning unit 30 acquires the importance assigning target information stored in the importance assigning target information storage unit 200 by the item extracting unit 20, and further obtains the importance assigning pre-stored in the importance assigning reference holding unit 300. Get the criteria.
FIG. 7 is an explanatory diagram illustrating an example of the importance assignment criterion.
Then, the importance assigning unit 30 assigns the importance assignment target information (importance degree) describing the item type, the item area, and the importance based on the correspondence between the item type, the item content, and the importance described in the importance assignment criterion. ) Is created, and the created importance assignment target information (including the importance) is stored in the importance assignment target information storage unit 200 (step S30).
FIG. 8 is an explanatory diagram illustrating an example of importance assignment target information (including importance).
For example, the importance assignment target information stored in the importance assignment target information storage unit 200 by the item extraction unit 20 is the importance assignment target information shown in FIG. In the case where the importance assignment criterion is the importance assignment criterion shown in FIG. 7, the importance assignment section 30 sets items such as “text”, “section title”, “figure”, and “caption” described in the importance assignment criterion shown in FIG. Item type and item area based on the correspondence between the type, item contents such as “including character string“ OCR ”” and “unspecified” and importance such as “high”, “medium”, “low”, “high” and “high” The importance assignment target information (including the importance) shown in FIG. 8 describing the importance is created, and the created importance assignment target information (including the importance) is stored in the importance assignment target information storage unit 200.
[0027]
[Operation of Output Unit 40]
The output unit 40 acquires the document image stored in the document image storage unit 100 by the input unit 10, and further obtains the importance assignment target information stored in the importance assignment target information storage unit 200 by the importance assignment unit 30 ( (Including importance). Then, the acquired document image and the information to be added with importance (including importance) are stored in the storage device 3 connected to the image processing apparatus 1 (step S40).
[0028]
In the above description of the operation, words such as “text”, “section title”, “figure”, and “caption” are used as item types, but any word that can be easily understood as a unit forming the contents of a document , Any word may be used.
The circumscribed rectangle “(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)” of the circumscribed rectangle of the item is used in the item area of the importance assignment target information, but it is the coordinate value of the circumscribed rectangle of the item Any format may be used as long as it can be easily understood.
Item types such as "Body", "Section title", "Figure", "Caption", etc., item areas such as "Body area", "Section title area", "Figure area", "Caption area", and "Body contents", "Section title contents", " The above method was used to extract the item contents such as “caption contents”, but the item types such as “body”, “section title”, “figure”, “caption”, etc., “body area”, “section title area”, “figure area”, “figure area” Other methods may be used as long as they can extract item areas such as "caption area" and item contents such as "text content", "section title content" and "caption content".
Although the specification items such as "including character string" OCR "" are used for the item content of the importance assignment standard, any specification item that can easily understand that the item content is specified is specified. Matters may be used.
Although a word such as "no designation" is used as the item content of the importance assignment criterion, any word may be used as long as it is easily understood that the item content is not specified. Also, “no designation” may be used for all item contents.
Words such as "high", "medium", and "low" were used as the importance of the importance assignment criteria, but any word that can easily understand the importance level and importance is used. May be. The number of levels of importance may be two (important or not important).
[0029]
<effect>
As described above, according to the first specific example, an item is extracted from a document image, and an importance is assigned to this item using a previously assigned importance assignment criterion. Has a significant effect.
That is, since the importance is automatically given to the item according to the content of the item which is a unit forming the content of the document, when a document is made public, the importance is relatively high for a certain viewer. Keep the "body", "figure" and "caption" private, or keep only the "body" with important keywords in addition to the "figure" and "caption" that are important to other viewers. Can be.
That is, disclosure / non-disclosure of items according to the degree of importance can be adjusted for each item content. This is because even if a user who does not know the contents of the document deeply sets the importance, for example, the body with important keywords may need to be kept private because the whole body may be important. Indicates that the importance can be set to some extent according to the importance of the contents of the document.
[0030]
<< Specific Example 2 >>
In the specific example 2, importance is assigned to each logical structure indicating the relationship between items.
[0031]
<Constitution>
FIG. 9 is a configuration diagram of the specific example 2.
The image processing device 1 a is connected to the input device 2 and the storage device 3, and has an input unit 10, a logical structure extraction unit 21, an importance assigning unit 31, an output unit 40, a document image storage unit 100, and importance assignment target information. The storage unit 200 includes an importance assignment reference holding unit 300.
Here, the input unit 10, the output unit 40, the document image storage unit 100, the importance level assignment target information storage unit 200, and the importance level assignment reference holding unit 300 are the same as those in the first embodiment, and thus description thereof is omitted. I do. However, the importance assignment criterion of the importance assignment criterion holding unit 300 in the specific example 2 is data of importance for the logical structure type and the logical structure content.
The logical structure extracting unit 21 extracts, from a document image of a document of a predetermined format stored in the document image storage unit 100, an item area serving as a unit for forming the content of the document based on the document format. At the same time, a logical structure indicating the relationship between the item area and other item areas in the document is obtained, and a logical structure type indicating the type of the logical structure and the contents of the logical structure are described. It has a function to extract the logical structure contents shown. The importance assigning unit 31 compares the logical structure type and the logical structure content extracted by the logical structure extracting unit 21 with the importance assigning criterion of the importance assigning criterion holding unit 300, and determines the importance for each logical structure. It has the function of giving.
[0032]
<motion>
FIG. 10 is a flowchart illustrating the operation of the second embodiment.
In addition, also in the specific example 2, description will be made using an example of the document image of FIG. 3 and an example of the components of FIG. The description of the input unit 10 and the output unit 40 that perform the same operations as in the first embodiment will be omitted. That is, in FIG. 10, steps S10 and S40 are the same as those in the first example, and thus description thereof will be omitted.
[0033]
[Operation of Logical Structure Extraction Unit 21]
The logical structure extraction unit 21 acquires the document image stored in the document image storage unit 100 by the input unit 10 and, for all or a part of the logical structure in the acquired document image, a logical structure type and a logical structure area ( The coordinate value set of the circumscribed rectangle of the item forming the logical structure and the logical structure content (character recognition result set of the item forming the logical structure) are extracted. Then, importance assignment target information describing the extracted logical structure type, logical structure area, and logical structure content is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200 (step S21). .
Note that the logical structure is a structure of an item based on the contents of a document. That is, it is information indicating what relationship each item has in the document when there are a plurality of items.
[0034]
FIG. 11 is an explanatory diagram illustrating an example of the logical structure.
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is the paper image shown in FIG. 3, the logical structure extraction unit 21 determines all or a part of the logical structure in the paper image shown in FIG. For the logical structure type such as “2.2. Section text” and “2.3. Section text”, and the logical structure area such as “2.2 section text area” and “2.3. Section text area” Extraction of the coordinate value set of the circumscribed rectangle of the item to be formed) and logical structure contents (character recognition result set of items forming the logical structure) such as "2.2. Section body content" and "2.3. Section body content" I do. However, the logical structure content is extracted only from the logical structure formed by the item whose item type is determined from the component whose component type is “text” shown in FIG. That is, the logical structure type is only the logical structure such as “2.2. Section text” or “2.3. Section text”. Then, the extracted logical structure types such as “2.2. Section text” and “2.3. Section text”, logical structure areas such as “2.2. Section text area” and “2.3. Section text area”, and Creates importance assignment target information describing logical structure contents such as “2.2. Section text content” and “2.3. Section text content”, and stores the created importance assignment target information in the importance assignment target information storage unit. 200.
[0035]
FIG. 12 is an explanatory diagram illustrating an example of the importance assignment target information.
As shown in FIG. 12, the logical structure area is indicated by a set of "(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)" of a circumscribed rectangle of items forming the logical structure. Also, the coordinate values of the circumscribed rectangles of the items forming the logical structure are arranged in order from the top to the bottom of the left column. Further, the character recognition results of the items forming the logical structure are arranged in the same order.
Hereinafter, logical structure types such as “2.2. Section text” and “2.3. Section text”, logical structure areas such as “2.2. Section text area” and “2.3. Section text area” and “2. Extraction of logical structure contents such as "2.2. Section text contents" and "2.3. Section text contents" will be described.
[0036]
The logical structure extraction unit 21 first extracts a set of black pixels by examining the connectivity of the black pixels in the paper image. If the height and width of the circumscribed rectangle of the extracted black pixel set is greater than an arbitrary value, the extracted black pixel set is determined to be a component whose component type is “figure”, and the extracted black pixel set The circumscribed rectangle is extracted as a “figure area”. Since the components of the figure can be considered as items that are the units forming the contents of the document, it is determined that the component with the component type of the figure is the item of the item type of the figure. , A “drawing area” which is a component area is an item area.
On the other hand, if the height and width of the circumscribed rectangle of the extracted black pixel set is smaller than an arbitrary value, the extracted black pixel set is determined to be a component having the component type “character”, and the extracted black pixel The circumscribed rectangle of the set is extracted as a “character area”. If the components of the “character” are arranged in the horizontal direction, and the horizontal distance between two adjacent components of the “character” is smaller than an arbitrary value, the configuration of the “character” arranged in the horizontal direction The element set is determined to be a component whose component type is “character string”, and a circumscribed rectangle of the component set of “character” is extracted as a “character string area”.
[0037]
Further, when the components of the “character string” are arranged in the vertical direction, and when the vertical distance between two adjacent components of the “character string” is smaller than an arbitrary value, the “character string” is arranged in the vertical direction. Is determined to be a component having a component type of “text”, and a circumscribed rectangle of the component set of “character string” is extracted as a “text region”. Subsequently, a character set is processed for a set of components of "character", which is a component of "sentence". If the obtained character recognition result includes punctuation, the component of "sentence" is an item type. Is determined to be an item of “text”, the “text area”, which is a component area, is set to “text area”, and the character recognition result is set to “text content”.
[0038]
On the other hand, in the obtained character recognition result, when the character string from the beginning is “number”, “.”, “Number”, “.”, The component element of “sentence” is an item whose item type is “section title” It is determined that the “sentence area”, which is a component area, is “section title area”, and the character recognition result is “section title content”. In addition, in the obtained character recognition result, when the first character is “figure”, the component of “sentence” is determined to be an item of item type “caption”, and “sentence area” which is a component area is determined. "As a" caption area "and the character recognition result as" caption content ". Here, when the character string from the top of “section title content” is “2.3.”, The item of “section title” has a logical structure whose logical structure type is “2.3. Section title”. And the item area “section title area” is set to “2.3. Section title area”, and the item content “section title content” is set to “2.3. Section title content”. In general, papers have a column structure. Items are formed from left to right between columns and from top to bottom within columns. Therefore, the logical structure type is “2.3. The item set of “Body” existing before the logical structure has the logical structure type of “2.2. Section body”, and exists after the logical structure of “2.3. Section title”. It is determined that the item set of “text” has a logical structure whose logical structure type is “2.3. Section text”, and the coordinate value sets of circumscribed rectangles of the items forming the logical structure are respectively described in “2.2. As a logical structure area such as “text area” and “2.3. Section text area”, character recognition result sets of items that form the logical structure are respectively referred to as “2.2. Section text content” and “2.3. And so on.
As described above, in the specific example 2, as a document in a predetermined format, a logical structure is extracted by paying attention to the number of a paper in which each section is sequentially numbered.
[0039]
[Operation of importance imparting section 31]
The importance level assigning unit 31 acquires the importance level assigning target information stored in the importance level assigning information storage unit 200 by the logical structure extracting unit 21, and further acquires the importance level stored in the importance level assigning reference holding unit 300 in advance. Get the grant criteria.
FIG. 13 is an explanatory diagram illustrating an example of the importance assignment criterion.
Then, based on the logical structure type, the logical structure content, and the correspondence between the importance and the logical structure type described in the importance level criterion, the importance level assigning unit 31 describes the importance level assignment target describing the logical structure type, the logical structure area, and the importance level. Information (including importance) is created, and the created importance assignment target information (including importance) is stored in the importance assignment target information storage unit 200 (step S31).
FIG. 14 is an explanatory diagram illustrating an example of importance assignment target information (including importance).
For example, the importance assignment target information stored in the importance assignment target information storage unit 200 by the logical structure extraction unit 21 is stored in advance in the importance assignment target information illustrated in FIG. When the importance assignment criterion is the importance assignment criterion illustrated in FIG. 13, the importance assignment unit 31 determines the logical structure type such as “XX section text” described in the importance assignment criterion illustrated in FIG. FIG. 14 describes the logical structure type, the logical structure area, and the importance based on the correspondence between the logical structure contents such as “including OCR” and “unspecified” and the importance such as “high” and “medium”. The importance assignment target information (including the importance) is created, and the created importance assignment target information (including the importance) is stored in the importance assignment target information storage unit 200. Here, “section XX” of “section XX body” indicates an arbitrary section number.
[0040]
In the above description of the operation, words such as “2.2. Section text” and “2.3. Section text” are used for the logical structure type, but it is easily understood that the item structure is based on the contents of the document. Any word that can be used may be used.
Although a set of “(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)” of a circumscribed rectangle of an item forming a logical structure in the logical structure area of the importance assignment target information is used, Any format may be used as long as it can be easily understood that it is a set of coordinate values of a circumscribed rectangle of the item forming the.
Logical structure types such as "2.2. Section text" and "2.3. Section text", logical structure areas such as "2.2. Section text area" and "2.3. Section text area", and "2.2. Although the above method was used to extract the logical structure contents such as “.Section text contents” and “2.3.Section text contents”, the logical structures such as “2.2.Section text” and “2.3.Section text” were used. Type, logical structure areas such as “2.2. Section body area” and “2.3. Section body area” and logical structure contents such as “2.2. Section body contents” and “2.3. Section body contents” Any method may be used as long as it can be extracted.
Although the specification items such as "including character string" OCR "" are used in the logical structure content of the importance assignment standard, what kind of specification items can be easily understood as the matter specifying the logical structure content. May be used.
Although a word such as “no designation” is used as the logical structure content of the importance assignment reference, any word may be used as long as it is easily understood that the logical structure content is not specified. In addition, “not specified” may be used for all logical structure contents.
Although words such as “high” and “medium” are used as the importance of the importance assignment criterion, any word may be used as long as the word of the importance level and the level of importance can be easily understood. . Further, the number of levels of importance may be two types, that is, important and not important.
[0041]
<effect>
As described above, according to the specific example 2, the logical structure indicating the relation between the items in the document is extracted from the document image, and the importance assigning criterion registered in advance for the logical structure is extracted. Is used to assign importance, and the following effects are obtained.
That is, importance is automatically assigned to the logical structure according to the contents of the logical structure, which is the structure of the item based on the contents of the document. Relatively high “2.2.Section text” and “2.3.Section text” are kept private, or only “2.3.Section text”, which is highly important to other viewers, is kept private. be able to. That is, the disclosure / non-disclosure of the logical structure can be adjusted according to the importance. This is because even if a user who does not know the contents of the document deeply sets the importance, for example, the body with important keywords needs to be kept private because the body itself may be important. Indicates that the importance can be set to some extent according to the importance of the contents of the document. In addition, since the logical structure of “2.3. Section body” is an item set of “Body”, specific example 1 has important keywords in the “Body section” of “2.3. Section body”. However, in the specific example 2, it is possible to set an accurate importance for the whole text, while there is a possibility that a high importance is set only for the “text” having “. Also, by using designation items such as “including the character string“ OCR ”in the content of the section XX title” as the logical structure content of the importance assignment standard, only the logical structure content of “section section XX title” is used. By searching, the corresponding “XX section text” can be efficiently derived.
[0042]
<< Specific Example 3 >>
In the specific example 3, the document type is obtained from the image information in the document, and the importance is given in consideration of the document type.
[0043]
<Constitution>
FIG. 15 is a configuration diagram of the third embodiment.
The illustrated image processing device 1b is connected to the input device 2 and the storage device 3, and includes an input unit 10, a document identification unit 52, an item extraction unit 22, an importance assignment unit 32, an output unit 40, and a document image storage unit 100. , An importance assignment target information storage unit 200 and an importance assignment reference holding unit 300.
Here, the input unit 10, the output unit 40, the document image storage unit 100, the importance assignment target information storage unit 200, and the importance assignment reference holding unit 300 are the same as those in the first and second specific examples, and therefore will be described here. Is omitted. However, the importance assignment criterion of the importance assignment criterion holding unit 300 of the specific example 3 is data of importance including the document type in addition to the item type and the item content.
[0044]
The document identification unit 52 is a functional unit that identifies a document type indicating a type of a document based on the format of the document from a document image of a predetermined format stored in the document image storage unit 100. The item extracting unit 22 has a function of performing the same processing as that of the item extracting unit 20 of the first embodiment based on the document identification result of the document identifying unit 52. Further, the importance assigning unit 32 compares the document type identified by the document identifying unit 52, the item type and the item content extracted by the item extracting unit 22 with the importance assigning criterion of the importance assigning criterion holding unit 300. In comparison, this is a functional unit that assigns importance to each item.
[0045]
<motion>
FIG. 16 is a flowchart illustrating the operation of the third example.
The description of the input unit 10 and the output unit 40 that perform the same operations as those of the first and second examples will be omitted. That is, in FIG. 16, the description of step S10 and step S40 is omitted.
[0046]
[Operation of Document Identification Unit 52]
The document identification unit 52 acquires the document image stored in the document image storage unit 100 by the input unit 10, and identifies the document type of the acquired document image. Then, importance assignment target information (document type only) describing the identified document type is created, and the created importance assignment target information (document type only) is stored in the importance assignment target information storage unit 200 (step S52). ).
FIG. 17 is an explanatory diagram illustrating an example of the document image in the specific example 3.
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is such a paper image as shown in FIG. 17 (this is a document type of “paper A” at the upper right end of the paper image shown in FIG. 3). The document identification unit 52 extracts the document type of “paper A” described in the upper right end of the paper image, and identifies the document type of the paper image. Then, importance assignment target information (document type only) describing the identified document type is created, and the created importance assignment target information (document type only) is stored in the importance assignment target information storage unit 200.
[0047]
FIG. 18 is an explanatory diagram illustrating an example of the importance assignment target information (only the document type).
Hereinafter, extraction of the document type will be described.
The document identification unit 52 first extracts a set of black pixels by examining the connectivity of the black pixels near the upper right end in the paper image. If the height and width of the circumscribed rectangle of the extracted black pixel set are greater than an arbitrary value, the extracted black pixel set is determined to be a component having a component type of “frame”, and the extracted black pixel set The circumscribed rectangle is extracted as a “frame area”. On the other hand, if the height and width of the circumscribed rectangle of the extracted black pixel set is smaller than an arbitrary value, the extracted black pixel set is determined to be a component having the component type “character”, and the extracted black pixel The circumscribed rectangle of the set is extracted as a “character area”. If the components of the “character” are arranged in the horizontal direction, and the horizontal distance between two adjacent components of the “character” is smaller than an arbitrary value, the configuration of the “character” arranged in the horizontal direction The element set is determined to be a component whose component type is “character string”, and a circumscribed rectangle of the component set of “character” is extracted as a “character string area”. Further, a character recognition process is performed on a component set of “character” which is a component of “character string”, and an obtained character recognition result (“paper A”) is extracted as a document type.
[0048]
[Operation of item extraction unit 22]
The item extracting unit 22 acquires the document image stored in the document image storing unit 100 by the input unit 10 and, for all or some of the items in the acquired document image, an item type and an item area (circumscribed rectangle of the item). ) And the item contents (character recognition result of the item) are extracted. An item is a unit that forms the content of a document. Further, the item extraction unit 22 acquires the importance level assignment target information (only the document type) stored in the importance level assignment target information storage unit 200 by the document identification unit 52, and obtains the document type and the extracted item type, item area, and The importance assignment target information describing the item contents is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200 (step S22).
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is the article image shown in FIG. 17, the item extracting unit 22 determines whether all or some of the items in the article image shown in FIG. Item types such as “body”, “section title”, “figure”, “caption”, etc., item areas such as “body area”, “section title area”, “figure area”, “caption area” (the coordinate values of the circumscribed rectangle of the item) and “ Item contents (character recognition results of items) such as body contents, section title contents, and caption contents are extracted. However, the item contents are extracted only for the items whose item type is determined from the components whose component type is “text” shown in FIG. That is, the item types are only items such as “text”, “section title”, and “caption”. Further, the item extraction unit 22 acquires the importance assignment target information (only the document type) shown in FIG. 18 stored in the importance assignment target information storage unit 200 by the document identification unit 52, and obtains the document type of “paper A”. Item types such as "text", "section title", "figure", "caption", etc., "text area", "section title area", "figure area", "caption area", etc., and "text content", "section title" The importance assignment target information that describes item contents such as "contents" and "caption contents" is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200.
FIG. 19 is an explanatory diagram illustrating an example of the importance level assignment target information in the specific example 3. In FIG. 19, the item area is indicated by the circumscribed rectangle “(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)” of the item. The item type, the item area, and the item content are arranged in order from top to bottom in the left column.
Item types such as "Body", "Section title", "Figure", "Caption", etc., item areas such as "Body area", "Section title area", "Figure area", "Caption area" and "Body contents", "Section title contents" The extraction of item contents such as "" and "caption contents" is the same as in the first embodiment.
[0049]
[Operation of Importance Assigning Unit 32]
The importance level assigning unit 32 acquires the importance level assigning target information stored in the importance level assigning target information storage unit 200 by the item extracting unit 22, and further obtains the importance level assigning level stored in advance in the importance level assigning reference holding unit 300. Get the criteria.
FIG. 20 is an explanatory diagram illustrating an example of the importance assignment criterion.
Then, the importance assigning unit 32 writes the document type, the item type, the item area, and the importance based on the correspondence between the document type, the item type, the item content, and the importance described in the importance assignment criterion. Assignment target information (including importance) is created, and the created importance assignment target information (including importance) is stored in the importance assignment target information storage unit 200 (step S32).
FIG. 21 is an explanatory diagram illustrating an example of importance assignment target information (including importance).
For example, the importance assignment target information stored in the importance assignment target information storage unit 200 by the item extraction unit 22 is the importance assignment target information shown in FIG. In the case where the importance assignment criterion is the importance assignment criterion shown in FIG. 20, the importance assigning unit 32 outputs the document type such as "article A" or "unspecified" described in the importance assignment criterion shown in FIG. Item types such as “section title”, “figure”, “caption”, etc., and item contents such as “including character string“ OCR ”” and “unspecified” and “high”, “low”, “medium”, “low”, “high”, “high” 21. Based on the correspondence of the degrees of importance such as "", the information on the degree of importance (including the degree of importance) shown in FIG. (Including importance) is stored in the importance assignment target information storage unit 200. That.
[0050]
In the above description of the operation, a word such as "thesis A" is used as the document type, but any word can be used as long as the word can be easily understood as the document type. Further, although the above method is used for extracting the document type, any method may be used as long as the document type can be extracted.
Further, although a word such as "not specified" is used as the document type of the importance assignment criterion, any word may be used as long as it is easily understood that the document type is not specified. Further, “not specified” may be used for all document types.
Other uses are the same as in the first embodiment.
[0051]
<effect>
As described above, according to the specific example 3, in addition to the configuration of the specific example 1, the document type is extracted based on the document format, and the importance of each item is assigned in consideration of the document type. Therefore, the following effects are obtained.
That is, items are automatically assigned importance in accordance with the document type and the contents of the items which are the units forming the contents of the document. Therefore, when a document is made public, the importance of a certain document is relatively low. It is possible to keep high “body”, “figure” and “caption” private, and to keep only “body” with important keywords in addition to “figure” and “caption” that are important in other documents. Yes (the "body" that has been kept private in one document can be made public in another document). That is, disclosure / non-disclosure of items according to importance can be adjusted for each document type and item content. Compared to the specific example 1, the document can be finely disclosed since the adjustment can be made for each document type.
[0052]
<< Specific Example 4 >>
In the specific example 4, a use type indicating the purpose of use of the document is obtained from the image information in the document, and importance is given in consideration of the use type.
[0053]
<Constitution>
FIG. 22 is a configuration diagram of the specific example 4.
The illustrated image processing device 1c is connected to the input device 2 and the storage device 3, and includes the input unit 10, the use identifying unit 53, the item extracting unit 23, the importance assigning unit 33, the output unit 40, and the document image storing unit 100. , An importance assignment target information storage unit 200 and an importance assignment reference holding unit 300.
Here, the input unit 10, the output unit 40, the document image storage unit 100, the importance assignment target information storage unit 200, and the importance assignment reference holding unit 300 are the same as those in the specific examples 1 to 3, and will be described here. Is omitted. However, the importance assignment criterion of the importance assignment criterion holding unit 300 of the specific example 4 is data of the importance including the use type in addition to the item type and the item content.
[0054]
The use identifying unit 53 is a function of identifying, from a document image in a predetermined format stored in the document image storage unit 100, a use type indicating a document used for what purpose based on the format of the document. Department. The item extracting unit 23 has a function of performing the same processing as that of the item extracting unit 20 of the first specific example, based on the usage identification result of the usage identifying unit 53. Further, the importance assigning unit 33 compares the use type identified by the use identifying unit 53, the item type and the item content extracted by the item extracting unit 23 with the importance assigning criterion of the importance assigning criterion holding unit 300. In comparison, this is a functional unit that assigns importance to each item.
[0055]
<motion>
FIG. 23 is a flowchart showing the operation of the fourth embodiment.
The description of the input unit 10 and the output unit 40 that perform the same operations as those of the first to third examples will be omitted. That is, in FIG. 23, the description of step S10 and step S40 is omitted.
[0056]
[Operation of usage identifying unit 53]
The use identifying unit 53 acquires the document image stored in the document image storage unit 100 by the input unit 10, and identifies the type of use of the acquired document image. Then, importance level assignment target information (only the usage type) describing the identified usage type is created, and the created importance level assignment target information (only the usage type) is stored in the importance level assignment target information storage unit 200 (step S53). ).
FIG. 24 is an explanatory diagram illustrating an example of the document image in the specific example 4.
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is a dissertation image shown in FIG. 24 (this is a case where the usage type of “internal publication” is described in the upper left corner of the dissertation image shown in FIG. 3). The usage identifying unit 53 extracts the usage type of “internal publication” described in the upper left corner of the paper image, and identifies the usage type of the paper image. Then, importance level assignment target information (only usage type) describing the identified usage type is created, and the created importance level assignment target information (only usage type) is stored in the importance level assignment target information storage unit 200.
[0057]
FIG. 25 is an explanatory diagram illustrating an example of the importance assignment target information (use type only).
Hereinafter, the extraction of the usage type will be described.
The use identifying unit 53 first extracts a set of black pixels by examining the connectivity of black pixels near the upper left end in the paper image. If the length and width of the circumscribed rectangle of the extracted black pixel set is greater than an arbitrary value, the extracted black pixel set is determined to be a component whose component type is “frame”, and the extracted black pixel set The circumscribed rectangle is extracted as a “frame area”. On the other hand, if the height and width of the circumscribed rectangle of the extracted black pixel set is smaller than an arbitrary value, the extracted black pixel set is determined to be a component having the component type “character”, and the extracted black pixel The circumscribed rectangle of the set is extracted as a “character area”. If the components of the “character” are arranged in the horizontal direction, and the horizontal distance between two adjacent components of the “character” is smaller than an arbitrary value, the configuration of the “character” arranged in the horizontal direction The element set is determined to be a component whose component type is “character string”, and a circumscribed rectangle of the component set of “character” is extracted as a “character string area”. Further, a character recognition process is performed on a component set of “character”, which is a component of “character string”, and the obtained character recognition result (“internal publication”) is extracted as a usage type.
[0058]
[Operation of item extraction unit 23]
The item extraction unit 23 acquires the document image stored in the document image storage unit 100 by the input unit 10 and, for all or some of the items in the acquired document image, an item type and an item area (circumscribed rectangle of the item). ) And the item contents (character recognition result of the item) are extracted. An item is a unit that forms the content of a document. Further, the item extraction unit 23 acquires the importance level assignment target information (only the usage type) stored in the importance level assignment target information storage unit 200 by the usage identification unit 53, and uses the usage type and the extracted item type, item area, and The importance assignment target information describing the item contents is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200 (step S23).
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is the article image shown in FIG. 24, the item extracting unit 23 determines whether all or some of the items in the article image shown in FIG. Item types such as “body”, “section title”, “figure”, “caption”, etc., item areas such as “body area”, “section title area”, “figure area”, “caption area” (the coordinate values of the circumscribed rectangle of the item) and “ Item contents (character recognition results of items) such as body contents, section title contents, and caption contents are extracted. However, the item contents are extracted only for the items whose item type is determined from the components whose component type is “text” shown in FIG. That is, the item types are only items such as “text”, “section title”, and “caption”. Further, the item extraction unit 23 acquires the importance assignment target information (only the use type) stored in the importance assignment target information storage unit 200 by the use identification unit 53, and extracts the use type of “internal publication” and the extracted “ Item types such as "body", "section title", "figure", "caption", etc., item areas such as "body area", "section title area", "figure area", "caption area", and "body content", "section title content", "caption" The importance assignment target information that describes item contents such as “contents” is created, and the created importance assignment target information is stored in the importance assignment target information storage unit 200.
[0059]
FIG. 26 is an explanatory diagram illustrating an example of importance assignment target information.
In FIG. 26, the item area is indicated by the circumscribed rectangle "(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)" of the item. The item type, the item area, and the item content are arranged in order from top to bottom in the left column.
Item types such as "Body", "Section title", "Figure", "Caption", etc., item areas such as "Body area", "Section title area", "Figure area", "Caption area", and "Body contents", "Section title contents", " The extraction of item contents such as "caption contents" is the same as in the first embodiment.
[0060]
[Operation of importance assigning section 33]
The importance level assigning unit 33 acquires the importance level assigning target information stored in the importance level assigning information storage unit 200 by the item extracting unit 23, and further obtains the importance level assigning level stored in advance in the importance level assigning reference holding unit 300. Get the criteria.
FIG. 27 is an explanatory diagram illustrating an example of the importance assignment criterion.
Then, based on the use type, the item type, and the correspondence between the item contents and the importance described in the importance provision criterion, the importance assigning unit 33 describes the importance type in which the use type, the item type, the item area, and the importance are described. Assignment target information (including importance) is created, and the created importance assignment target information (including importance) is stored in the importance assignment target information storage unit 200 (step S33).
FIG. 28 is an explanatory diagram illustrating an example of importance assignment target information (including importance).
For example, the importance assignment target information stored in the importance assignment target information storage unit 200 by the item extraction unit 23 is the importance assignment target information shown in FIG. In the case where the importance assignment criterion is the importance assignment criterion shown in FIG. 27, the importance assignment section 33 writes the usage type such as “internal publication” or “no designation” described in the importance assignment criterion shown in FIG. Item types such as “section title”, “figure”, “caption”, etc., and item contents such as “including character string“ OCR ”” and “unspecified” and “high”, “low”, “medium”, “low”, “high”, “high” Based on the correspondence of the degrees of importance such as "", the degree of importance provision target information (including the degree of importance) shown in FIG. (Including importance) is stored in the importance assignment target information storage unit 200. To.
[0061]
In the above description of the operation, a word such as “internal publication” is used as the usage type, but any word may be used as long as it is easily understood that the usage type is used. Further, although the above-described method is used for extracting the usage type, any method may be used as long as the usage type can be extracted.
Further, although a word such as "no designation" is used as the usage type of the importance assignment criterion, any word may be used as long as it is easily understood that the usage type is not specified. Further, “unspecified” may be used for all application types.
Other uses are the same as in the first embodiment.
[0062]
<effect>
As described above, according to the specific example 4, in addition to the configuration of the specific example 1, a usage type is extracted based on a document format, and the importance of each item is given in consideration of the usage type. Therefore, the following effects are obtained.
That is, importance is automatically assigned to an item according to the type of use and the content of the item, which is a unit forming the content of the document. It is possible to keep high “body”, “figure” and “caption” private, or to keep only “body” with important keywords in addition to “figure” and “caption” that are highly important for other uses. Yes (a "body" that has been kept private for one use can be made public for another use). That is, the disclosure / non-disclosure of the item according to the degree of importance can be adjusted for each use type and item content. Compared to the specific example 1, the document can be disclosed in a more detailed manner, since the adjustment can be made for each application type.
[0063]
<< Specific Example 5 >>
In Example 5, a use type indicating the purpose of use of the document and a document type indicating what kind of document the document is obtained from the image information in the document. The importance is given in consideration of the document type and the document type.
[0064]
<Constitution>
FIG. 29 is a configuration diagram of the specific example 5.
The illustrated image processing device 1d is connected to the input device 2 and the storage device 3, and has an input unit 10, a use / document identification unit 54, an item extraction unit 24, an importance assignment unit 34, an output unit 40, and a document image storage. It comprises a unit 100, an importance assignment target information storage unit 200, and an importance assignment reference holding unit 300.
Here, the input unit 10, the output unit 40, the document image storage unit 100, the importance assignment target information storage unit 200, and the importance assignment reference holding unit 300 are the same as those in the specific examples 1 to 4, and will be described here. Is omitted. However, the importance assignment criterion of the importance assignment criterion holding unit 300 of the specific example 5 is data of importance including the document type and the use type in addition to the item type and the item content.
[0065]
The use / document identification unit 54, from the document image in a predetermined format stored in the document image storage unit 100, based on the format of this document, a use type indicating what use the document is used for, This is a functional unit for identifying a document type indicating the type of document. The item extracting unit 24 has a function of performing the same processing as that of the item extracting unit 20 of the first specific example based on the use type / document type identification result of the use / document identification unit 54. Further, the importance assigning unit 34 stores the usage type / document type identified by the application / document identifying unit 54 and the item type and item content extracted by the item extracting unit 22 in the importance assigning criterion holding unit 300. This is a functional unit that assigns importance to each item as compared with the importance assignment reference.
[0066]
<motion>
FIG. 30 is a flowchart showing the operation of the specific example 5.
The description of the input unit 10 and the output unit 40 that perform the same operations as those of the first to fourth embodiments will be omitted. That is, in FIG. 30, the description of step S10 and step S40 is omitted.
[0067]
[Operation of Usage / Document Identification Unit 54]
The use / document identification unit 54 acquires the document image stored in the document image storage unit 100 by the input unit 10, and identifies the use type and the document type of the acquired document image. Then, importance assignment target information (only use / document type) describing the identified usage type and document type is created, and the created importance assignment target information (only use / document type) is stored in the importance assignment target information storage unit. 200 (step S54).
FIG. 31 is an explanatory diagram illustrating an example of a document image in the specific example 5.
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is a dissertation image shown in FIG. 31 (this is the usage type of “internal publication” in the upper left part of the dissertation image shown in FIG. The use / document identification unit 54 has a use type of “internal publication” and an upper right end described in the upper left corner of the paper image. The document type of “thesis A” described in the above is extracted, and the usage type and the document type of the thesis image are identified. Then, importance assignment target information (only use / document type) describing the identified usage type and document type is created, and the created importance assignment target information (only use / document type) is stored in the importance assignment target information storage unit. 200.
[0068]
FIG. 32 is an explanatory diagram illustrating an example of importance assignment target information (only use / document type).
Hereinafter, extraction of the usage type and the document type will be described.
The use / document identification unit 54 first extracts a set of black pixels by examining the connectivity of black pixels near the upper left corner in the dissertation image. If the length and width of the circumscribed rectangle of the extracted black pixel set is greater than an arbitrary value, the extracted black pixel set is determined to be a component whose component type is “frame”, and the extracted black pixel set The circumscribed rectangle is extracted as a “frame area”. On the other hand, if the height and width of the circumscribed rectangle of the extracted black pixel set is smaller than an arbitrary value, the extracted black pixel set is determined to be a component having the component type “character”, and the extracted black pixel The circumscribed rectangle of the set is extracted as a “character area”. If the components of the “character” are arranged in the horizontal direction, and the horizontal distance between two adjacent components of the “character” is smaller than an arbitrary value, the configuration of the “character” arranged in the horizontal direction The element set is determined to be a component whose component type is “character string”, and a circumscribed rectangle of the component set of “character” is extracted as a “character string area”. Further, a character recognition process is performed on a component set of “character”, which is a component of “character string”, and the obtained character recognition result (“internal publication”) is extracted as a usage type. Subsequently, a set of black pixels is extracted by examining the connectivity of the black pixels near the upper right end in the paper image, and the character recognition result (“paper A”) obtained by performing the same processing is used as the document type. Extract as
[0069]
[Operation of item extraction unit 24]
The item extraction unit 24 acquires the document image stored in the document image storage unit 100 by the input unit 10 and, for all or some of the items in the acquired document image, an item type and an item area (circumscribed rectangle of the item). ) And the item contents (character recognition result of the item) are extracted. An item is a unit that forms the content of a document. Further, the item extracting unit 24 acquires the importance level assignment target information (only the usage type / document type) stored in the importance level assignment target information storage unit 200 by the usage / document identification unit 54, and extracts the usage type, the document type, and the extraction. Then, importance level assignment target information describing the created item type, item area, and item content is created, and the created importance level assignment target information is stored in the importance level assignment target information storage unit 200 (step S24).
For example, when the document image stored in the document image storage unit 100 by the input unit 10 is the article image shown in FIG. 31, the item extracting unit 24 determines whether all or some of the items in the article image shown in FIG. Item types such as “body”, “section title”, “figure”, “caption”, etc .; Item contents (character recognition results of items) such as body contents, section title contents, and caption contents are extracted. However, the item contents are extracted only for the items whose item type is determined from the components whose component type is “text” shown in FIG. That is, the item types are only items such as “text”, “section title”, and “caption”. Further, the item extraction unit 24 acquires the importance assignment target information (only the use / document type) shown in FIG. 32 stored in the importance assignment target information storage unit 200 by the use / document identification unit 54, and outputs the “internal disclosure”. ", The document type of" Thesis A "and the extracted item types such as" Body "," Section title "," Figure ", and" Caption "," Body area "," Section title area "," Figure area ", and" Caption area ". And the like, and item contents such as "text contents", "section title contents", and "caption contents" are created, and the created importance assignment information is stored in the importance assignment information storage unit 200. To be stored.
[0070]
FIG. 33 is an explanatory diagram illustrating an example of the importance level assignment target information in the specific example 5. In FIG. 33, the item area is indicated by the circumscribed rectangle "(upper left X coordinate, upper left Y coordinate, lower right X coordinate, lower right Y coordinate)" of the item. The item type, the item area, and the item content are arranged in order from top to bottom in the left column.
Item types such as "Body", "Section title", "Figure", "Caption", etc., item areas such as "Body area", "Section title area", "Figure area", "Caption area", and "Body contents", "Section title contents", " The extraction of item contents such as "caption contents" is the same as in the first embodiment.
[0071]
[Operation of Importance Assigning Unit 34]
The importance level assigning unit 34 acquires the importance level assigning target information stored in the importance level assigning target information storage unit 200 by the item extracting unit 24, and further acquires the importance level assigning level stored in the importance level assigning reference holding unit 300 in advance. Get the criteria.
FIG. 34 is an explanatory diagram illustrating an example of the importance assignment criterion.
Then, the importance assigning unit 34 determines the use type, the document type, the item type, the item area, the importance, The importance assignment target information (including the importance) describing the degree is created, and the created importance assignment target information (including the importance) is stored in the importance assignment target information storage unit 200 (step S34).
FIG. 35 is an explanatory diagram illustrating an example of importance assignment target information (including importance).
For example, the importance assignment target information stored in the importance assignment target information storage unit 200 by the item extraction unit 24 is the importance assignment target information shown in FIG. In the case where the importance assignment criterion is the importance assignment criterion shown in FIG. 34, the importance assignment section 34 sets the usage type such as “internal publication” or “no designation” described in the importance assignment criterion shown in FIG. , "No designation", etc., item types such as "Body", "Section title", "Figure", "Caption", etc., and item contents such as "including character string" OCR "", "No designation" and "High", "High" Based on the correspondence of importance such as “low”, “medium”, “high”, “low”, “high”, “high”, etc., the importance shown in FIG. 35 describing the use type, document type, item type, item area, and importance Create assignment target information (including importance) and create the importance assignment target information Storing (including severity) the importance grantee information storage unit 200.
[0072]
In the above description of the operation, a word such as “internal publication” is used as the usage type, but any word may be used as long as it is easily understood that the usage type is used. Although a word such as "article A" is used as the document type, any word may be used as long as the word can be easily understood as the document type.
Although the above method is used for extracting the use type and the document type, any method may be used as long as the use type and the document type can be extracted.
Although a word such as "unspecified" is used as the usage type of the importance assignment criterion, any word may be used as long as it is easily understood that the usage type is not specified. Further, “unspecified” may be used for all application types.
Although a word such as "no designation" is used as the document type of the importance assignment criterion, any word may be used as long as it is easily understood that the document type is not specified. Further, “not specified” may be used for all document types.
Other uses are the same as in the first embodiment.
[0073]
<effect>
As described above, according to the specific example 5, in addition to the configuration of the specific example 1, the application type and the document type are extracted based on the document format, and the importance of each item is also taken into consideration in consideration of the application type and the document type. Has the following effects.
That is, since the items are automatically assigned importance in accordance with the type of use, the type of document, and the content of the item which is a unit forming the content of the document, the document is disclosed to the company (use type is ""Internalpublication"), keep the relatively important "text", "figure", and "caption" in one document private, or add the important "figure", "caption" It is possible to keep only the “body” having a certain keyword private (the “body” kept private in one document can be made public in another document).
Further, it is possible to adjust the disclosure / non-disclosure when the document is disclosed outside the company (the use type is “external publication”). That is, the disclosure / non-disclosure of the item according to the degree of importance can be adjusted for each use type, document type, and item content. Compared with the specific examples 3 and 4, since the adjustment can be made for each use type and document type, more detailed disclosure of a document can be realized.
[0074]
《Usage form》
In each of the above specific examples, a document in a format as shown in FIG. 3, FIG. 17, FIG. 24, and FIG. 31, for example, is targeted as a document in a predetermined format, but a document in another format may be used. That is, if the format of the document is known in advance, by extracting each item area corresponding to the format, processing can be performed in the same manner as the above specific examples.
In the specific examples 3, 4, and 5, the configuration of the document identification and the use identification is added to the configuration of the specific example 1. However, these specific examples are applied to the configuration of the specific example 2. You may. In this case, it can be realized by replacing the item extraction units 22, 23, and 24 of the specific examples 3, 4, and 5 with the logical structure extraction unit 21 of the specific example 2.
[0075]
In each of the above specific examples, by assigning importance to each item, the document is used for disclosure / non-disclosure of the document, but can be applied to various other uses. For example, it may be used as a high or low security level such that a digital watermark is added to a portion with high importance. Further, the color level (increase / decrease) of the document data, the compression level (compression / non-compression), the resolution level (high resolution / low resolution), the omission level (omitted / omitted), and the summary level depend on the difference in importance. The present invention can be applied to processing corresponding to various levels such as (summarize / not summarize).
[0076]
Also, as a target document, in school relations, it can be applied to an in-house report, a grade sheet, a notification slip, and the like. In medical institutions, it can be applied to medical records, prescriptions, and the like. Further, in a company, it can be applied to confidential documents, specifications, source codes, circuit diagrams, and the like. Further, in a financial institution or the like, it can be applied to various documents such as an application form.
[Brief description of the drawings]
FIG. 1 is a configuration diagram illustrating a specific example 1 of an image processing apparatus according to the present invention.
FIG. 2 is a flowchart illustrating an operation of a specific example 1.
FIG. 3 is an explanatory diagram illustrating an example of a document image according to a specific example 1.
FIG. 4 is an explanatory diagram illustrating an example of a component of a specific example 1.
FIG. 5 is an explanatory diagram showing an example of items of a specific example 1.
FIG. 6 is an explanatory diagram showing an example of importance level assignment target information of a specific example 1.
FIG. 7 is an explanatory diagram illustrating an example of an importance assignment criterion according to a specific example 1.
FIG. 8 is an explanatory diagram showing an example of importance assignment target information (including importance) of the first specific example.
FIG. 9 is a configuration diagram of a specific example 2.
FIG. 10 is a flowchart illustrating an operation of a specific example 2;
FIG. 11 is an explanatory diagram illustrating an example of a logical structure.
FIG. 12 is an explanatory diagram showing an example of importance level assignment target information of a specific example 2.
FIG. 13 is an explanatory diagram illustrating an example of an importance assignment criterion according to a specific example 2.
FIG. 14 is an explanatory diagram showing an example of importance assignment target information (including importance) in Specific Example 2.
FIG. 15 is a configuration diagram of a specific example 3.
FIG. 16 is a flowchart illustrating an operation of a specific example 3.
FIG. 17 is an explanatory diagram showing an example of a document image in specific example 3.
FIG. 18 is an explanatory diagram showing an example of importance assignment target information (only a document type) in a specific example 3.
FIG. 19 is an explanatory diagram showing an example of importance level assignment target information in specific example 3.
FIG. 20 is an explanatory diagram illustrating an example of an importance assignment criterion in a specific example 3.
FIG. 21 is an explanatory diagram showing an example of importance assignment target information (including importance) in Specific Example 3.
FIG. 22 is a configuration diagram of a specific example 4.
FIG. 23 is a flowchart showing an operation of a specific example 4.
FIG. 24 is an explanatory diagram illustrating an example of a document image in a specific example 4.
FIG. 25 is an explanatory diagram showing an example of importance assignment target information (only a use type) in a specific example 4.
FIG. 26 is an explanatory diagram illustrating an example of importance level assignment target information in a specific example 4.
FIG. 27 is an explanatory diagram illustrating an example of an importance level assignment criterion in a specific example 4.
FIG. 28 is an explanatory diagram showing an example of importance assignment target information (including importance) in Specific Example 4.
FIG. 29 is a configuration diagram of a specific example 5;
FIG. 30 is a flowchart showing an operation of a specific example 5;
FIG. 31 is an explanatory diagram showing an example of a document image in specific example 5.
FIG. 32 is an explanatory diagram showing an example of importance level assignment target information (only use / document type) in specific example 5;
FIG. 33 is an explanatory diagram showing an example of importance level assignment target information in a specific example 5.
FIG. 34 is an explanatory diagram illustrating an example of an importance assignment criterion in a specific example 5.
FIG. 35 is an explanatory diagram showing an example of importance assignment target information (including importance) in Specific Example 5;
[Explanation of symbols]
20, 22, 23, 24 Item extraction unit
21 Logical structure extraction unit
30, 31, 32, 33, 34 Importance assigning unit
100 Document image storage
200 Importance level information storage
300 Holder for assigning importance level

Claims (10)

For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and what kind of the item area is And an item extraction unit that extracts an item type indicating the item type indicating the content of the item area,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between the item type and the importance of the item content,
An importance assigning unit that assigns an importance to each item by comparing the item type and the item content extracted by the item extracting unit with an importance assigning criterion of the importance assigning criterion holding unit; An image processing apparatus characterized by the above-mentioned. For a document in a predetermined format, an item area serving as a unit for forming the content of the document is extracted from an image of the document based on the format of the document, and the item area is used as another item in the document. A logic for obtaining a logical structure indicating how the logical structure is related to the area, and extracting a logical structure type indicating the type of the logical structure and a logical structure content indicating the contents of the logical structure A structure extraction unit, and in advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between the logical structure type and the importance of the logical structure content,
The logical structure type and the logical structure content extracted by the logical structure extracting unit are compared with the importance assigning criterion of the importance assigning criterion holding unit, and an importance assigning unit that assigns an importance to each logical structure. An image processing apparatus comprising: For a document in a predetermined format, a document identification unit that identifies a document type indicating what kind of document the document is based on the format of the document from an image of the document,
For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and what kind of the item area is And an item extraction unit that extracts an item type indicating the item type indicating the content of the item area,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between the document type, the item type, and the item content, and
The document type identified by the document identification unit, and the item type and item content extracted by the item extraction unit are compared with the importance assignment criterion of the importance assignment criterion holding unit, and important items for each item are compared. An image processing apparatus comprising: an importance assigning unit that assigns a degree. For a document in a predetermined format, a document identification unit that identifies a document type indicating what kind of document the document is based on the format of the document from an image of the document,
For a document in a predetermined format, an item area serving as a unit for forming the content of the document is extracted from an image of the document based on the format of the document, and the item area is used as another item in the document. A logic for obtaining a logical structure indicating how the logical structure is related to the area, and extracting a logical structure type indicating the type of the logical structure and a logical structure content indicating the contents of the logical structure A structure extraction unit,
An importance assignment reference holding unit that holds an importance assignment reference indicating a relationship between the document type, the logical structure type, and the logical structure content in advance;
The document type identified by the document identification unit, the logical structure type and the logical structure content extracted by the logical structure extraction unit are compared with the importance assignment criterion of the importance assignment criterion holding unit, and each logical An image processing apparatus comprising: an importance assigning unit that assigns an importance to each structure. For a document of a predetermined format, from a picture of the document, a use identification unit that identifies a use type indicating what kind of use the document is used based on the format of the document,
For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and what kind of the item area is And an item extraction unit that extracts an item type indicating the item type indicating the content of the item area,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between importance levels for use types, item types, and item contents,
The application type identified by the application identification unit and the item type and item content extracted by the item extraction unit are compared with the importance assignment criterion of the importance assignment criterion holding unit, and the importance is assigned to each item. An image processing apparatus comprising: an importance assigning unit that assigns a degree. For a document of a predetermined format, from a picture of the document, a use identification unit that identifies a use type indicating what kind of use the document is used based on the format of the document,
For a document in a predetermined format, an item area serving as a unit for forming the content of the document is extracted from an image of the document based on the format of the document, and the item area is used as another item in the document. A logic for obtaining a logical structure indicating how the logical structure is related to the area, and extracting a logical structure type indicating the type of the logical structure and a logical structure content indicating the contents of the logical structure A structure extraction unit,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between importance levels for use types, logical structure types, and logical structure contents,
The use type identified by the use identification unit, the logical structure type and the logical structure content extracted by the logical structure extraction unit are compared with the importance assignment reference of the importance assignment reference holding unit, and each logical An image processing apparatus comprising: an importance assigning unit that assigns an importance to each structure. For a document in a predetermined format, from the image of the document, a use type indicating what use the document is used based on the format of the document, and a document indicating what the document is A use / document identification unit for identifying the type;
For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and what kind of the item area is And an item extraction unit that extracts an item type indicating the item type indicating the content of the item area,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between importance levels for use types, document types, item types, and item contents,
The use type and document type identified by the use / document identification unit, and the item type and item content extracted by the item extraction unit are compared with the importance assignment criterion of the importance assignment criterion holding unit, An image processing apparatus comprising: an importance assigning unit that assigns an importance to each item. For a document in a predetermined format, from the image of the document, a use type indicating what use the document is used based on the format of the document, and a document indicating what the document is A use / document identification unit for identifying the type;
For a document in a predetermined format, an item area serving as a unit for forming the content of the document is extracted from an image of the document based on the format of the document, and the item area is used as another item in the document. A logic for obtaining a logical structure indicating how the logical structure is related to the area, and extracting a logical structure type indicating the type of the logical structure and a logical structure content indicating the contents of the logical structure A structure extraction unit,
In advance, an importance assignment criterion holding unit that holds an importance assignment criterion indicating a relationship between importance levels for use types, document types, logical structure types, and logical structure contents,
The use type and document type identified by the use / document identification unit are compared with the logical structure type and logical structure content extracted by the logical structure extraction unit, and the importance assignment standard of the importance assignment standard holding unit is compared. And an importance assigning unit that assigns an importance to each logical structure. For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and the type of the item area is determined. And an item type indicating the content of the item area are extracted,
Next, the item type and the item content are compared in advance with an importance assignment criterion indicating the relationship between the item type and the item content and the importance is assigned to each item. Processing method. A computer constituting the image processing apparatus,
For a document of a predetermined format, an item area that is a unit forming the content of the document is extracted from an image of the document based on the format of the document, and what kind of the item area is And an item extraction unit that extracts an item type indicating the item type indicating the content of the item area,
The item type and item content extracted by the item extraction unit are compared in advance with an importance assignment criterion indicating the relationship between the item type and the item content, and the importance is assigned to each item. A program for controlling an image processing apparatus to function as an assigning unit.

Images