JPH10269310A - Character recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To shorten processing time without increasing the burden of an operator by correcting a format to another format when there are plural formats that have similar characteristics and also when an error occurs in an identification result due to various factors such as the deterioration of an image quality at the time of identifying a format by using an image characteristic of a ruled line, etc. SOLUTION: A layout analyzing part 3 analyzes a document layout from the image data of a document that is acquired by an image inputting part 1 and sorts elements that constitute the document into each different item to be recognized, and a character recognizing part 5 performs character recognition of each sorted item. A slip identifying part 6 identifies a slip format from the layout analyiis result and character recognition result and makes a displaying part 8 show the identification result. An operator decides the displayed content of the displaying part and when there is an error in the identified format, a correcting part 10 corrects it to another format.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device, and more particularly to a character recognition device in an image processing system in which various kinds of image data are input in a mixed manner.

[0002]

2. Description of the Related Art Various kinds of image data are inputted, and OCR is performed.
2. Description of the Related Art In a so-called image processing system in which reading is performed by an (optical character reader) or filing is performed by a filing system, it is necessary to classify and process various kinds of input image data.
In order to perform this classification, the operator classifies the form in advance, adds a sheet such as a batch form for identifying the form, adds an ID for classification to the form, and recently, A method of automatically specifying a format based on features such as ruled lines is employed.

[0003]

However, in the case of the above-mentioned methods, the work for classification is required in advance, and the work is imposed on the operator, so that the burden on the operator increases. Will be.
In addition, in the method of automatically identifying formats using image features such as ruled lines, there are a plurality of formats having similar features, and the identification result may be incorrect due to various factors such as deterioration of image quality. May occur.

The present invention has been made in view of the above problems, and has as its object to provide a character recognition apparatus capable of correcting an error in a recognition result without increasing the burden on an operator. It is to provide a device.

[0005]

According to the present invention, there is provided a character recognition apparatus comprising: an image input unit for obtaining an image of a document; a layout analysis unit for analyzing a layout of the document obtained by the image input unit; A character recognition unit that performs character recognition for each item of the layout, a form identification unit that identifies a form format from the layout analysis result and the character recognition result, and a display unit that displays the form format identified by the form identification unit. And a correction unit for correcting the form format displayed on the display unit.

In the character recognition device having the above configuration, the layout analysis section analyzes the layout of the document from the image data of the document acquired by the image input section, separates the elements constituting the document into items to be recognized, and performs character recognition. The unit performs character recognition for each of the separated items. The form identification unit identifies the form format from the layout analysis result and the character recognition result, and displays the identification result on the display unit. Then, the operator determines the display content of the display unit, and if there is an error in the identified format,
The correction section corrects to another format.

Another character recognition device according to the present invention includes an image input unit for acquiring an image of a document, a layout analysis unit for analyzing a layout of the document acquired by the image input unit, and each of the layouts. A character recognition unit that performs character recognition of the item, an item search unit that determines the content of each item of the layout, and associates each item, and a display unit that displays a result of association of each item by the item search unit. The configuration includes a correction unit for correcting the association of each item displayed on the display unit.

In another character recognition apparatus having the above structure, a layout analysis unit analyzes the layout of the document from the image data of the document acquired by the image input unit, and separates the elements constituting the document into items to be recognized. The character recognition unit performs character recognition on each of the separated items. The item search unit determines the content of each item of the layout of each element whose character has been recognized, associates each item, and displays the result on the display unit. Then, the operator determines the display content of the display unit, and if the association of each item is incorrect, the correction unit corrects the association.

[0009]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing a first embodiment of the present invention. In FIG. 1, the character recognition device according to the present embodiment includes an image input unit 1, an image data holding unit 2, a layout analysis unit 3, a layout analysis result holding unit 4,
It comprises a character recognition unit 5, a form identification unit 6, a character recognition result holding unit 7, a display unit 8, a format data holding unit 9, a correction unit 10, and an identification result holding unit 11, The elements are interconnected via a bus line 12.

Here, the image input unit 1 is for obtaining an image of a document, and is composed of an image scanner, a video camera, an image file and the like. The image data holding unit 2 is for holding the image data acquired by the image input unit 1. As the image data holding unit 2, a storage medium such as a memory or a hard disk is used.

The layout analysis unit 3 analyzes the layout of the image obtained by the image input unit 1 and separates the elements constituting the document into items to be recognized, for example, an address column and a date of birth column. The layout analysis result holding unit 4 is for holding the analysis result of the layout analysis unit 3. As the layout analysis result holding unit 4, a storage medium such as a memory and a hard disk is used.

The character recognizing unit 5 performs character recognition of each item separated by the layout analyzing unit 3. The form identification unit 6 determines the content of each item of the layout of each element recognized by the character recognition unit 5, and associates each item. The character recognition result holding unit 7 is for holding character data recognized by the character recognition unit 5. As the character recognition result holding unit 7, a storage medium such as a memory and a hard disk is used.

The format data holding unit 9 holds the characteristics of a form for identifying a format. As the format data holding unit 9, a storage medium such as a memory and a hard disk is used. The identification result holding unit 11 holds a result of identifying a format and a result of performing character recognition. As the identification result holding unit 11, a storage medium such as a memory and a hard disk is used.

The display unit 8 displays the result of the format identification, and the contents such as the layout analysis result and the character recognition result as required. The correction unit 10 causes the display unit 8 to display the format identification result, and, based on the display contents, determines that the format is not correct, and when the operator inputs that fact, the format data holding unit 9.
Display section 8 displays information about the format held in
In order to correct the format by a correction operation by the operator.

The storage medium constituting the image data holding unit 2, the layout analysis result holding unit 4, the character recognition result holding unit 7, the format data holding unit 9, and the identification result holding unit 11 is separately provided for each component. May be used, but it is more cost-effective to use a single storage medium in common and allocate and use the storage area for each component.

Next, the operation of character recognition in the character recognition device according to the first embodiment having the above configuration will be described. FIG. 2 shows an example in which the format identification result is modified.

First, the image data obtained by the image input unit 1 is held in the image data holding unit 2. The stored image data is analyzed for features such as ruled lines by a layout analysis unit 3 and stored in a layout analysis result storage unit 4. In the layout analysis result, the image data is subdivided into items for character recognition. Then, each of the subdivided items is subjected to character recognition by the character recognition unit 5, and the result is stored in the character recognition result storage unit 7.

The form identification unit 6 selects a corresponding format from the format data stored in the format data holding unit 9 by using the layout identification result and the character recognition result, and sends the selected format to the identification result holding unit 11. Store. The identification result is displayed on the display unit 8 so that the operator can confirm it. At this time, the display unit 8
, The contents such as the layout analysis result and the character recognition result are also displayed as necessary.

When the operator looks at the identification result displayed on the display unit 8 and determines that the format is different from the expected one, the information to that effect is given to the correction unit 10. Then, receiving the information, the correction unit 10 causes the display unit 8 to display a list of the format data stored in the format data holding unit 9. The operator looks at this list of format data and selects the correct format from the list.

As a result of the above-described processing, as an example, in FIG. 3, what was initially identified as format 1 is changed to format 2 by format correction by the operator. At this time, item D 'is item C'''''.

As described above, in the image processing system in which various types of image data are input in a mixed manner, the correction unit 10 for changing the form format displayed on the display unit 8 to another format is provided in the character recognition device. In this way, if there is an error in the identified format, the operator can correct it to the correct format with a simple operation of simply selecting the desired format, and perform operations such as restarting the identification process from the beginning Is no longer required, so that the burden on the operator can be reduced.

FIG. 4 is a block diagram showing a second embodiment of the present invention. In FIG. 4, a character recognition device according to the present embodiment includes an image input unit 21 and an image data holding unit 22.
Form information holding unit 23, layout analysis unit 24,
A layout analysis result holding unit 25, a character recognition unit 26,
It comprises a character recognition result holding unit 27, an item search unit 28, an item search result holding unit 29, a display unit 30, and a correction unit 31, and these components are mutually connected via a bus line 32. ing.

Here, the image input section 21 is for acquiring an image of a document, and is composed of an image scanner, a video camera, an image file and the like. The image data holding unit 22 holds the image data acquired by the image input unit 21. As the image data holding unit 22, a storage medium such as a memory and a hard disk is used.

The form information holding unit 23 is for holding information such as item names that are considered to be included in a form specified in advance. As the form information holding unit 23, a storage medium such as a memory and a hard disk is used. The layout analysis unit 24 analyzes the layout of the image acquired by the image input unit 21 and converts the elements constituting the document into
For example, it is divided into items to be recognized, such as an address column and a date of birth column.

The layout analysis result holding section 25 is for holding the analysis result of the layout analysis section 24. As the layout analysis result holding unit 25, a storage medium such as a memory and a hard disk is used. The character recognizing unit 26 performs character recognition of each item separated by the layout analyzing unit 24. Character recognition result holding unit 27
Is for holding the character data recognized by the character recognition unit 26. As the character recognition result holding unit 27, a storage medium such as a memory and a hard disk is used.

The item search unit 28 determines the contents of each item of the layout of each element recognized by the character recognition unit 26 and associates each item. The item search result holding unit 29 holds the result of the item search performed by the item search unit 28. As the item search result holding unit 29, a storage medium such as a memory and a hard disk is used.

The display unit 30 displays the results of the item search, and the contents such as the layout analysis result and the character recognition result as required. The correction unit 31 causes the display unit 30 to display the search result of the item, and, based on the displayed content, determines that the item is not correct. This is for modifying the item search result.

The storage media constituting the image data holding unit 22, the form information holding unit 23, the layout analysis result holding unit 25, the character recognition result holding unit 27, and the item search result holding unit 29 are as follows. Although separate storage media may be used, it is more cost-effective to use a single storage medium in common and to allocate and use storage areas for each component.

Next, the operation of character recognition in the character recognition device according to the second embodiment having the above configuration will be described.

First, in FIG. 2 showing a form layout to be subjected to character recognition, the components of the form layout are roughly divided into items (number, product name, amount, total) and item entry columns. The item entry column is a portion for entering the content corresponding to the item. The item and the item entry column are associated with each other. In the figure, hatched portions are numbers and their entry columns. The number entry field has a number attribute such as “four digits” depending on the association. This attribute is included in the form information.

FIG. 5 shows an example in which the item search result is modified. First, the image data acquired by the image input unit 21 is held in the image data holding unit 22. The stored image data is analyzed for features such as ruled lines by a layout analysis unit 24 and stored in a layout analysis result storage unit 25. In the layout analysis result, the image data is subdivided into items for character recognition. Each of the subdivided items is subjected to character recognition by the character recognition unit 26, and the result is stored in a character recognition result holding unit 27.
Is held.

The item search unit 28 is a layout analysis unit 24
Using the layout analysis result by the character recognition unit 26 and the character recognition result by the character recognition unit 26, the item names stored in the form information storage unit 23 are assigned to the respective item columns and associated with the item entry columns, and the results are stored in the item search results. It is stored in the unit 29. The item search result by the item search unit 28 is displayed on the display unit 30 so that the operator can confirm it.

At this time, the display unit 30 displays
Contents such as a layout analysis result and a character recognition result are also displayed. Looking at the identification result displayed on the display unit 30, if the operator determines that the association between the item and the item entry column is different from the expected one, the operator gives correction information by, for example, clicking the mouse. Then, in response to this correction information, the correction unit 31 corrects the association between the item and the item entry field.

As a result of the processing described above, in FIG. 3, the item entry column at the lower right is initially associated with item C as the item entry column C "", but the item D is changed by the operator's correction operation. Has been changed to the item entry column D '.

As described above, in the image processing system in which various kinds of image data are input in a mixed manner, the display unit 30
The correction unit 31 for correcting the association between the item displayed in the item and the item entry field is provided in the character recognition device, so that when the association between the item and the item entry field is incorrect, the operator can easily perform the operation. To correct the association, and there is no need to perform an operation such as re-performing the identification process from the beginning, so that the burden on the operator can be reduced.

In this embodiment, the color of the display on the display unit 30 is not described, but the result of the association of each item may be displayed in different colors. For example, as shown in FIG. 6, different items are displayed in different colors, and the item entry fields associated with the items are displayed in the same color. In this manner, by displaying the association between the item and the item entry column in different colors, the association between the items can be clearly displayed to the operator, so that the association can be confirmed at a glance.

Further, the display unit 30 may be configured to display information of each item on an item column or an item entry column. For example, as shown in FIG. 7, the item name “number” is displayed above the item column (number), and “the“ number ”entry column 1, number, four digits” is displayed above the item entry column. , The information of each item is displayed. As described above, the information of each item is displayed above the item column or the item entry column and the association is clearly indicated to the operator, so that the result of the association can be confirmed at a glance.

Further, in the correction unit 31, the display unit 30
It is also possible to select an item field displayed by the above, display a list of item names from the form information, and select an item name from this list. That is, the correction unit 31 displays a list of candidate item names for the selected item according to an instruction of the operator. Then, the operator selects a correct item name from among them.

As described above, by selecting the item column displayed on the display unit 30, displaying a list of item names from the form information, and selecting the item name from this list, the recognition result of the item name is obtained. Even if incorrect, the item name can be easily changed or corrected. FIG. 8 shows an example of the display. In this display example, candidate item names are displayed with the item name “number” selected.

Further, when the attribute of the item column or the item entry column is changed in the correction section 31, it is also possible to perform the processing after character recognition again on the changed column. As a result, for example, when "article name" is changed to "amount" and "article name description" is changed to "amount description", hiragana is displayed in "amount". Can be prevented. The re-recognition may be automatic or manual according to an instruction of the operator. According to this,
Even when the attribute of the item column or the entry column is changed, a recognition result corresponding to the new attribute can be obtained.

Further, it is also possible to copy the attribute of a certain item column or item entry column to another item column or item entry column. FIG. 9 shows an example of attribute copying.
In FIG. 9, it has not been determined in the first result. However, since the third and fourth lines have the same format as the second line, the second line is selected, and the format is copied to the third and fourth lines.

At this time, the copying procedure is as follows:
The lines may or may not be simultaneous. Also, when the format is the same line, regardless of the operator's instruction,
Copying is also conceivable. In this example, the row is copied, but it may be a column or a single column. According to this, it is possible to easily change the attribute of the item column or the item entry column.

Although the character recognition device according to the first embodiment and the character recognition device according to the second embodiment have been described as being different from each other, the character recognition device to which both the first and second embodiments are applied is described. Of course, it may be. According to such a character recognition device, since both functions of the first and second embodiments can be satisfied at the same time, the effect is extremely large.

[0045]

As described in detail above, according to the character recognition apparatus of the present invention, in an image processing system to which various kinds of image data are input, the character recognition apparatus is identified from the layout analysis result of the document and the character recognition result. If there is an error in the format, it can be corrected to another format, so that it is possible to shift to the subsequent processing without performing operations such as re-performing the identification process from the beginning, so always give the correct recognition result. And the processing time can be reduced.

According to another character recognition device of the present invention, in an image processing system to which various types of image data are input, the contents of each item of the layout of each element whose character has been recognized are determined. By associating items and correcting the association if the association is incorrect, the process can proceed to subsequent processes without performing operations such as re-performing the identification process from the beginning. And the processing time can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a diagram illustrating an example in which a format identification result is modified.

FIG. 3 is an explanatory diagram for correcting a format identification result.

FIG. 4 is a block diagram showing a second embodiment of the present invention.

FIG. 5 is a diagram showing an example in which an item search result is modified.

FIG. 6 is an explanatory diagram of a first modification of the second embodiment.

FIG. 7 is an explanatory diagram of a second modification of the second embodiment.

FIG. 8 is an explanatory diagram of a third modification of the second embodiment.

FIG. 9 is an explanatory diagram of Modification 4 of the second embodiment.

[Explanation of symbols]

 1,21 Image input unit 3,24 Layout analysis unit 5,26 Character recognition unit 6 Form identification unit 8,30 Display unit 10,31 Correction unit 28 Item search unit

Claims (7)

[Claims] An image input unit for acquiring an image of the document; a layout analyzing unit for analyzing a layout of the document acquired by the image input unit; and a character recognition unit for performing character recognition of each item of the layout. A form identification unit for identifying a form format based on an analysis result by the layout analysis unit and a recognition result by the character recognition unit; a display unit for displaying a form format identified by the form identification unit; and the display unit. And a correction unit for correcting the form format displayed on the document. 2. An image input unit for acquiring an image of a document, a layout analyzing unit for analyzing a layout of the document acquired by the image input unit, and a character recognition unit for performing character recognition of each item of the layout An item search unit that determines the content of each item of the layout and associates each item; a display unit that displays a result of association of each item by the item search unit; and a display unit that is displayed on the display unit. A character recognition device comprising: a correction unit for correcting the association of each item. 3. The character recognition device according to claim 2, wherein the display unit displays the result of the association of each item by the item search unit in different colors. 4. The character recognition device according to claim 2, wherein the display unit displays a result of determining the content of each item by the item search unit on the item. 5. The correction unit selects one item displayed on the display unit, displays a list of item names registered in advance on the display unit, and selects from the list. 3. The character recognition device according to claim 2, wherein the content of the item is corrected by the following. 6. The character recognition according to claim 2, wherein when the attribute of the item column or the item entry column is changed by the correction unit, processing after the character recognition is performed again on the item whose attribute has been changed. apparatus. 7. The character recognition device according to claim 2, wherein the correction unit can copy an analysis result of a certain column, row, or column to another column, row, or column.