JP2008250817A - Business form automatically discriminating device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent a collation error of business forms by checking and collating features between a previously registered business form image and an actual business form to be sorted. <P>SOLUTION: A business form discrimination processing part 15 generates feature data indicating a feature of image data from the image data input by an image input processing part 14. Then, the business form discrimination processing part 15 extracts feature image data similar to the input image data based on the generated feature data and the feature data stored in a business form discrimination data storage memory 12. The business form discrimination processing part 15 specifies a feature quantity difference indicating quantity of a feature difference from each other in regions divided into a predetermined number based on the generated feature data and the feature data described in the extracted feature image data. A similarity degree display part 16 displays the feature quantity difference on each specified region. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an automatic form identification apparatus for inputting a wide variety of form images on an image basis and identifying the form format from the form images.

  In a so-called core business form image processing system that inputs a wide variety of form images, recognizes characters in the character recognition target area, and filing the images, the input form image is first adapted to the business. It is necessary to identify the type of form. There are several methods for identifying, but one of them is a ruled line feature identifying method. In this ruled line feature identification method, the ruled line feature of the form image to be identified is extracted, the feature is classified and registered in advance as the format of the image, and the ruled line of the form image is input when the form image is input. And the ruled line features of a plurality of registered formats are identified, and the form type is identified (see, for example, Patent Document 1).

By the way, the form is frequently revised so that a column is added or the layout and size of the table are changed as necessary. An existing form is also used when another new form is created.
JP-A-11-282956

  In the conventional form image processing system using the ruled line feature identification method described above, a form image is identified by comparing a pre-registered form image with a form image to be actually classified according to a predetermined criterion. However, there is a problem that collation mistakes such as a case where the same form is collated with the ruled line feature alone, but the case where the forms are different, and the opposite case occur.

  Accordingly, in view of the above-described problems, the present invention has an object to provide a form automatic identification device capable of preventing a matching error between a form image registered in advance and a form image to be actually classified. To do.

  In order to solve the above-mentioned problem, the invention described in claim 1 is characterized in that the feature data indicating the feature of the image data of the form, the feature image data storing the feature image data corresponding to the image data, and the image data of the form An image data input unit that inputs the feature data, a feature data generation unit that generates feature data indicating features of the image data from the image data input by the image data input unit, and feature data generated by the feature data generation unit And an extraction unit that extracts the feature image data similar to the image data input from the image data input unit based on the feature data stored in the feature data storage unit, and the feature data generation unit Is divided into a predetermined number based on the feature data generated by the feature data and the feature data described in the feature image data extracted by the extraction unit. A feature amount difference specifying unit that specifies a feature amount difference indicating an amount of difference between the features in the region, and a display unit that displays the feature amount difference for each region specified by the feature amount difference specifying unit. .

  According to this configuration, the feature amount difference specifying unit divides the data into a predetermined number based on the feature data generated by the feature data generation unit and the feature data described in the feature image data extracted by the extraction unit. Since the feature amount difference indicating the amount of the difference between the features in the selected area is specified, and the display unit displays the feature amount difference for each region specified in the feature amount difference specifying unit, the registered form Since the feature between the image and the form image to be actually classified can be confirmed and collated, a form collation error can be prevented.

  In the invention described in claim 2, in the invention described in claim 1, the display unit displays the feature amount difference for each region. The image data input from the image data input unit in the selected area and the image data included in the feature image data extracted by the extraction unit are enlarged and displayed.

  According to this configuration, in addition to displaying the feature amount difference for each region, when the region in the displayed feature amount difference is pressed, the display unit inputs from the image data input unit in the pressed region. Since the image of the image data and the image of the image data of the feature image data extracted by the extraction unit are enlarged and displayed, the form image registered in advance and the form image to be actually classified Since it is possible to check the characteristics between the two in detail and collate them, it is possible to further prevent the form from being collated.

  As described above, according to the present invention, since it is possible to check and collate the characteristics between a pre-registered form image and a form image to be actually classified, it is possible to prevent a form collation error. .

  Next, the best mode for carrying out the present invention will be described.

  FIG. 1 is a block diagram showing the configuration of a form automatic identification apparatus according to the present invention. As shown in FIG. 1, the form automatic identification device includes an input image memory 11, a form identification data storage memory 12, a form identification result storage memory 13, an image input processing unit 14, a form identification processing unit 15, The similarity display unit 16 is provided. Here, the form identification processing unit 15 has the following functions when a CPU (not shown) executes a control program stored in a ROM (not shown).

  The input image memory 11 is formed of a video RAM, and stores form image data input from the image input processing unit 14.

    The form identification data storage memory 12 is formed by a hard disk device, and stores data (hereinafter referred to as form identification data) indicating the characteristics of the form to be identified.

  The form identification data is data obtained by pairing the start position of each ruled line and the end position of the ruled line with image data. For example, as shown in FIG. 2, if there is a ruled line surrounded by vertical and horizontal lines as shown by reference C in the image data of a certain form, the form identification data in this form data is It consists of a set of coordinates Pa (Xa, Ya), coordinates Pb (Xb, Yb) at position b, coordinates Pc (Xc, Yc) at position c, and coordinates Pd (Xd, Yd) at position d (Pa, Pb, Pc) , Pd) and image data are paired.

  The form identification result storage memory 13 is formed of RAM, and the form identification data extracted by the form identification processing unit 15 from all the form identification data stored in the form identification data storage memory 12 as will be described later. Is stored.

  The image input processing unit 14 is formed by an image scanner such as a digital camera, inputs an input form image as image data (hereinafter referred to as input image data), and writes the input image data into the input image memory 11.

  The form identification processing unit 15 calculates the start position and end position for all ruled lines in the form from the input image data of the form stored in the input image memory 11, and starts and ends the calculated ruled lines. Ruled line feature data including positions is generated.

  Further, the form identification processing unit 15 generates the ruled line feature data previously generated for each area equally divided into nine parts as shown in FIG. 3 with respect to the form identification data stored in the form identification data storage memory 12. Based on the above, the difference between the feature amounts for each region (hereinafter referred to as the feature amount difference for each region) is determined as large, medium, small, and none, and the determined region-specific feature amount difference is stored in a RAM (not shown). To do. Here, a large feature amount difference for each region means that the number of ruled lines in the corresponding region is different even by one, and a medium feature amount difference means that the number of ruled lines is the same and the length of the ruled line. The difference between the regions is 50% or more, and the characteristic difference by region is small when the number of ruled lines is the same and the length difference between the ruled lines is less than 50% and 10% or more. “No difference in feature amount by region” means that the number of ruled lines is the same and the difference in the lengths of the ruled lines is less than 10% of each other. The length of the ruled line is calculated by the distance between the start position of the ruled line and the end position of the ruled line, and the number of ruled lines is the number of the calculated ruled line length.

  Further, as described above, the form identification processing unit 15 converts the form identification data stored in the form identification data storage memory 12 into the ruled line feature data generated previously for each equally divided area. Based on the area-specific feature quantity difference determined as large, medium, small, or none, the similarity of the form identification data to the ruled line feature data is calculated. For example, as a reference value for calculating similarity, an area having a large area-specific feature amount difference is set to “0”, an area-by-area feature quantity difference is set to “4”, and an area having a small area-specific feature amount difference is set to “4”. When “7” is set and an area having no area-specific feature difference difference is “10”, the form identification processing unit 15 determines the ruled line for the form identification data according to the total result of the values in all nine areas. The similarity to the feature data is calculated.

  In addition, the form identification processing unit 15 performs the above-described process, and the feature of the ruled line of the input image data input from the image input processing unit 14 is determined from the form identification data stored in the form identification data storage memory 12. Five form identification data are extracted from the most possessed form, and the extracted form identification data is stored in the form identification result storage memory 13.

  The similarity display unit 16 is formed of a liquid crystal display device and displays an image feature display screen shown in FIG. As shown in FIG. 4, the image feature display screen includes an input image display field 161 for displaying an input image, an extracted image display field 162 for displaying an image of the extracted form identification data, and the above-described nine items. When one of the nine areas displayed in the area-specific feature quantity difference display input field 163 for displaying the area-specific feature quantity difference display area 163 and the area-specific feature quantity difference display input field 163 is pressed, An input area image display field 164 for enlarging and displaying the input image data of this area and an extraction area image display field 165 for enlarging and displaying the form identification data of the area are provided.

  Next, the operation of the form automatic identification apparatus according to the embodiment of the present invention will be described.

  FIG. 5 is a flowchart showing the operation of the automatic form identification device according to the embodiment of the present invention. First, the image input processing unit 14 inputs an input form image as input image data, writes the input screen data in the input image memory 11, and the input image data is input to the form identification processing unit 15. A notification to that effect is input (step S1).

  When the notification is input from the image input processing unit 14, the form identification processing unit 15 reads the input image data stored in the input image memory 11, and the start position and end position for all ruled lines of this image data And ruled line feature data including the start positions and end positions of all the calculated ruled lines is generated (step S2).

  Next, the form identification processing unit 15 determines the length of the ruled line and the number of ruled lines in the vertical direction and the horizontal direction from the ruled line start position and the ruled line end position described in the ruled line feature data as shown in FIG. Identifies each area equally divided into nine equal parts. Thereafter, the form identification processing unit 15 reads out the form identification data stored in the form identification data storage memory 12, and determines the length of the ruled line and the ruled line from the ruled line start position and ruled line end position described in the form identification data. Is specified for each of the equally divided areas in the vertical and horizontal directions.

  Thereafter, the form identification processing unit 15 compares the read form identification data with the ruled line feature data generated previously for each of the equally divided nine areas, and determines the feature value difference for each area as large, medium, or small. Then, using the reference value for calculating the degree of similarity, the degree of similarity of the form identification data with respect to the ruled line feature data is calculated using the reference value for calculating the degree of similarity (step S3).

  The form identification processing unit 15 calculates the similarity to the ruled line feature data for all the form identification data stored in the form identification data storage memory 12 as described above, and then stores it in the form identification data storage memory 12. From the form identification data that has been entered, five pieces of form identification data having the features of the ruled lines of the input image data input from the image input processing unit 14 (from the highest total value to five) are extracted, The extracted form identification data is stored in the form identification result storage memory 13 from the one having the highest characteristic (step S4).

  Next, the form identification processing unit 15 reads out the input image data of the image input from the image input processing unit from the input image memory 11 and the form identification data having the highest feature amount from the form identification result storage memory 13. Is read. In addition, the form identification processing unit 15 reads the area-specific feature amount differences in the nine areas calculated previously from the RAM.

  After that, the form identification processing unit 15 displays the image feature display screen shown in FIG. 4 to be displayed on the similarity display unit 16 based on the read input image data, form identification data, and area-specific feature value difference values in the nine areas. Display data is generated, and this display data is output to the similarity display unit 16.

  When the display data is input from the form identification processing unit 15, the similarity display unit 16 displays an image feature display screen shown in FIG. At this time, each of the input area image display field 164 and the extraction area image display field 165 includes an image of input image data corresponding to the area pressed in the area-specific feature amount difference display input field 163, and a form identification. Data image is not displayed.

  The form identification processing unit 15 then selects the input region when the user presses, for example, a location indicated by the symbol A in the region-specific feature amount difference display input field 163 displayed on the image feature display screen. In addition to displaying the image of the input image data in the area corresponding to the image display field 164 and generating the display data for displaying the image of the form identification data of the area corresponding to the extraction area image display field 165, the generated display data is displayed. It outputs to the similarity display part 16 (step S5).

  When the display data is input from the form identification processing unit 15, the similarity display unit 16 displays the image of the input image data in the area corresponding to the input area image display field 164 as shown in FIG. An image feature display screen for displaying an image of the form identification data in the area corresponding to the extracted area image display field 165 is displayed (step S6).

  The image displayed in the input area image display field 164 of the image feature display screen shown in FIG. 4 is the image of the input image data in the area indicated by the symbol A, and the image displayed in the extraction area image display field 165. Is an image of the form identification data in the area indicated by the symbol A, and as shown in FIG. 4, the number of horizontal ruled lines is the same and the length difference between the vertical ruled lines is 50%. If the difference is less than 10% or more, the region-specific feature amount difference is small.

  Incidentally, the image displayed in the input area image display field 164 of the image feature display screen shown in FIG. 6 is an image of the input image data in the area indicated by reference numeral B, and is also displayed in the extraction area image display field 165. The image is an image of the form identification data in the area indicated by the symbol B, and has a large difference in area-specific feature amount when the number of horizontal ruled lines is different as shown in FIG.

  Thereafter, when the process of confirming the characteristics of the form for the input image data and other form identification data stored in the form identification result storage memory 13 is continued by the user, the form identification processing unit 15 and The similarity display unit 16 continues the same process as described above. On the other hand, when the process for confirming the feature of the form is finished, the process is finished.

  According to the embodiment, based on the feature data generated by the form identification processing unit 15 and the feature data described in the extracted feature image data, the difference between the features in each of the nine divided regions is shown. Since the region-specific feature amount difference indicating the amount is specified and the similarity display unit 16 displays the region-specific feature amount difference indicating the amount of difference between the features in the nine divided regions, it is registered in advance. Since the features between the form image and the form image to be actually classified can be confirmed and collated, it is possible to prevent a form collation error.

  In addition, according to the embodiment, when the similarity display unit 16 displays the feature amount difference for each of the nine divided regions, when the region in the displayed region-specific feature amount difference is pressed, this pressing is performed. Since the image of the image data input from the image input processing unit in the area and the image of the image data included in the form identification data extracted by the form identification processing unit 15 are enlarged and displayed, it has been registered in advance. Since the features between the form image and the form image to be actually classified can be confirmed and collated in detail, it is possible to further prevent the form from being collated.

  In the form automatic identification device according to the above-described embodiment, the similarity display unit 16 displays the feature amount difference for each region with the display portion being displayed in monochrome. You may make it display.

It is a block diagram which shows the structure of the form automatic identification device of the Example which concerns on this invention. It is a figure for demonstrating form identification data. It is a figure for demonstrating nine area | regions which show the feature-value difference according to area | region. It is a figure which shows the display screen which the similarity display part in FIG. 1 displays. It is a flowchart which shows operation | movement of the form automatic identification device of the Example which concerns on this invention. It is a figure which shows the display screen which the similarity display part in FIG. 1 displays.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Input image memory 12 Form identification data storage memory 13 Form identification result storage memory 14 Image input process part 15 Form identification process part 16 Similarity display part

Claims (4)

A feature image data storage unit for storing feature data indicating features of image data of a form and feature image data corresponding to the image data;
An image data input unit for inputting the image data of the form;
A feature data generating unit that generates feature data indicating the characteristics of the image data from the image data input by the image data input unit;
Based on the feature data generated by the feature data generation unit and the feature data stored in the feature data storage unit, the feature image data similar to the image data input from the image data input unit is obtained. An extractor for extracting;
Based on the feature data generated by the feature data generation unit and the feature data described in the feature image data extracted by the extraction unit, the difference between the features of each other in an area divided into a predetermined number A feature quantity difference identifying unit that identifies a feature quantity difference indicating a quantity;
A display unit for displaying a feature amount difference for each region specified by the feature amount difference specifying unit;
A form automatic identification device comprising:   In addition to displaying the feature amount difference for each region, the display unit, when a region in the displayed feature amount difference is selected, the image data input from the image data input unit in the selected region, The form automatic identification device according to claim 1, wherein the image data included in the feature image data extracted by the extraction unit is enlarged and displayed.   The form automatic identification device according to claim 1, wherein the display unit displays different colors in the area depending on a difference in feature amount. The form automatic identification device according to claim 1, wherein the feature is a ruled line feature based on a ruled line of the form, and a difference between the features is specified by a number and a length of the ruled line.

Images