JP2008083760A - Business form reader - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To appropriately discriminate a business form based on feature information and to correctly read character information recorded on the business form by invalidating an area which includes a large amount of feature information in the image data of input business form, and discriminating the business form by collating the feature information with that registered in advance. <P>SOLUTION: A business form reader 10 has; an image input means for inputting image data of a business form; a feature information extraction means to extract feature information from the image data; and a business form discrimination means for discriminating the business form by collating the feature information extracted by the feature information extraction means against the feature information registered in advance. When there exists an area including a large amount of the feature information in the image data, the feature information extraction means invalidates the feature information on this area. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a form reading apparatus.

  Traditionally, financial institutions such as banks, post offices, credit unions, consumer finance companies, etc. have required information when updating customer information such as address changes and requesting financial transactions such as bank transfers at sales counters and counters. Items described in a form such as a written request slip are input to a computer such as a data management server. In this case, a form reading device having an OCR (Optical Character Reader) function for reading the character information recorded in the form is used to automatically input the information recorded in the form into the computer without manual intervention. Has been.

However, since the operations handled by the financial institution's branch are diverse, the forms accepted at the counter and counter are also diverse. Therefore, the image data of various forms is acquired in advance by a form reading device having an OCR function, the characteristics of the ruled lines of the form are registered as ruled line information, and the character information recorded in the form received at the window or counter Is read out, the ruled line information of the form is checked against previously registered ruled line information to identify the type of the form (for example, see Non-Patent Document 1). After identifying the type of the form in this way, the form reading device reads form data such as name, amount, and account number described in the form and transmits it to the computer.
“Oki Technical Review” published by Oki Electric Industry Co., Ltd., July 2002, No.191, Vol. 69, no. 3, p. 98-101 ("Image / OCR component")

  However, in the conventional form reading device, the form is identified based on the characteristic of the ruled line. Therefore, if the interval between the ruled lines is narrow or the ruled line is complicated, the characteristic of the ruled line is appropriately determined. Could not be identified. For example, in the conventional form reading device, if black pixels are continuous for a certain amount or more, they are recognized as ruled lines, so that the portion where white characters are printed in the black solid area or the password is hidden. It will be recognized that there is a ruled line in the portion for the purpose. That is, if all the information of a part where the interval between ruled lines is narrow or the ruled lines are complicated is read, the identification result of the form may be affected and the form may be erroneously recognized.

  However, in the conventional form reading device, erroneous recognition is prevented by removing noise. That is, at the time of registering the form, the registration is performed with noise removed. However, in this case, the ruled line itself may be removed as noise, and the form may not be properly identified.

  The present invention solves the problems of the conventional form reading device, and if there is an area containing a large amount of feature information in the image data of the input form, the area is invalidated and the feature information is registered in advance. Provides a form reader that can identify forms based on feature information, and can accurately identify forms based on feature information and accurately read text information recorded on forms. The purpose is to do.

  Therefore, in the form reading apparatus of the present invention, image input means for inputting image data of the form, feature information extraction means for extracting feature information from the image data, and feature information extracted by the feature information extraction means. A form reader having form identification means for identifying the form by collating with pre-registered feature information, wherein the feature information extraction means has an area in which a large amount of feature information is included in the image data. If it exists, the feature information of the area is invalidated.

  In another form reading apparatus according to the present invention, the form identifying means collates the feature information outside the area with the feature information registered in advance.

  In still another form reading apparatus of the present invention, the feature information is ruled line information, and the region is a region where ruled lines exist at a density equal to or higher than a threshold value.

  In still another form reading apparatus according to the present invention, the area further includes an area where white characters are printed in a black solid area or an area for hiding a password.

  Still another form reading apparatus according to the present invention further includes registered feature information storage means for storing pre-registered feature information.

  According to the present invention, in the form reading device, if there is a region containing a large amount of feature information in the image data of the input form, the region is invalidated and the feature information is collated with pre-registered feature information. The form is identified. As a result, the form can be appropriately identified based on the feature information, and the character information recorded on the form can be read accurately.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a form reading apparatus according to an embodiment of the present invention.

  In the figure, reference numeral 10 denotes a form reading device disposed in a bank, post office, credit union, a head office of a financial institution such as a consumer financial company, a branch office or the like, a form handling center of the financial institution, a data entry center, or the like. It is. The form reading device 10 is a request slip or the like in which necessary items are described when a customer requests a bank account or counter for notification or renewal of a new account, notification of a cash card PIN, change of address, etc. at a financial institution. This is a computer system for acquiring image data of an input form 11 as an identification target form and inputting items described in the input form 11 as entry data based on the image data.

  The form reading device 10 includes CPU, MPU and other arithmetic means, magnetic disk, storage means such as semiconductor memory, keyboard, mouse, push button, touch panel and other input means, CRT, liquid crystal display, LED (Light Emitting Diode). ) A computer having display means such as a display, a communication interface, and the like. The form reading device 10 is a personal computer, for example, but may be any type of computer such as a server or a workstation. Further, the form reading device 10 may be a single device or a device such as a distributed server in which a plurality of computers are connected to each other.

  The form reading apparatus 10 according to the present embodiment has an image input processing unit 12 as image input means for inputting an image of the input form 11 as image data, and input image data for storing input image data, from the viewpoint of function. An input image memory 13 as storage means, a feature information extraction section 14 as feature information extraction means for extracting feature information from the image data, a feature information storage memory 15 as feature information storage means for storing the extracted feature information, Form identification processing unit 16 as form identification means for identifying the input form 11, form identification information storage memory 21 as registered feature information storage means for storing characteristic information of various registered forms, and the form reading device 10 receives an input operation by an operator who operates the form reading device 10 and a control unit 22 that controls the operation of the entire document It takes the input unit 23, and a display unit 24 for displaying information to the operator. The image input processing unit 12, the input image memory 13, the feature information extraction unit 14, the feature information storage memory 15, the form identification processing unit 16, the form identification information storage memory 21, the control unit 22, the input unit 23, and the display unit 24. Are connected to each other via a communication line network such as a data bus, a system bus, a LAN (Local Area Network), and an intranet.

  Here, the image input processing unit 12 is a flat head scanner, a sheet feed scanner, a drum scanner, a non-contact type stand scanner, or the like, but may be any kind of scanner. The image input processing unit 12 is preferably a color scanner capable of reading a color image, but may be a monochrome scanner. In the present embodiment, the image input processing unit 12 includes items to be input to the form reading device 10, that is, a light source composed of a fluorescent lamp that irradiates light to the input form 11 on which entry data is described, and the input An image sensor such as a line type image sensor including a CCD (Charge Coupled Device) that receives reflected light from the surface of the form 11 is provided.

  Furthermore, it is desirable that the image input processing unit 12 has a hopper, a conveyance path, a stacker, and the like and can sequentially read a plurality of input forms 11. In this case, the hopper stores a plurality of input forms 11 in a stacked state. The transport path is provided with a transport device such as a belt and a roller, is formed in the image input processing unit 12, and a plurality of input forms 11 stacked in the hopper are placed on the top or the top by a paper feed device. When sequentially sent from the bottom, the fed input form 11 is conveyed to the image sensor. Further, the transport path transports the input form 11 for which image data has been acquired by the image sensor and discharges it to the stacker.

  In addition, a printing unit including a print head is disposed on the downstream side of the image reading unit in the conveyance path, and a serial number is printed in a serial number description column at a corner or a back side margin of the input form 11. It is desirable. The serial number is a number indicating the order in which the reading processing is performed by the image input processing unit 12 in the order in which the input forms 11 are read.

  The input form 11 is a form having a format different for each business and describing necessary items corresponding to each business. Here, the business is a business performed by a financial institution in response to a notification or renewal requested by a customer at a window or counter, but includes a business performed by a financial institution in addition to a business corresponding to a customer's request. May be. The input form 11 is sorted for each business, and the image data of the input form 11 corresponding to the business designated by the operator operating the input unit 23 is acquired by the image input processing unit 12.

  Further, the feature information extraction unit 14 extracts feature information for identifying what kind of the input form 11 is from the image data stored in the input image memory 13. The feature information is, for example, ruled line information as information including the shape, length, number, configuration, and the like of ruled lines printed on the input form 11. The feature information of the input form 11 extracted by the feature information extraction unit 14 is stored in the feature information storage memory 15.

  Further, the form identification processing unit 16 collates the feature information of the input form 11 stored in the feature information storage memory 15 with the feature information of various forms stored in the form identification information storage memory 21. The input form 11 is identified. In the form identification information storage memory 21, feature information of various forms having different formats for each business is registered and stored in advance. Therefore, by comparing with the feature information stored in the form identification information storage memory 21, the type of the input form 11 can be specified, and what format is provided can be determined.

  The input unit 23 receives an input operation by an operator who operates the form reading apparatus 10, and corresponds to input means such as a keyboard, a mouse, a push button, and a touch panel. Further, the display unit 24 displays various types of information to the operator, and corresponds to display means such as a CRT, a liquid crystal display, and an LED display.

  Next, the operation of the form reading apparatus 10 configured as described above will be described. Here, an operation for identifying the input form 11 will be described.

  FIG. 2 is a diagram showing a first example of an input form in the embodiment of the present invention, FIG. 3 is a diagram showing a second example of the input form in the embodiment of the present invention, and FIG. 4 is a diagram showing an embodiment of the present invention. FIG. 5 is a diagram showing image data of the second example of the input form in the form, FIG. 5 is a diagram showing a range in which the feature information is invalidated in the image data of the second example of the input form in the embodiment of the present invention, FIG. These are the flowcharts which show operation | movement of the form reading apparatus in embodiment of this invention.

  First, when the operator operates the input unit 23 to start an operation for inputting items described in the input form 11 as entry data to the form reading apparatus 10, the form reading apparatus 10 performs image input processing. Execute. In the image input process, the control unit 22 controls the image input processing unit 12 to cause the image input processing unit 12 to read the input form 11. Then, the image data of the input form 11 acquired by the image input processing unit 12 is stored in the input image memory 13.

  Here, the input form 11 is, for example, as shown in FIG. In this case, the input form 11 is a type of form used by a customer to request opening a new account, etc., and the area 11a-1 in which white characters are printed in the black solid area and the secret number are hidden. Region 11a-2 for the purpose. The image input processing unit 12 recognizes the regions 11a-1 and 2 as regions containing a large amount of ruled lines because black pixels are continuous for a certain level or more as ruled lines. End up.

  FIG. 3 shows another example of the input form 11. The input form 11 shown in FIG. 3 has a region 11b where ruled lines exist at a density equal to or higher than a predetermined threshold. The threshold is a density corresponding to a state in which a predetermined number is included in a predetermined range. For example, when the input form 11 has an A4 size, three thresholds in a 1 cm square range. This is a density corresponding to a state in which a ruled line is included. The threshold value is set in advance, but can be changed as appropriate.

  The image input processing unit 12 itself has a configuration with a relatively low resolution, such as a non-contact type stand scanner, or in an environment that is easily affected by ambient light and shadows. In the case of being placed, adjacent ruled lines are crushed and connected in the region 11b, so that the image data of the input form 11 is as shown in FIG. There is a possibility that. That is, there is a possibility that the region 11b where the ruled line exists at a density equal to or higher than the threshold becomes a black solid region.

  Next, the form reading device 10 executes feature information extraction processing. In the feature information extraction process, the feature information extraction unit 14 extracts feature information from the image data of the input form 11 stored in the input image memory 13 and stores it in the feature information storage memory 15. Here, description will be made assuming that the feature information is ruled line information including the shape, length, number, and configuration of ruled lines. In this case, the feature information extraction unit 14 invalidates ruled line information as feature information existing in a range 11c as shown in FIG. 5 including a region 11b where ruled lines exist at a density equal to or higher than a threshold. Therefore, the feature information extraction unit 14 extracts only ruled line information about ruled lines existing outside the range 11c from the image data of the input form 11 as feature information.

  Next, the form reading device 10 executes form identification processing. In the form identification process, the form identification processing unit 16 sequentially extracts feature information of various forms registered in advance in the form identification information storage memory 21 and inputs the input form 11 stored in the input image memory 13. And feature information extracted by the feature information extraction unit 14 from the image data. And information (for example, ID etc.) which specifies the feature information of the form whose coincidence rate is equal to or higher than a predetermined threshold is displayed on the display unit 24. As a result, the input form 11 is identified, and the type thereof can be grasped as information for specifying the characteristic information of the pre-registered form stored in the form identification information storage memory 21.

Next, a flowchart will be described.
Step S1: An image input process is performed.
Step S2: Feature information extraction processing is performed.
Step S3 A form identification process is performed and the process is terminated.

  As described above, in the present embodiment, the form reading device 10 invalidates the feature information in the area where the intervals between the ruled lines are narrow or complicated, and based on the feature information in the other area, 11 is identified. Therefore, the input form 11 can be appropriately identified based on the feature information. Therefore, the character information recorded on the input form 11 can be read accurately.

  In addition, a stable identification result of the input form 11 can be obtained regardless of the type of the image input processing unit 12 and the environment in which the image data of the input form 11 is acquired.

  In addition, this invention is not limited to the said embodiment, It can change variously based on the meaning of this invention, and does not exclude them from the scope of the present invention.

It is a block diagram which shows the structure of the form reading apparatus in embodiment of this invention. It is a figure which shows the 1st example of the input form in embodiment of this invention. It is a figure which shows the 2nd example of the input form in embodiment of this invention. It is a figure which shows the image data of the 2nd example of the input form in embodiment of this invention. It is a figure which shows the range which invalidates characteristic information in the image data of the 2nd example of the input form in embodiment of this invention. It is a flowchart which shows operation | movement of the form reading apparatus in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Form reader 11a-1, 11a-2, 11b Area | region 12 Image input process part 14 Feature information extraction part 16 Form identification process part 21 Form identification information storage memory

Claims (5)

(A) image input means for inputting image data of a form;
(B) feature information extracting means for extracting feature information from the image data;
(C) A form reading device having form identification means for identifying the form by comparing feature information extracted by the feature information extraction means with previously registered feature information,
(D) The form reading device, wherein the feature information extraction unit invalidates the feature information of the area when there is an area containing a large amount of feature information in the image data. The form reading device according to claim 1, wherein the form identifying unit collates feature information outside the area with feature information registered in advance. The form reading apparatus according to claim 1, wherein the feature information is ruled line information, and the area is an area where ruled lines exist at a density equal to or higher than a threshold value. The form reading apparatus according to claim 3, wherein the area includes an area where white characters are printed in a black solid area or an area for hiding a password. The form reading device according to claim 1, further comprising registered feature information storage means for storing pre-registered feature information.

Images