JP2010256994A - Character recognition device, character recognition method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a character recognition device, a character recognition method, and a program capable of improving identification of characters and a recognition process and efficiency of reuse of data afterwards by performing re-synthesis about an image including only information to be needed by a user, and an image which makes a front and a back pages of an original in one page from a plurality of images of the original read by a scanner etc. <P>SOLUTION: A part of an area of the plurality of images is extracted, then a synthesized image which reconstructs a plurality of areas is produced, then a character within the produced synthesized image is recognized. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a character recognition device, a character recognition method, and a program for digitizing and managing paper documents and forms.

  In a conventional character recognition device, when a paper document or a form is digitized and processed, a plurality of pages read by an image reading device such as a scanner are combined as one image, and character recognition of the combined image is performed. Such processing and management are performed.

  For example, in the technique described in Patent Document 1, based on the size of a plurality of forms read by a scanner or the like, a combined image is generated by combining form images in series in the vertical direction or the horizontal direction, and the combined image is generated. A technique for recognizing a form based on this is disclosed.

JP 2004-178108 A

  However, in a conventional character recognition device (eg, Patent Document 1), only a specific area in a plurality of forms read by a scanner or the like is cut out and re-created as a new image including only information required by the user. It had the problem that it could not be synthesized. For this reason, in the character recognition device, when one related form covers a plurality of pages, there is a problem that the data size becomes enormous by storing an image including all of them in a storage area. . Specifically, in insurance contract-related documents, there are a wide variety of documents and forms such as contracts, medical checkups, insurance claims application forms, doctor's diagnosis forms, etc. On the other hand, there are matters that are common among insurers such as precautions, etc. When acquiring all as images, there was a problem that the data size would increase due to common items that do not need to be stored in particular . Further, in the character recognition device, a form in which a plurality of items (for example, a contractor's address, name, phone number, date of birth, etc.) is not managed as an image extracted for each item cannot be managed. For example, when checking and correcting images, all items of information are disclosed to the worker, and personal information written on the form may be leaked through the worker. It was.

  Further, in a conventional character recognition device (Patent Document 1 or the like), when there are reading areas on both front and back sides of a form, the front image and the back image are highly related, and even if it is necessary to manage them together, the page unit Therefore, there is a problem that the front and back must be associated manually. Specifically, when the total of the items on the front side of the form is the value of the item on the back side of the form, there is a problem that the consistency of the front and back values cannot be confirmed if the front and back sides are separate images. Specifically, when the front and back of the document are a series of data and are only separated into front and back due to paper space, the recognition result is output as text or CSV (Comma-Separated Values). In this case, there is a problem that the front and back cannot be output as one data. Specifically, in the case of a character recognition device having a function of improving character recognition by using a handwritten character's unique wrinkle, the character's wrinkle is extracted from one page to improve the accuracy of character recognition processing. However, when front and back documents written by the same person are processed as separate pages, it is necessary to build a mechanism for dealing with front and back images, compared to a case where both front and back sides are managed centrally. There was a problem of requiring processing time and data size. Further, in the character recognition device, the type of the composite image is classified according to the layout of the composite image, and character recognition is performed according to the type of the classification, but the front and back are managed as separate images. However, even when the layouts are not similar, there is a problem in that images of forms that are similar only on one side are sorted as the same type.

  The present invention has been made in view of the above-described problems. From a plurality of original images read by a scanner or the like, an image including only information required by the user, or an image in which the front and back sides of the original are one page. It is an object of the present invention to provide a character recognition device, a character recognition method, and a program that can improve character identification, recognition processing, and subsequent data reuse efficiency.

  In order to achieve such an object, a character recognition device of the present invention is a character recognition device including at least a storage unit and a control unit, and the storage unit stores an image of a document read by an image reading device. An image storage unit configured to extract a partial region of the plurality of images stored in the image storage unit, and generate a composite image obtained by reconstructing the plurality of regions Means and a recognition means for recognizing characters in the composite image generated by the image reconstruction means.

  The character recognition device of the present invention is characterized in that in the character recognition device described above, the plurality of images are the images on both the front and back sides of the document.

  Further, the character recognition device of the present invention is the character recognition device described above, wherein the control unit includes layout acquisition means for acquiring layout information of the document from the image stored in the image storage means, The image reconstruction unit extracts the region based on the layout information acquired by the layout acquisition unit.

  The character recognition device according to the present invention is characterized in that, in the character recognition device described above, the layout information is information relating to the similarity of the layout of the document.

  Further, the character recognition device of the present invention is the above-described character recognition device, wherein the control unit generates image reconstruction information generating means for generating image reconstruction information representing the arrangement and size of the region on the composite image. , And the image reconstruction unit generates the composite image based on the image reconstruction information generated by the image reconstruction information generation unit.

  The character recognition device according to the present invention is characterized in that, in the character recognition device described above, the image reconstruction information is information relating to a predetermined area designated by a user.

  The character recognition device according to the present invention is characterized in that in the character recognition device described above, the document is a form.

  A character recognition method according to the present invention is a character recognition method executed in a character recognition device including at least a storage unit and a control unit, and the storage unit stores an image of a document read by the image reading device. Image storage means for extracting a partial area of the plurality of images stored in the image storage means and generating a composite image reconstructed from the plurality of areas, which is executed in the control unit. An image reconstruction step; and a recognition step for recognizing characters in the composite image generated in the image reconstruction step.

  The program according to the present invention is a program for causing a character recognition device including at least a storage unit and a control unit to execute, and the storage unit stores an image of a document read by the image reading device. An image reconstructing step for extracting a partial region of the plurality of images stored in the image storage unit and generating a composite image in which the plurality of regions are reconstructed in the control unit; A recognition step of recognizing characters in the composite image generated in the image reconstruction step.

  According to the present invention, it is possible to extract a partial area of a plurality of images stored in the storage unit, generate a composite image in which the plurality of areas are reconstructed, and recognize characters in the generated composite image it can. As a result, the present invention has an effect of omitting a management mechanism for a plurality of pages, which has been conventionally required, that is, a plurality of pages of data can be managed as a set of data. As a result, the present invention not only reduces paper storage costs, but also facilitates the retrieval of digitized images and the processing of data, thereby improving the character recognition searchability. .

  According to the present invention, the plurality of images can be images on both the front and back sides of the document. Thus, according to the present invention, when the front and back sides of a document are handled individually, there is an effect that processing and management that can be handled as a series of related data can be performed between the back and front data that could not be processed. Specifically, the present invention reconstructs specific areas on the front and back sides of a document as a one-page image, and processes the image to perform all necessary processing and management for the one-page image. The function can be used as it is, and the data size can be reduced.

  Further, according to the present invention, document layout information can be acquired from an image stored in the storage unit, and an area can be extracted based on the acquired layout information. Thereby, the present invention has an effect that an image of a desired item can be efficiently extracted from a form or the like having a predetermined format.

  Further, according to the present invention, the layout information can be information relating to the similarity of the layout of the document. As a result, the present invention has an effect that documents of the same format can be efficiently extracted from a plurality of documents. As a result, the present invention produces an effect of synthesizing front and back images of similar documents with similar layouts that have been difficult to identify in the past as a composite image of one page and enabling subsequent processing such as character recognition.

  Further, according to the present invention, it is possible to generate image reconstruction information indicating the arrangement and size of a region on a combined image, and to generate a combined image based on the generated image reconstruction information. Thereby, the present invention has an effect that a synthesized image according to a predetermined arrangement can be efficiently generated from the extracted image.

  According to the present invention, the image reconstruction information can be information related to a predetermined area designated by the user. As a result, the present invention has an effect of collecting only the portions required by the user and generating an image excluding redundant portions. That is, the present invention cuts out only necessary portions of a read original according to a user's instruction, and creates and manages an image obtained by synthesizing them, thereby reducing the overall image size and not losing essential information. The effect of doing.

  According to the present invention, the document can be a form. As a result, the present invention has an effect that the management of paper forms on which personal information or the like is described can be efficiently performed as data. Specifically, the present invention generates and manages an image composed only of specific items (for example, name, address, telephone number, etc.) from contract documents of a plurality of persons, and thereby, personal information is obtained from only individual images. It cannot be obtained, and has the effect of being useful as a countermeasure against leakage of personal information and the like.

FIG. 1 is a block diagram showing an example of the configuration of a character recognition device to which the present invention is applied. FIG. 2 is a flowchart showing an example of character recognition processing in the present embodiment. FIG. 3 is a diagram showing an example of the image reconstruction information in the present embodiment. FIG. 4 is a diagram illustrating an example of a composite image in the present embodiment. FIG. 5 is a flowchart showing an example of image reconstruction processing in the present embodiment.

  Embodiments of a character recognition device, a character recognition method, and a program according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

  Hereinafter, the configuration and processing of the present invention will be described in detail.

[Configuration of character recognition device]
Next, the configuration of the character recognition device will be described with reference to FIG. FIG. 1 is a block diagram showing an example of the configuration of a character recognition apparatus to which the present invention is applied, and conceptually shows only the portion related to the present invention. Moreover, in FIG. 1, the arrow has shown the direction through which data flows.

  The character recognition device 100 generally includes a control unit 102 (layout acquisition unit 102a, image reconstruction information generation unit 102b, image reconstruction unit 102c, format definition unit 102d, recognition unit 102e, confirmation correction unit 102f, output conversion unit). 102g) and a storage unit 106 (image file 106a, layout information file 106b, composite image file 106c, format definition file 106d, final result file 106e). Here, the control unit 102 is a CPU or the like that comprehensively controls the entire character recognition apparatus 100. The storage unit 106 is a device that stores various databases and tables. Each part of these character recognition apparatuses 100 is connected so that communication is possible via arbitrary communication paths.

  The storage unit 106 is a storage unit, and stores various databases and tables (an image file 106a, a layout information file 106b, a composite image file 106c, a format definition file 106d, a final result file 106e, and the like).

  Among the components of the storage unit 106, in FIG. 1, an image file 106 a is an image storage unit that stores an image of a document read by the image reading device 200. Here, the document may be a form. As the image reading device 200, a scanner device, a digital camera, a WEB camera, or the like can be used.

  The layout information file 106b is layout information storage means for storing document layout information for extracting a target area from the document. Here, the layout information file 106b may be information on the similarity of the layout of the document.

  The composite image file 106c is a composite image storage unit that stores the composite image.

  In FIG. 1, a format definition file 106d is format definition storage means for storing a format definition for character reading for a composite image.

  The final result file 106e is final result storage means for storing the final result.

  The control unit 102 also has an internal memory for storing a control program such as an OS (Operating System), a program defining various processing procedures, and necessary data. And the control part 102 performs the information processing for performing various processes by these programs. In FIG. 1, the control unit 102 is functionally conceptually illustrated in terms of the function of layout acquisition unit 102 a, image reconstruction information generation unit 102 b, image reconstruction unit 102 c, format definition unit 102 d, recognition unit 102 e, confirmation correction unit 102 f, and output conversion unit. 102g is comprised.

  Among these, the layout acquisition unit 102a is a layout acquisition unit that acquires layout information of a document from an image stored in the image file 106a. Here, the layout information may be information regarding the similarity of the layout of the document. The layout acquisition unit 102a may store the acquired layout information in the layout information file 106b.

  The image reconstruction information generation unit 102b is an image reconstruction information generation unit that generates image reconstruction information indicating the arrangement and size of a partial region of an image on a composite image. Here, the image reconstruction information may be information regarding a predetermined area designated by the user.

  The image reconstruction unit 102c is an image reconstruction unit that extracts partial areas of a plurality of images stored in the image file 106a and generates a composite image obtained by reconstructing the plurality of areas. Here, the image reconstruction unit 102c may extract a region based on the layout information acquired by the layout acquisition unit 102a. Further, the image reconstruction unit 102c may generate a composite image based on the image reconstruction information generated by the image reconstruction information generation unit 102b. The image reconstruction unit 102c may store the composite image in the composite image file 106c. The plurality of images may be images on both the front and back sides of the document.

  The format definition unit 102d is a format definition unit that sets a format definition for character reading of a composite image. Here, the format definition unit 102d may store the set format definition in the format definition file 106d.

  The recognition unit 102e is a recognition unit that recognizes characters in the composite image generated by the image reconstruction unit 102c. Here, the recognition unit 102e may recognize characters in the composite image generated by the image reconstruction unit 102c based on the format definition stored in the format definition file 106d.

  The confirmation / correction unit 102f is a confirmation / correction unit that confirms and corrects characters in the composite image based on the recognition result of the characters in the composite image by the recognition unit 102e.

  The output conversion unit 102g is an output conversion unit that converts the composite image confirmed and corrected by the confirmation correction unit 102f into a final result in a data format to be output. Here, the output conversion unit 102g may store the final result in the final result file 106e.

  This is the end of the description of the configuration of the character recognition apparatus 100.

[Character recognition processing]
Next, an example of character recognition processing in the present embodiment will be described in detail with reference to FIGS. FIG. 2 is a flowchart showing an example of character recognition processing in the present embodiment.

  As shown in FIG. 2, the control unit 102 stores the image of the paper (form) read by the image reading device 200 in the image file 106a (step SA-1).

  Then, the layout acquisition unit 102a acquires form layout information from the image stored in the image file 106a, and stores the layout information in the layout information file 106b (step SA-2). Here, the layout information may be information related to the similarity of the layout of the form.

  The image reconstruction information generation unit 102b represents the arrangement and size of a partial area of the image, that is, a predetermined area designated by the user, on the composite image generated by the image reconstruction unit 102c. Image reconstruction information is generated (step SA-3).

  Here, an example of the image reconstruction information will be described with reference to FIG. FIG. 3 is a diagram showing an example of the image reconstruction information in the present embodiment.

  As shown in FIG. 3, the image reconstruction information includes the number N of pre-reconstructed input images, that is, the number of images of the form stored in the image file 106a used for generating a group of combined images. It is out. Here, the number N of pre-reconstructed input images is 2 when the images on both the front and back sides of the form are made into a single composite image. Further, the image reconstruction information includes the post-reconstruction input image size, that is, the output size of the composite image. Here, in the present embodiment, the image reconstruction information is A4. Further, as shown in FIG. 3, the image reconstruction information is image reconstruction information I (I = 1,..., N), that is, which region of the I-th image is arranged in the composite image. Including information. For example, in the two-sided image of the form in FIG. 3, the image reconstruction information 1 (front) is the item A of the size (Sx, Sy) starting from the coordinate start position (Xa, Ya) of the image 1 (form 1). This is information indicating that the image is arranged at a size (Hx, Hy) starting from the coordinate start position (X1, Y1) of the composite image. In addition, in the front and back side images of the form in FIG. 3, the image reconstruction information 2 (back side) is the entire image of the image 2 (form 2), that is, the entire area image on the back side, and the coordinate start position (X2 , Y2) is information indicating the arrangement in the size (Fx, Fy) starting from.

  Returning to FIG. 2 again, the image reconstruction unit 102c extracts partial areas of the plurality of images stored in the image file 106a based on the layout information stored in the layout information file 106b, and performs image reconstruction. Based on the image reconstruction information generated by the information generation unit 102b, a composite image in which the front and back areas of the form are reconstructed is generated and stored in the composite image file 106c (step SA-4).

  Here, an example of the composite image in the present embodiment will be described with reference to FIG. FIG. 4 is a diagram illustrating an example of a composite image in the present embodiment.

  As shown in FIG. 4, the composite image shows an image generated based on the image reconstruction information shown in FIG. 3, and a partial image on the front side of the form (item A in image 1) and the back side of the form The entire image (the entire image of the image 2) is combined with each other arranged vertically.

  Also, an example of details of the image reconstruction process in the present embodiment will be described with reference to FIG. FIG. 5 is a flowchart showing an example of image reconstruction processing in the present embodiment.

  First, as shown in FIG. 5, the image reconstruction unit 102c reads the number N of pre-reconstructed input images included in the image reconstruction information generated by the image reconstruction information generation unit 102b (step SB-1).

  Then, the image reconstruction unit 102c clears the count K of the counter that counts the number of input images, that is, sets it to 0 (step SB-2).

  Then, the image reconstruction unit 102c initializes the composite image area (composite image file 106c) in the storage area (step SB-3).

  Then, the image reconstruction unit 102c adds 1 as the number of input images to the count K of the counter (step SB-4).

  Then, the image reconstruction unit 102c determines whether the count K is greater than the number N of pre-reconstructed input images (step SB-5).

  If the image reconstruction unit 102c determines in step SB-5 that the count K is greater than the number N of pre-reconstructed input images (step SB-5: Yes), the composite image is a composite image region (composite image). The file 106c) is output (for example, stored) (step SB-6), and the process is terminated.

  On the other hand, if the image reconstruction unit 102c determines in step SB-5 that the count K is equal to or less than the number N of pre-reconstructed input images (step SB-5: No), the I = Kth image stored in the image file 106a. Image K and the layout information of the image K stored in the layout information file 106b (step SB-7), and the image reconstruction information K related to the image K generated by the image reconstruction information generation unit 102b is obtained. Obtain (Step SB-8).

  Then, the image reconstruction unit 102c performs a layout matching process for extracting a target area from the image K based on the layout information acquired in Step SB-7 (Step SB-9). Here, the image reconstruction unit 102c may collate the front and back of the form to be combined with the forms having similar layouts on the front and back sides based on the layout information related to the similarity of the layouts.

  Then, the image reconstructing unit 102c acquires an extracted image constituting a composite image from the image K acquired in step SB-7 based on the layout matching process performed in step SB-9 (step SB- 10).

  Then, the image reconstruction unit 102c pastes the extracted image into the composite image based on the image reconstruction information K acquired in Step SB-8, and shifts the processing to Step SB-4 (Step SB-11). .

  Returning to FIG. 2 again, the format definition unit 102d sets a format definition for character reading for the composite image and stores it in the format definition file 106d (step SA-5).

  Then, the recognition unit 102e recognizes the characters in the composite image generated by the image reconstruction unit 102c based on the format definition stored in the format definition file 106d (step SA-6).

  Then, the confirmation / correction unit 102f determines whether a predetermined area (for example, name, address, etc.) of the form designated by the user is included based on the recognition result of the characters in the composite image by the recognition unit 102e. The characters in the image are confirmed and corrected (step SA-7).

  The output conversion unit 102g converts the composite image confirmed and corrected by the confirmation correction unit 102f into a final result in a data format (for example, JPEG, GIF, etc.) to be output to a printer or the like (step SA-8). . Here, the output conversion unit 102g may store the final result in the final result file 106e.

  Above, description of an example of the character recognition process in this Embodiment is finished.

[Other embodiments]
Although the embodiments of the present invention have been described so far, the present invention is not limited to the above-described embodiments, but can be applied to various different embodiments within the scope of the technical idea described in the claims. It may be implemented.

  For example, although the case where the character recognition device 100 performs processing in a stand-alone form has been described as an example, the processing is performed in response to a request from a client terminal configured in a separate housing from the character recognition device 100, and the processing You may comprise so that a result may be returned to the said client terminal.

  In addition, among the processes described in the embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method.

  In addition, unless otherwise specified, the processing procedures, control procedures, specific names, information including registration data for each processing, parameters such as search conditions, screen examples, and database configurations shown in the above documents and drawings Can be changed arbitrarily.

  Moreover, regarding the character recognition apparatus 100, each component shown in the figure is functionally conceptual, and need not be physically configured as illustrated.

  For example, all or some of the processing functions provided in each device of the character recognition device 100, particularly the processing functions performed by the control unit 102, are interpreted and executed by a CPU (Central Processing Unit) and the CPU. It may be realized by a program to be executed, or may be realized as hardware by wired logic. The program is recorded on a recording medium to be described later, and is mechanically read by the character recognition device 100 as necessary. That is, the storage unit 106 such as a ROM or HD stores a computer program for performing various processes by giving instructions to the CPU in cooperation with an OS (Operating System). This computer program is executed by being loaded into the RAM, and constitutes the control unit 102 in cooperation with the CPU.

  The computer program may be stored in an application program server connected to the character recognition apparatus 100 via an arbitrary network, and may be downloaded in whole or in part as necessary. is there.

  The program according to the present invention can also be stored in a computer-readable recording medium. Here, the “recording medium” is an arbitrary “portable physical medium” such as a flexible disk, a magneto-optical disk, a ROM, an EPROM, an EEPROM, a CD-ROM, an MO, a DVD, a Blu-ray Disc, or the like. It includes a “communication medium” that holds a program in a short period of time, such as a communication line or a carrier wave when a program is transmitted via a network represented by a LAN, WAN, or the Internet.

  The “program” is a data processing method described in an arbitrary language or description method, and may be in any format such as source code or binary code. The “program” is not necessarily limited to a single configuration, but is distributed in the form of a plurality of modules and libraries, or in cooperation with a separate program represented by an OS (Operating System). Including those that achieve the function. Note that a well-known configuration and procedure can be used for a specific configuration for reading a recording medium, a reading procedure, an installation procedure after reading, and the like in each device described in the embodiment.

  The various databases and the like stored in the storage unit 106 are memory devices such as RAM, SSD, and ROM, fixed disk devices such as hard disks, and storage means such as flexible disks and optical disks. Stores programs, tables, databases, web page files, etc.

  The character recognition apparatus 100 is connected to an information processing apparatus such as a known personal computer or workstation, and software (including programs, data, etc.) for realizing the method of the present invention is installed in the information processing apparatus. It may be realized.

  Furthermore, the specific form of distribution / integration of the devices is not limited to that shown in the figure, and all or a part of them may be functional or physical in arbitrary units according to various additions or according to functional loads. Can be distributed and integrated.

  As described above in detail, according to the present invention, from a plurality of original images read by a scanner or the like, an image including only information required by the user, an image with the front and back sides of the original as one page, Can be provided to provide a character recognition device, a character recognition method, and a program that can increase the efficiency of character identification, recognition processing, and subsequent data reuse. The present invention is extremely useful when implemented in the field of image processing related to images.

100 character recognition device 102 control unit
102a Layout acquisition unit
102b Image reconstruction information generation unit
102c Image reconstruction unit
102d Format definition part
102e recognition unit
102f Confirmation and correction part
102g Output conversion unit 106 Storage unit
106a image file
106b Layout information file
106c Composite image file
106d Format definition file
106e Final result file 200 Image reading device

Claims (9)

A character recognition device comprising at least a storage unit and a control unit,
The storage unit
Image storage means for storing an image of a document read by the image reading device;
With
The control unit
Image reconstruction means for extracting a partial area of the plurality of images stored in the image storage means and generating a composite image obtained by reconstructing the plurality of areas;
Recognition means for recognizing characters in the composite image generated by the image reconstruction means;
A character recognition device comprising: The character recognition device according to claim 1,
The character recognition apparatus according to claim 1, wherein the plurality of images are the images on both sides of the document. In the character recognition device according to claim 1 or 2,
The control unit
Layout acquisition means for acquiring layout information of the document from the image stored in the image storage means;
With
The character recognizing device, wherein the image reconstructing means extracts the region based on the layout information acquired by the layout acquiring means. The character recognition device according to claim 3,
The character recognition apparatus according to claim 1, wherein the layout information is information related to the similarity of the layout of the originals. In the character recognition device according to any one of claims 1 to 4,
The control unit
Image reconstruction information generating means for generating image reconstruction information representing the arrangement and size of the region on the composite image;
With
The character recognizing device, wherein the image reconstruction means generates the composite image based on the image reconstruction information generated by the image reconstruction information generation means. In the character recognition device according to any one of claims 1 to 5,
The character recognizing apparatus, wherein the image reconstruction information is information relating to a predetermined area designated by a user. In the character recognition device according to any one of claims 1 to 6,
A character recognition apparatus, wherein the original is a form. A character recognition method executed in a character recognition device including at least a storage unit and a control unit,
The storage unit
Image storage means for storing an image of a document read by the image reading device;
With
Executed in the control unit,
An image reconstruction step of extracting a partial region of the plurality of images stored in the image storage unit and generating a composite image in which the plurality of regions are reconstructed;
A recognition step for recognizing characters in the composite image generated in the image reconstruction step;
A character recognition method comprising: A program for causing a character recognition device including at least a storage unit and a control unit to execute the program,
The storage unit
Image storage means for storing an image of a document read by the image reading device;
With
In the control unit,
An image reconstruction step of extracting a partial region of the plurality of images stored in the image storage unit and generating a composite image in which the plurality of regions are reconstructed;
A recognition step for recognizing characters in the composite image generated in the image reconstruction step;
A program for running

Images