JP4092849B2 - Form recognition device and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a form recognition device and a recording medium for recognizing a form image with reference to various form formats set in advance and determining the form type of the form image according to the recognition state.
[0002]
[Prior art]
Conventionally, in a business computer device that reads a form image and processes data, after recognizing the form image with reference to various preset form formats, the recognition rate is recognized for each form format. Accordingly, the type of the form image is determined. In other words, the form format set for each form is read one by one, and the operation of recognizing the form image based on this form format is repeated for the set number of forms, and the recognition rate for character recognition for each form is compared. The form with the highest recognition rate is determined as the form type of the form image.
[0003]
[Problems to be solved by the invention]
In this way, in the method of obtaining and comparing the recognition rate for each form format, it is necessary to read out the form format information set in advance separately and recognize characters, so the type is determined after reading the form image. A series of processes performed until confirmation is required for the number of times corresponding to the number of forms format information set. Therefore, as the number of settings increases, the number of processes increases, and it takes time to determine the form type, which causes a reduction in processing efficiency and adversely affects the entire business process.
[0004]
An object of the present invention is to simply synthesize a form group and dynamically synthesize a plurality of form format definition information corresponding to the designated group. By recognizing a form image based on this synthesized format, A series of processes from reading a form image to determining its type can be executed in a single process so that the form type can be determined efficiently.
[0005]
[Means for Solving the Problems]
The form recognition apparatus of the present invention is preset. Consists of multiple fields In the form recognition device that recognizes the form image with reference to the form format and determines the type of the form image according to the recognition state in which the form is recognized for each form format. Corresponding to each field of the form, field definition information consisting of the area coordinates and character type of the field Form format As The form definition storage means for storing, the form group storage means for storing the correspondence relationship between the form group and each form belonging to the group, and each form associated with the designated group when the form group is designated. Read from the form group storage means and for each read form The field definition information defined for the form Reading means from the form definition storage means, and each form read by the reading means Field Read each definition information , A plurality of field definition information in which the field definition information is sequentially written as synthetic field definition information Definition synthesis means for creating stored synthesis format definition information, and synthesis format definition information obtained by this definition synthesis means Synthetic field definition information Character recognition means for recognizing the form image based on the image, and the recognition result recognized by the character recognition means for each form Field And the recognition status of the divided recognition result is the best. Field Each form format corresponding to the designated group based on the designation of the form group is provided with a form type confirming means for confirming the form as the form image type. Field Definition information Synthesize The feature is that the form type is fixed.
[0006]
The present invention may be as follows.
(1) The definition combining means categorizes and synthesizes the format definition information for each form when creating the combined format definition information, and the character recognition means determines the form based on the combined format definition information. When the image is recognized as characters, the characters are recognized for each form category.
(2) The form format definition information set for each form includes basic definition information and field definition information that defines each field constituting the form, and the definition composition unit creates composition format definition information. At this time, the basic definition information and the field definition information are combined separately.
(3) The definition synthesis means analyzes the content of the basic definition information when creating the synthesis format definition information, creates a plurality of synthesis format definition information based on the analysis result, and the character recognition means The form image is recognized for each of the plurality of composite format definition information created based on the analysis result of the basic definition information, and the form type determination means recognizes the most based on the recognition result for each composite format definition information. A form having a good state is determined as the type of the form image.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
(First embodiment)
Hereinafter, a first embodiment of the present invention will be described with reference to FIGS.
FIG. 1 is a block diagram showing an overall configuration of a data processing apparatus with a form recognition function in this embodiment.
The data processing apparatus with a form recognition function reads a form image, recognizes characters in the read form image, and determines the type of the form according to the recognition rate. In other words, when a business to be processed is specified, each form format definition information corresponding to the business is dynamically synthesized, the form image is recognized according to the synthesized format definition information, and the recognition result is divided into each form. By comparing the recognition rates, the type of the form is determined according to the recognition rate, and a series of processes from reading the form image to determining the type is performed once. It is characterized by being able to be executed by processing and expecting a significant improvement in processing efficiency.
Prior to detailed description of the features of this embodiment, the hardware configuration of this embodiment will be described below.
[0009]
The CPU 1 is a central processing unit that controls the overall operation of the data processing apparatus with a form recognition function in accordance with the operating system and various application software in the storage device 2. The storage device 2 stores a database, character fonts, and the like in addition to an operating system and various application software, and includes a recording medium 3 configured by a magnetic, optical, semiconductor memory, and the drive system thereof. The recording medium 3 is a fixed medium such as a hard disk or a portable medium such as a removable CD-ROM, floppy disk, RAM card, or magnetic card. The program and data in the recording medium 3 are loaded into a RAM (for example, a static RAM) 4 as required by the control of the CPU 1, and the data in the RAM 4 is saved in the recording medium 3. Further, the recording medium may be provided on the side of an external device such as a server, and the CPU 1 can directly access and use the program / data in the recording medium via the transmission medium.
In addition, the CPU 1 can capture a part or all of the data stored in the recording medium 3 from another device via a transmission medium, and can newly register or add to the recording medium 3. That is, the transmission control unit 5 receives a program / data transmitted from another device constituting the computer communication system via a wired transmission line such as a communication line or cable or a wireless transmission line such as radio waves, microwaves, and infrared rays. And can be installed in the recording medium 3. Furthermore, the program / data may be stored and managed on the external device side such as a server, and the CPU 1 can directly access and use the program / data on the external device side via a transmission medium.
[0010]
On the other hand, the CPU 1 is connected with the input / output peripheral devices, ie, the transmission control unit 5, the input unit 6, the display unit 7, and the image scanner 8 through the bus line, and the CPU 1 performs these operations according to the input / output program. Control. The transmission control unit 5 is a communication interface including, for example, a communication modem, an infrared module, an antenna, and the like. The input unit 6 is an operation unit constituting a pointing device such as a keyboard, a touch panel, a mouse, or a touch input pen, and inputs character string data and various commands. The display unit 7 is a liquid crystal, CRT, or plasma display device that performs full color display, and the image scanner 8 is a stationary scanner that reads a form image.
[0011]
FIG. 2 is a diagram showing various files stored in the data processing apparatus.
The business information file F1 defines various forms used in the business according to the business content. FIGS. 7A and 8B illustrate the content of the business information file F1. . 7A shows forms {1}, {2}, {3} used in the business (1), and FIG. 8 (B) shows a form {1} used in the business (2). , (4) are shown. In this way, the business information file F1 is configured to store a correspondence relationship with each form used in the business for each group (business).
The basic definition file F2 stores basic definition information that defines the form size, the presence / absence of edge mark for alignment, etc. corresponding to each business form, and the field definition file F3 corresponds to each form. It is the structure which memorize | stores the definition information which defines each field which comprises.
[0012]
3 to 6, (A) is a diagram conceptually showing the formats of forms (1), (2), (3), and (4), and (B) is forms (1) and (2). The storage contents of the basic definition file F2 corresponding to ▼, (3), and (4) are shown. (C) shows the stored contents of the field definition file F3 corresponding to the forms (1), (2), (3), and (4).
In the basic definition file F2, the horizontal size and vertical size are set as the form size, and when there is information indicating the presence or absence of an edge mark and an edge mark, the start position and interval of the mark are set. The presence / absence of a timing mark is set. For example, in the basic format of the form (1), the form size is set to 200 mm in width, 150 mm in length, with edge marks, its start (20 mm, 20 mm), interval (150 mm, 110 mm), and no timing mark. In the field definition file F3, field information indicating “character type”, “number of characters”, and “area” is set for each field. For example, the form {circle over (1)} has three fields, “type number” is set as the character type of each field, “26” is set as the number of characters, and the upper left starting point for specifying each area Coordinates, lower right end point coordinates are set.
[0013]
The composite basic definition file F4 stores composite basic definition information obtained by combining the basic definitions of each form used in the business. FIG. 7B shows the composite basic definition of the business (1). Information, FIG. 8B exemplifies synthetic basic definition information of the business (2). A synthesis method for synthesizing the contents of the basic definition file F2 associated with each business-compatible form will be described later. FIG. 7C is a diagram showing a state in which the fields of the form {circle around (1)}, {circle around (2)}, and {circle around (3)} are conceptually superimposed. FIG. It is the figure which showed the state which overlap | superposed each field of no.
The synthetic field definition file F5 stores synthetic field definition information obtained by synthesizing the field definitions of each form used in the business. FIG. 7D shows the synthetic field definition of the business (1). Information, FIG. 8D exemplifies the composite field definition information of the business (2).
Here, when combining the contents of the basic definition file F2 associated with each business-supported form, field definition information for each form and each field corresponding to a composite field number that is a series of serial numbers. Are sequentially written to generate composite field definitions for a plurality of forms. Each time field definition information for one form is written in the composite field definition file F5, the correspondence between the field definition information writing range and the form is stored and managed.
[0014]
The form image file F6 temporarily stores the form image read by the image scanner 8, and the CPU 1 recognizes characters based on the form image in the form image file F6.
The recognition data file F7 stores and holds the recognition result for each form when the form image is recognized by referring to the definition information of each form format. FIG. 9 shows the data structure of the recognition data file F7. Is shown. That is, the recognition data file F7 is configured to store “field number”, “recognition data”, “recognition rate”, “field number”,... For each form and each field. In this case, the CPU 1 reads the contents of the recognition data file F7 for each form and compares the sum of the “recognition rates” of each field to determine the form having the largest value as the type of the form. I have to.
[0015]
Next, the form recognition process in the first embodiment will be described with reference to the flowcharts shown in FIGS. Here, a program for realizing each function described in these flowcharts is stored in the recording medium 3 in the form of a readable program code, and the CPU 1 sequentially executes operations according to the program code. To do. Further, the CPU 1 can sequentially execute the operation according to the above-described program code transmitted via the transmission medium. The same applies to other embodiments to be described later. In addition to a recording medium, an operation peculiar to this embodiment can be executed using a program / data supplied externally via a transmission medium.
[0016]
FIG. 10 is a flowchart showing the overall operation of the form recognition process.
First, when a company organization designates its own business when processing a form, the CPU 1 searches the business information file F1 based on the designated business, and acquires a plurality of forms used in the business as selection candidates. (Step A1).
Assuming that the business (1) is designated, forms (1), (2), (3) are selected as candidates for the business as shown in FIG. Then, the basic definition file F2 and the field definition file F3 are accessed based on the selected plurality of forms, and the corresponding basic definition information and field definition information are acquired (step A2). Definition information is stored in a work area in the RAM 4 (step A3).
In this state, a basic definition synthesis process is performed for synthesizing a plurality of basic definition information corresponding to the designated job, the synthesized result is stored in the synthesized basic definition file F4, and a plurality of field definition information corresponding to the designated job is synthesized. While performing the field definition synthesis process, the synthesis result is stored in the synthesis field definition file F5 (steps A4 to A6).
[0017]
FIG. 11 is a flowchart showing the basic definition synthesis process.
First, the basic definitions of all the forms in the work area acquired as selection candidates are read out and compared, and it is checked whether all the “form sizes” are the same (step B1). Here, if they are the same, the basic size is determined as the composite size (step B2), but if all are not the same, the value is the largest for each of the vertical and horizontal widths constituting the form size of each basic definition. A large size is extracted and determined as a combined size (step B3).
Now, the basic definitions for forms (1), (2), and (3) are set as shown in FIGS. 3A to 5A, and the maximum size is the width (200 mm). Therefore, it is the combined form size obtained by combining the basic definitions of forms (1), (2), and (3).
[0018]
Next, among the basic definitions of all the forms acquired as selection candidates, “edge mark presence / absence” is compared to check whether all are set to “edge mark present” (step B4). If it is set to “Yes”, it is determined as a basic composite definition (step B5). If any “No Edge Mark” is included in each basic definition, “No Edge Mark” is set. The basic composite definition is determined (step B6). In the above example, “no edge mark” is set as the basic composition definition.
Finally, among the basic definitions of all the forms, “presence / absence of timing mark” is compared, and it is checked whether “all with timing mark” is set (step B7). Then, it is determined as a basic composition definition (step B8). If even one “no timing mark” is included, “no timing mark” is determined as a basic composition definition (step B9). In this case, “no timing mark” is set as the basic composition definition.
The respective basic synthesis definition information synthesized in this way is stored in the synthesis basic definition file F4 (step A5 in FIG. 10).
[0019]
FIG. 12 is a flowchart showing the field definition synthesis process.
First, an initial value “1” is set in “synthesis field number N” and “write pointer M” used in this field definition synthesis process (step C1), and an initial value “1” is set in “read pointer n”. Is set (step C2). Here, “synthesis field number N” is a serial number set in the synthesis field definition file F5. The “write pointer M” is a pointer for designating the write address of the composite field definition file F5, and the “read pointer n” is a field for all the forms set as selection candidates in the work area of the RAM 4. This is a pointer for addressing and reading definitions one by one.
Then, the field definition for one form corresponding to the value of “read pointer n” out of all the forms set as selection candidates in the work area of the RAM 4 is designated as a range to be read (step C3). Initially, the first form field definition is specified.
[0020]
In this state, the nth form field definition indicated by “read pointer n” is acquired (step C4), and this field definition information is written in the address area of the composite field definition file F5 indicated by “write pointer M” ( Step C5). Then, “synthesis field number N” is written in the address area of the synthesis field definition file F5 indicated by “write pointer M” (step C6).
Then, after performing an update process of “+1” for the values of “combined field number N”, “write pointer M”, and “read pointer n” (step C7), the unwritten field definition information is Then, it is determined whether or not the work area still remains (step C8), and if there is, the process returns to step C4 to read the next field definition information designated by the “read pointer n”, and the above operation is repeated thereafter. As a result, field definition information for one form is sequentially written in the composite field definition file F5 corresponding to the “composite field number N”.
[0021]
When the writing of field definition information for one form is completed, the writing range becomes the address range of N1 to Nn. Therefore, the form number is associated with the address range of N1 to Nn, and the combined field definition file. Write to F5 (step C9).
Next, it is checked whether there are any other unprocessed forms (step C10). Now that the first form has been processed, it is determined that there is an unprocessed form, the process returns to step C2, the value of “read pointer n” is initialized to “1”, and the next form is designated. After that (step C3), the field definition information is read from the head of the designated form, and is written in the composite field definition file F5 together with the “composite field number N” (steps C4 to C8).
Thereafter, the synthesizing process for sequentially writing the field definition information to the synthetic field definition file F5 is repeated for all of the candidates for selection (steps C2 to C10).
[0022]
After the composite basic definition file F4 and the composite field definition file F5 are generated in this way, the process proceeds to step A7 in FIG. 10, and the form image read by setting the form document to the image scanner 8 is the form image file. In addition to being stored in F6, the form image is recognized for each form and for each field based on the contents of the synthetic basic definition file F4 and the synthetic field definition file F5 (step A8).
In this case, the recognition result recognized for each form and for each field is sequentially written to the recognition data file F7 for each form and for each field. At this time, the recognition rate is calculated for each field, and the recognition data file It is set at the corresponding position of F7 (steps A9, A10).
In the next step A11, recognition result analysis processing for analyzing the contents of the recognition data file F7 is performed.
[0023]
FIG. 13 is a flowchart showing the recognition result analysis process. This analysis process divides the recognition result into each form, and compares the recognition rates of the divided recognition results to obtain a form with the highest recognition rate. Is determined as the form type of the form image.
That is, since the recognition result recognized for each form and field is written in the recognition data file F7, first, the recognition data for each form is separated for each form (step D1), and for each form. The recognition rates corresponding to each field are summed (step D2). Then, the total recognition rates calculated for each form are compared, the form having the largest value is specified, and this form is determined as the type of form image read this time (step D3).
Then, the recognition data corresponding to the confirmed form is renumbered in accordance with the basic definition and field definition of the original form, and stored as processing target data in the work memory (step D4).
Note that the CPU 1 can execute normal form processing based on data stored as a processing target in the work memory.
[0024]
As described above, in the first embodiment, when a business to be processed is designated and input, the business information file F1 is referred to based on the designated business, and each form corresponding to the business is specified. For each identified form, each form format definition information read from the basic definition file F2 and the field definition file F3 is synthesized, and the form image is recognized according to the synthesized format definition information, and the recognition result for each form. The recognition rate is divided and the form with the best recognition rate is determined as the form type of the form image. By simply specifying the work, each form format definition information corresponding to the work can be moved. Can be combined, and a series of processes from reading a form image to determining its type can be executed in a single process. Type can be efficiently determined.
[0025]
Also, when creating composite format definition information, the format definition information is divided and combined for each form, and when recognizing the form image based on this composite format definition information, for each form section Since character recognition is performed, further improvement in processing efficiency can be expected.
The form format preset for each form consists of basic definition information and field definition information. When creating composite format definition information, it is synthesized separately for each basic definition information and for each field definition information. Since it did in this way, synthetic | combination format definition information can be handled like normal format information.
[0026]
(Second Embodiment)
A second embodiment of the present invention will be described below with reference to FIGS.
In the first embodiment described above, one composite format definition information is created based on a plurality of form format information corresponding to a designated job. However, the second embodiment is based on a difference in basic definition information. Thus, a plurality of composite format definition information is created. Here, components that are basically the same in both embodiments are denoted by the same reference numerals, description thereof will be omitted, and the following description will focus on the features of the second embodiment.
[0027]
FIG. 14 is a flowchart showing the form recognition processing in the second embodiment, and FIGS. 15 and 16 show specific examples in this case. FIG. The contents of the corresponding business information file F1 are shown.
First, when a processing task is designated, the business information file F1 is referenced based on the designated business, each form corresponding to the designated business is acquired (step E1), and the basic definition file F2 is obtained based on each obtained form. And basic definition information for each form is acquired (step E2).
Then, the basic definition information of each form is compared, and the forms are classified based on the “presence / absence of edge mark” (step E3). Now, when the business (1) is designated, the basic definition information in the form (1), the form (2), and the form (3) corresponding to the designated business is shown in (B) of FIGS. As shown, only the form (1) is “with edge mark”, so the form classification is divided into the classification of the form (1) only, and the classification including the form (2) and the form (3). It is done.
[0028]
Then, a basic definition synthesis process for synthesizing the basic definition information of each form for each classification is performed (step E4). That is, FIG. 15B and FIG. 16A exemplify the contents of the synthetic basic definition file F4 synthesized according to the classification corresponding to the business (1). The basic definition is synthesized based on the classification consisting only of 1 ▼, and the basic definition is synthesized based on the classification including the forms (2) and (3) of “no edge mark”. Since the synthesis of the basic definition is the same as that in the first embodiment described above, the description thereof is omitted.
[0029]
Next, a field definition synthesis process for synthesizing the field definition information of each form for each classification is performed. The synthesis in this case is also the same as that in the first embodiment described above (step E5). FIG. 15 (C, D) and FIG. 16 (B, C) show specific examples in this case.
When the process of combining the basic definition and field definition for each classification is completed in this way, a recognition process for recognizing the form image for each classification is performed based on the combined definition information for the number of classifications (step E6). The recognition results for the number of classifications are analyzed to determine the form type (step E7). Note that the character recognition and determination process in this case is the same as that in the first embodiment.
[0030]
As described above, in the second embodiment, when synthesizing each form format definition information corresponding to a designated job, a plurality of composite format definition information is created based on the contents of the basic definition information, The form image is recognized for each composite format definition information, and the form type is determined based on the recognition result for each composite format definition, so even if there are many different types of forms mixed together. The form type can be determined efficiently without reducing the rate.
[0031]
In each of the above-described embodiments, a plurality of forms corresponding to the business are set in the business information file F1, but the form group is not limited to the business, and may be arbitrary, for example, a department unit may be a form group. .
Further, the combination of the basic definition and the field definition is not limited to the above-described combining method, and is arbitrary.
In the second embodiment described above, the form is classified based on “the presence / absence of an edge mark”. However, the form may be classified based on “the presence / absence of a timing mark”.
[0032]
【The invention's effect】
According to the present invention, simply by specifying a form group, the definition information of a plurality of form formats associated with the specified group is called up and dynamically combined. By recognizing the form image based on the character and confirming the type, a series of processes from reading the form image to confirming the type can be executed in a single process, making it easy to determine the form type. It is possible to execute well.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the overall configuration of a data processing apparatus with a form recognition function.
FIG. 2 is a view showing various files stored in the data processing apparatus.
3A is a diagram conceptually showing the format of a form {circle around (1)}, FIG. 3 (B) shows the storage contents of a basic definition file F2 corresponding to the form {circle around (1)}, and FIG. The figure which showed the memory content of the field definition file F3 corresponding to 1 ▼.
4A is a diagram conceptually showing the format of a form {circle around (2)}, FIG. 4 (B) shows the storage contents of the basic definition file F2 corresponding to the form {circle around (2)}, and FIG. The figure which showed the memory content of the field definition file F3 corresponding to 2 ▼.
FIG. 5A is a diagram conceptually showing the format of a form {circle around (3)}, FIG. 5 (B) shows the storage contents of the basic definition file F2 corresponding to the form {circle around (3)}, and (C) is the form {circle around (7)}. The figure which showed the memory content of the field definition file F3 corresponding to 3 ▼.
6A is a diagram conceptually showing the format of a form {circle around (4)}, FIG. 6 (B) shows the storage contents of the basic definition file F2 corresponding to the form {circle around (4)}, and (C) is the form {circle around (7)}. The figure which showed the memory content of the field definition file F3 corresponding to 4 ▼.
7A shows an example of the contents of the business information file F1 and shows the forms (1), (2), and (3) used in the business (1), and (B) shows the business The composite basic definition file F4, (C) of (1) is a diagram showing a state in which the fields of the form (1), (2), (3) are conceptually superimposed, and (D) is a work The figure which illustrated the contents of synthetic field definition file F5 of 1 ▼.
FIG. 8A shows the contents of the business information file F1 and shows the forms (1) and (4) used in the business (2), and (B) shows the business (2). The composite basic definition file F4, (C) is a diagram showing a state in which the fields of the form (1), (4) are conceptually superimposed, and (D) is the composite field definition file F5 of the business (1). The figure which illustrated the contents of.
FIG. 9 is a diagram showing a data configuration of a recognition data file F7.
FIG. 10 is a flowchart showing the overall operation of form recognition processing.
FIG. 11 is a flowchart for explaining in detail the basic definition synthesis process shown in FIG. 10;
12 is a flowchart for explaining in detail the field definition synthesis process shown in FIG. 10;
13 is a flowchart for explaining the recognition result analysis process shown in FIG. 10 in detail.
FIG. 14 is a flowchart showing form recognition processing according to the second embodiment.
FIG. 15A is a diagram showing the contents of a business information file F1 corresponding to business (1), and FIG. 15 (B) is based on a classification consisting only of a form “1” with “edge mark”. The contents of the combined basic definition file combined are shown, (C) is a diagram showing a state in which the fields of the form combined corresponding to the classification are conceptually superimposed, and (D) is the classification The figure which illustrated the contents of synthetic field definition file F5 synthesized correspondingly.
FIG. 16A shows the contents of a combined basic definition file synthesized based on a classification including forms “2” and “4” of “with edge mark”, and FIG. The figure which showed the state which overlap | superposed the field of each form combined correspondingly conceptually, (C) is the figure which illustrated the content of the synthetic | combination field definition file F5 synthesize | combined corresponding to the said classification | category.
[Explanation of symbols]
1 CPU
2 storage devices
3 recording media
5 Transmission control unit
6 Input section
7 Display section
8 Image scanner
F1 business information file
F2 basic definition file
F3 field definition file
F4 composite basic definition file
F5 composite field definition file
F6 form image file
F7 recognition data file

Claims (2)

In the form recognition apparatus for recognizing the form image with reference to the form format composed of a plurality of fields set in advance, and determining the type of the form image according to the recognition state of character recognition for each form format,
A form definition storage means for storing field definition information consisting of area coordinates and character types of each field corresponding to each field of the form as a form format;
A form group storage means for storing a correspondence relationship between the form group and each form belonging to the group;
When a form group is designated, each form associated with the designated group is read from the form group storage means, and the field definition information defined corresponding to the form for each read form is A reading means for reading from the form definition storage means;
Definition synthesizing means for reading out the field definition information of each form read by the reading means and creating composite format definition information in which a plurality of field definition information in which the field definition information is sequentially written is stored as synthetic field definition information When,
A character recognition means for recognizing a form image based on the combined field definition information of the combined format definition information obtained by the definition combining means;
The form recognition means for dividing the recognition result recognized by the character recognition means into fields for each form, and for confirming the form in the field with the best recognition state of the divided recognition results as the form image type. A form recognition apparatus characterized in that a form type is determined by synthesizing field definition information of each form format corresponding to the designated group based on the designation of the form group. A computer of a form recognition apparatus that recognizes characters of a form image by referring to a form format composed of a plurality of fields set in advance and determines the type of the form image according to the recognition state of character recognition for each form format In addition,
A form definition storage means for storing, as a form format, field definition information consisting of field coordinates and character types corresponding to each field of the form for each form ;
Form group storage means for storing the correspondence between the form group and each form belonging to the group;
When a form group is designated, each form associated with the designated group is read from the form group storage means, and the field definition information defined corresponding to the form for each read form is Reading means for reading from the form definition storage means,
Definition synthesizing means for reading out the field definition information of each form read by the reading means and creating composite format definition information in which a plurality of field definition information in which the field definition information is sequentially written is stored as synthetic field definition information ,
Character recognition means for recognizing a form image based on the composite field definition information of the composite format definition information obtained by the definition composition means,
Functions as form type confirmation means that divides the recognition result recognized by the character recognition means into fields for each form, and confirms the form in the field with the best recognition state of the divided recognition results as the form image type. A computer-readable recording medium on which a program for causing the program to be recorded is recorded.

Images