JP2019191904A - Work support system, work support method, and program - Google Patents

Abstract

To reduce burdens in inspection work for forms.SOLUTION: A system comprises: an image data acquisition section for acquiring image data of a filled-out form; an entry content identification section for performing image processing and character recognition processing for the image data and identifying the entry content of the form; and an entry content inspection section for matching the entry content with a prescribed entry rule and inspecting whether or not the entry content conforms to the entry rule.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a business support system, a business support method, and a program.

  In recent years, a technology has been developed to construct a business support system by connecting IT devices such as personal computers used for business via a communication network so that daily business can be performed paperlessly (for example, patents). Reference 1).

JP 2017-021628 A

  However, there are businesses that handle handwritten forms, such as a window business at a financial institution and an application form reception business at a store. In such a case, the person in charge needs to input the contents of the handwritten form manually or by performing OCR processing into the business support system.

  The person in charge must check whether or not the contents entered in the form are appropriate, but it is necessary to check all the places where the form is filled, and the work load is large.

  The present invention has been made in view of the above problems, and an object of the present invention is to reduce the burden of inspection work on a form.

  An operation support system according to an embodiment of the present invention includes an image data acquisition unit that acquires image data of a completed form, performs image processing and character recognition processing on the image data, and specifies the entry contents of the form An entry content specifying unit for checking the entry content with a predetermined entry rule, and checking whether the entry content meets the entry rule.

  In addition, the problems disclosed by the present application and the solutions thereof will be clarified by the description in the column of the embodiment for carrying out the invention and the description of the drawings.

  It becomes possible to reduce the burden of the inspection work of the form.

1 is an overall configuration diagram of a system including a business support system according to the present embodiment. It is a figure which shows the memory | storage device of the work assistance system which concerns on this embodiment. It is a figure which shows the master table for rules which concerns on this embodiment. It is a figure which shows the software structure of the business assistance system which concerns on this embodiment. It is a figure which shows the business manual registration template which concerns on this embodiment. It is a figure which shows the workflow set using the business manual registration template which concerns on this embodiment. It is a figure which shows the outline | summary setting template which concerns on this embodiment. It is a figure which shows the detailed setting template which concerns on this embodiment. It is a figure which shows the AI model production | generation template which concerns on this embodiment. It is a figure which shows the preparation process performed with respect to the AI model which concerns on this embodiment. It is a figure which shows the procedure of the reading process of the form which concerns on this embodiment. It is a figure which shows the procedure of the reading process of the form which concerns on this embodiment. It is a figure which shows the function structure of the business assistance system which concerns on this embodiment. It is a figure which shows the reading process of the form which concerns on this embodiment. It is a figure which shows the check result of the form which concerns on this embodiment. It is a figure which shows a mode that the work assistance system which concerns on this embodiment detects use of a pencil. It is a figure which shows a mode that the business support system which concerns on this embodiment detects use of an ink penetration mark. It is a figure which shows a mode that the work assistance system which concerns on this embodiment detects use of a correction liquid. It is a flowchart which shows the flow of the process which concerns on this embodiment.

  At least the following matters will become apparent from the description of this specification and the accompanying drawings. Hereinafter, the present invention will be described in accordance with an embodiment thereof with reference to the accompanying drawings.

== Overall structure ==
FIG. 1 shows an overall configuration diagram of a system including a business support system 300 according to an embodiment of the present invention. The business support system 300 according to the present embodiment is communicably connected to the business terminal 100 and the scanner 200 through a network 500 such as the Internet, a LAN (Local Area Network), and a telephone network.

  As will be described in detail later, the business support system 300 analyzes image data obtained by digitizing a handwritten form 900 used in various business processes, recognizes the entry contents of the form 900, and uses a separately defined form 900. The information processing apparatus is capable of promoting work efficiency by checking whether there are any deficiencies in the contents of the form 900 with reference to the entry rules and processing according to the work workflow (execution order).

  The business support system 300 includes a workflow function that defines a business execution order, an AI-OCR function that digitizes a handwritten form 900 using artificial intelligence (AI), and a rule engine function that checks digital information. By linking these functions organically, it is possible to effectively support paperwork in various departments such as general affairs, human resources, and sales (see FIG. 5).

  Here, the scanner 200 is an image reading device that generates digital image data from a form 900 filled in by hand.

  The business terminal 100 is an information processing device such as a personal computer or a mobile terminal used by a person in charge when performing business processing.

  1 illustrates a configuration in which the scanner 200 is connected to the business support system 300 via the network 500 so that the scanner 200 can communicate with the business support system 300. However, when the business support system 300 or the business terminal 100 has the function of the scanner 200, FIG. The scanner 200 may not be provided individually.

  The business terminal 100 may be not only one but also a plurality. Alternatively, the person in charge may perform the paperwork directly using the business support system 300 without using the business terminal 100.

== Business support system ==
Next, the business support system 300 will be described.

  As illustrated in FIG. 1, the business support system 300 according to the present embodiment includes, as an example, a CPU 310, a memory 320, a communication device 330, a storage device 340, an input device 350, an output device 360, and a recording medium reading device 370. An information processing apparatus such as a computer is used.

  The storage device 340 is a nonvolatile storage device such as a hard disk drive or an SSD (Solid State Drive). As shown in FIG. 2, the storage device 340 includes a business support system control program 700 executed by the business support system 300, a business manual registration template 600, a rule definition template 610, a rule master table 613 (rule definition template). (Some elements included in 610), AI model generation template 620, form layout 630, and the like are stored (details will be described later).

  Various functions of the business support system 300 described later are realized by reading the business support system control program 700 and various data stored in the storage device 340 into the memory 320 and executing or processing them by the CPU 310.

  The recording medium reading device 370 reads the business support system control program 700 and various data recorded on the recording medium 800 such as a CD-ROM, DVD, USB (registered trademark) memory, and stores them in the storage device 340.

  The communication device 330 exchanges various data and programs with the business terminal 100, the scanner 200, and other computers (not shown) via the network 500. For example, the above-described business support system control program 700 and data such as templates are stored in another computer (not shown), and the business support system 300 downloads these data and the business support system control program 700 from this computer. can do.

  Alternatively, the above-described business manual registration template 600, rule definition template 610, AI model generation template 620, form layout 630, and rule master table 613 may be constructed in another computer.

  The input device 350 is a device for inputting data such as a keyboard and a mouse. The output device 360 is a device that outputs data such as a display or a printer.

  The business support system 300 may be realized by a single computer, or may be realized by a plurality of computers connected to be communicable with each other.

  Next, the business support system control program 700, the business manual registration template 600, the rule definition template 610, the AI model generation template 620, the form layout 630, and the rule master table 613 will be described with reference to FIGS.

<Business support system control program>
The business support system control program 700 has a workflow program 710, a rule engine program 720, an image processing program 730, and an AI model 740, as shown in FIG.

  The workflow program 710 is a program for realizing a workflow function in the business support system 300 according to the present embodiment. As described above, the workflow function is a function for managing the work procedure (workflow) of the paperwork, and provides a user interface for the person in charge to perform the work according to the workflow. In the configuration example illustrated in FIG. 1, a user interface is provided to the business terminal 100.

  Note that the workflow realized in the present embodiment conforms to BPMN (Business Process Model Notation) 2.0. BPMN 2.0 is a business process notation standardized by OMG (Object Management Group), a non-profit international consortium in which various companies, universities, and various non-profit organizations participate. Ability to use data related to processes and link with external systems, and easily share business processes between persons in different positions, such as office workers, managers, system developers, etc. It is said that.

  The workflow program 710 is generated using a business manual registration template 600 as shown in FIG.

  The business manual registration template 600 is a program for enabling generation of the workflow program 710 even without programming knowledge and skills.

  As shown in FIG. 5, the business manual registration template 600 according to the present embodiment accepts input of information specifying the work content of the paperwork in each cell of the spreadsheet software as an example. Then, the business manual registration template 600 is selected based on the information specifying the contents of the paperwork entered in each cell when the column labeled “Create workflow file” in FIG. 5 is selected and clicked with the mouse. A workflow program 710 is generated.

  Various work contents can be flexibly described in the business manual registration template 600. For example, a description that links the workflow function, the AI-OCR function, and the rule engine function with each other, or a description that branches subsequent office processing depending on the result of the work in progress are possible.

  In the example shown in FIG. 5, the business support system 300 analyzes the image data of the handwritten form 900 read by the scanner 200, checks and corrects the content by the person in charge, and then uses the rule engine function to A work process workflow is described in which the entered content is inspected, and if there is no error, a predetermined approval process is performed, and if there is an error, an action corresponding to the incomplete content is performed.

  The business support system 300 can also display the workflow described in the business manual registration template 600 in FIG. 5 as a visual flowchart as shown in FIG. (1) to (9) in FIG. 6 respectively correspond to No1 to No9 in FIG.

  The workflow shown in FIG. 5 and FIG. 6 will be described. First, in the business manual registration template 600 of FIG.

  No. 2 defines that OCR processing of the form 900 is performed in conjunction with the AI-OCR function. In other words, the business support system 300 acquires image data obtained by scanning the form 900 from the scanner 200, and performs image processing and character recognition processing of the image data using artificial intelligence, whereby each entry of the form 900 is performed. It is defined to specify (recognize) the contents of the column.

  In No. 3, it is defined that the reading result of the form 900 is displayed on the user interface so that the person in charge confirms it and, if necessary, is corrected.

  In No. 4, it is defined that the contents entered in the form 900 that has been inspected or modified by the operator are checked against a predetermined entry rule in conjunction with the rule engine function.

  In No. 5, branch processing is defined according to whether there is a defect.

  In No6 and No7, the process when there is no defect is defined.

  Moreover, in No8 and No9, the process when there is a defect is defined.

  The workflows shown in FIGS. 5 and 6 are examples, and the business support system 300 according to the present embodiment can flexibly define various workflows.

  Next, the rule engine program 720 shown in FIGS. 2 and 4 will be described.

  The rule engine program 720 is a program for realizing a rule engine function in the business support system 300 according to the present embodiment. The rule engine function is a function for inspecting information to be checked by applying check logic or conversion logic to the information to be checked. Specifically, it is a function for checking whether or not the contents of the form 900 match the entry rules.

  The rule engine program 720 is generated using a rule definition template 610 as shown in FIG.

  The rule definition template 610 is a program for enabling generation of the rule engine program 720 even without programming knowledge and skills.

  The rule definition template 610 according to the present embodiment includes an outline setting template 611 shown in FIG. 7 and a detailed setting template 612 shown in FIG.

  When generating the rule engine program 720, first, an entry rule to be set is input to the outline setting template 611 shown in FIG. In this embodiment, entry rules for each entry field of the form 900 are input.

  In the example shown in FIG. 7, an entry is made that it is determined that the form 900 is incomplete when a financial institution name or a branch name is not described in a predetermined entry column in the form 900 or when there is an error in the description content. Rules are set.

  Then, when the column “Create inspection rule details” is selected and clicked with a mouse or the like, the business support system 300 creates a detailed setting template 612 shown in FIG. 8.

  In the outline setting template 611, the entry rule of the form 900 is input in an expression close to the check work actually performed by the operator, but in the detailed setting template 612, the entry rule is expressed by a logical expression.

  Then, the operator checks the content of the detailed setting template 612, corrects or changes the content as appropriate, and selects and clicks the column “CSV output” with a mouse or the like. Then, the business support system 300 generates a rule engine program 720.

  As described above, when the rule engine program 720 is generated, the entry rule can be easily set by first creating the outline setting template 611. Further, by subsequently generating the detailed setting template 612, it becomes possible to accurately set the entry rule.

  Note that the summary setting template 611 shown in FIG. 7 and the detailed setting template 612 shown in FIG. 8 include a cell described as “financial institution TBL”, which is the rule master table 613 shown in FIGS. It is.

  As an example, the rule master table 613 is a table recording a list of financial institutions (various financial institutions including banks, credit unions, agricultural cooperatives, etc.) that can be entered in the form 900 as shown in FIG. Information about stores where accounts can be opened, such as head offices, branch offices, and branch offices of financial institutions, is also recorded in the rule master table 613.

  Next, the image processing program 730, the AI model 740, and the form layout 630 illustrated in FIGS. 2 and 4 will be described.

  The image processing program 730 and the AI model 740 are programs for realizing the AI-OCR function in the business support system 300 according to the present embodiment. The AI-OCR function is a function for recognizing characters and symbols entered in each entry column of the form 900 by analyzing the image data of the form 900 obtained from the scanner 200 using artificial intelligence (AI). It is. With the AI-OCR function, entry errors and omissions in the form 900 can be detected with high accuracy, and pencils, mechanical pencils, correction fluids, correction tapes, ink penetration marks (rubber stamps), etc. It is also possible to detect whether or not specific writing instruments and seals that are prohibited from being used are used.

  The image processing program 730 is a program for performing various image processing such as binarization, edge detection, color detection, and deformation on the image data obtained from the scanner 200.

  The AI model 740 is a program for recognizing the type of the form 900 and the contents of entry such as characters and symbols (circles and check marks) from the image data obtained from the scanner 200. As shown in FIG. It is generated using the generation template 620.

  The AI model generation template 620 is a program for enabling generation of the AI model 740 without programming knowledge or skills. In the present embodiment, CNN (Convolutional Neural Network) is adopted as the AI model.

  As shown in FIG. 9, the AI model generation template 620 specifies the structure of the AI model 740 (the number of layers, the number of nodes in the input layer, the intermediate layer, the output layer, etc.) for each cell of the spreadsheet software as an example. Accept input of information.

  When information specifying the structure of the AI model 740 such as the number of layers and the number of nodes is input to a predetermined input field of the AI model generation template 620, the business support system 300 generates a plurality of Python scripts, and these Python scripts Is used to generate an AI model 740.

  After the AI model 740 is generated in this way, various forms 900 are input to the AI model 740, and the types of the forms 900 and the characters to be entered in the forms 900 are learned.

  The form layout 630 is data representing a layout for each type of form 900 learned in this way. The form layout 630 also includes data representing the position of the entry field that is input by handwriting.

  Next, when the business support system 300 executes the image processing program 730 and the AI model 740, specific processing contents when reading the entry contents of the form 900 from the image data acquired from the scanner 200 are shown in FIGS. This will be described with reference to FIG.

  First, as illustrated in FIGS. 10A, 10 </ b> B, and 10 </ b> C, the business support system 300 performs learning of the form 900 using the AI model 740 and various settings of the form layout 630 and the image processing program 730.

  FIG. 10A shows that various forms 900 are input to the AI model 740 and learning is performed so that the business support system 300 can recognize the type of the form 900. The learning result is stored in the storage device 340 as a form layout 630.

  FIG. 10B shows that the position of the entry field to be handwritten is set in the form layout 630 for each type of form 900. This setting operation is performed by an operator.

  FIG. 10C shows parameter information such as various correction data for improving the accuracy of character recognition by correcting the color, size, distortion, etc. of the paper of the form 900, and check marks and circles in the form 900. In order to detect the use of writing tools and seals that are prohibited from being used in the form 900, such as correction information, pencils, ink penetrating marks, rubber stamps, etc., information indicating the position is scanned by the scanner. This indicates that color information specifying the color appearing in the image data when read at 200 is set. These pieces of information are set in the image processing program 730 as entry rules for the form 900.

  Note that the image processing program 730 according to the present embodiment can handle color information expressed in any of RGB (Red Green Blue) expression, HSV (Hue Saturation Value) expression, and HSB (Hue Saturation Brightness) expression. Any expression can be converted into another expression.

  According to such an aspect, color information can be defined in an expression that is easy for a color information setting operator to use, so that erroneous setting of color information can be prevented and more detailed setting work with high detection accuracy can be easily performed. It becomes possible.

  Next, as shown in FIGS. 11 (1) to 11 (5) and FIG. 12, the business support system 300 executes the image processing program 730 and the AI model 740, thereby filling the form 900 from the image data acquired from the scanner 200. Read the contents.

  Specifically, as shown in FIG. 11 (1), the business support system 300 first executes the image processing program 730 and uses the various parameter information set in FIG. The image data is distorted, tilted, and color corrected.

  Then, as shown in FIG. 11B, the business support system 300 executes the AI model 740 and identifies the type of the form 900 from the image data.

  Next, as shown in FIG. 11 (3), the business support system 300 executes the image processing program 730 and performs character recognition from the form 900 based on the information set in the form layout 630 in FIG. 10 (B). Specify the range (entry column).

  Then, as shown in FIGS. 10 (4) and 12, the business support system 300 executes the image processing program 730, and cuts out characters such as characters, circles, and dots that are entered in the entry field for character recognition.

  Next, as shown in FIGS. 11 (5) and 12, the business support system 300 executes the AI model 740, recognizes the extracted characters and symbols, and converts them into character codes.

  As described above, in the business support system 300 according to the present embodiment, the AI-OCR function can recognize handwritten characters in the free description field in the form 900 and can improve the accuracy of the character recognition rate. It has become.

<Functional configuration>
Next, a functional configuration diagram of the business support system 300 is shown in FIG. The business support system 300 according to the present embodiment has the functions of an image data acquisition unit 301, an entry content identification unit 302, a workflow control unit 303, a template unit 304, a confirmation result reception unit 305, and an entry content check unit 306.

  Each of these functions is realized by executing the business support system control program 700 according to the present embodiment by the hardware of the business support system 300.

  The image data acquisition unit 301 acquires image data of a form 900 in which predetermined items have been entered in each entry field. An example of the image data of the form 900 acquired by the image data acquisition unit 301 is shown in the column (1) “Recognition by AI-OCR” in FIG. The form 900 shown in (1) of FIG. 14 is a form 900 of the type “account transfer request form 4”.

  Next, the entry content specifying unit 302 will be described. The entry content specifying unit 302 specifies the entry contents in each entry column of the form 900 by performing image processing and character recognition processing using artificial intelligence on the image data of the form 900.

  Specifically, the entry content specifying unit 302 specifies the type of the form 900 (account transfer request form 4) read by the scanner 200 from the image data by using the AI model 740, and refers to the form layout 630 for the character. The entry field to be recognized is specified, and the characters entered in the entry field are recognized.

  The workflow control unit 303 defines the execution order of tasks to be performed using the form 900. For example, the workflow control unit 303 provides, for example, the business terminal 100 or the output device 360 with a user interface that allows the worker to perform business according to the execution order.

  The screens shown in (1) to (4) of FIG. 14 and FIG. 15 show user interfaces that the workflow control unit 303 presents to the business terminal 100 in accordance with the business execution order.

  (1) of FIG. 14 shows a state in which the workflow control unit 303 displays this image data according to the workflow when the image data acquisition unit 301 acquires the image data from the scanner 200.

  The “recognition result display & correction input” field in (2) of FIG. 14 shows a state in which the entry content of the form 900 specified by the entry content specifying unit 302 is displayed. The workflow control unit 303 displays the entry content of the form 900 as shown in (2) of FIG. 14 at the timing when the entry content specification unit 302 completes the specification of the entry content of the form 900.

  The operator can check whether there is an error in the reading result of the form 900 by looking at the display content.

  The confirmation result accepting unit 305 accepts the confirmation result by the operator of the contents entered in the form 900 presented on the user interface as shown in (2) of FIG. 14 by the workflow control unit 303. As shown in (2) of FIG. 14, when there is an error in the recognition result of the form 900 by the entry content specifying unit 302, the confirmation result receiving unit 305 can receive the correction of the content.

  The entry content checking unit 306 compares the entry content of the form 900 recognized by the entry content specifying unit 302 or the entry content reflecting the contents of correction by the operator with a predetermined entry rule. (3) “Automatic inspection by rule engine” column in FIG. 14).

  The automatic inspection by the entry content inspection unit 306 is started by an instruction from the workflow control unit 303 after the confirmation result reception unit 305 receives the confirmation result by the operator.

  Here, the entry rule to be collated with the entry contents of the form 900 is set not only in the entry rule (financial institution name, branch name, etc.) of the form 900 defined using the rule definition template 610 but also in the image processing program 730. The written entry rules (color information, etc.) are also included.

  Further, the entry content checking unit 306 displays an inspection result on the user interface as to whether or not the entry content of the form 900 conforms to the entry rule. This is shown in (4) “Check result check” column of FIG.

  In FIG. 15A, the use of the correction liquid is detected as a result of the inspection by the entry content inspection unit 306, and the message “NG” is output in the “Inspection result [image] No correction liquid used” column. It has been shown.

  FIG. 15B shows that the entry of the form 900 with a pencil is detected, and the message “NG” is output in the “Inspection result [image] No pencil writing” column.

  FIG. 15C shows that the use of the ink permeation mark is detected, and the message “NG” is output in the “inspection result [image] no use of seal mark” column.

  FIG. 15D shows that an incorrect entry date has been detected, and a message “The entry date is invalid” is output in the “Check result [required] entry date OK” column. Yes.

  In FIG. 15E, it is detected that the check points for selecting the financial institution are duplicated, and a message “Multiple financial institutions are selected” is displayed in the “Check result [correlation] selection of financial institution” column. It is shown that it is output.

  In FIG. 15F, a mismatch between the branch name and the branch code is detected, and a message “Branch code and branch name do not match” is output in the “Inspection result [correlation] branch code” column. It is shown.

  As described above, the image processing program 730 includes a specific type of writing instrument (pencil, correction liquid, correction tape, etc.) or a specific type of seal (ink penetration mark, etc.) that is prohibited from being used in the form 900. Color information specifying the color appearing in the image data when used is set as an entry rule.

  Therefore, the entry content checking unit 306 detects a color area (pixel) corresponding to the color information from each entry column of the form 900 specified by the entry content specifying unit 302, thereby detecting the specific type of writing instrument or It can be detected that the seal has been used.

  This will be described in detail with reference to FIGS.

  FIG. 16 is a diagram for explaining a case where the entry content inspection unit 306 detects the use of a pencil, and FIG. 17 is a diagram for explaining a case where the entry content inspection unit 306 detects the use of an ink penetration mark. FIG. 18 is a diagram for explaining a case where the entry content checking unit 306 detects the use of the correction fluid.

  FIG. 16A is an enlarged view of the name entry field in the image data of the form 900, and shows a state where the name is written in pencil in this entry field. Note that FIG. 16D shows a state in which a name is entered using a ballpoint pen, and the hue, saturation, and brightness are different from those of a written mark with a lead containing graphite, such as a pencil. It is shown.

  In FIG. 16B, the entry content checking unit 306 compares the color of each pixel in the entry field with the color information set in the image processing program 730, and only the pixels that match the color information of the pencil character are displayed. The state of extracting is shown.

  FIG. 16C shows a state in which the entry content checking unit 306 compares the number of pixels matching the color information with a threshold value. The entry content checking unit 306 determines that a pencil is being used when the number of pixels matching the color information exceeds a threshold value.

  By such an aspect, it becomes possible to detect pencil writing on the form 900 at high speed with simple logic.

  FIG. 17A is an enlarged view of the stamp field in the image data of the form 900, and shows a seal imprinted using an ink penetrating mark. Note that FIG. 17D shows a seal imprinted using vermilion, and the hue, saturation, and brightness are different from those of the ink penetrating seal.

  In FIG. 17B, the entry content checking unit 306 compares the color of each pixel in the stamp field with the color information set in the image processing program 730, and only the pixels that match the color information of the ink penetrating mark are displayed. The state of extracting is shown.

  FIG. 17C shows a state in which the entry content checking unit 306 compares the number of pixels matching the color information with a threshold value. When the number of pixels matching the color information exceeds the threshold value, the entry content checking unit 306 determines that the ink penetration mark has been used.

  By such an aspect, it is possible to detect the use of the ink penetrating mark on the form 900 at high speed with simple logic.

  FIG. 18A is an enlarged view of the name entry field in the image data of the form 900, and shows a state in which the correction liquid is used.

  In FIG. 18B, the entry content checking unit 306 compares the color of each pixel in this entry column with the color information set in the image processing program 730, and only the pixels that match the color information of the correction liquid are displayed. The extracted state is shown.

  FIG. 18C shows a state in which the entry content checking unit 306 compares the number of pixels that match the color information with a threshold value. When the number of pixels matching the color information exceeds the threshold, the entry content checking unit 306 determines that the correction liquid is used in the form 900.

  By such an aspect, it becomes possible to detect the use of the correction liquid for the form 900 at high speed with simple logic.

  In addition to the above, for example, a pencil for writing with a lead containing graphite or viscosity, like a pencil, a pen that can make ink transparent on paper, a correction tape, etc. When it is desired to detect a seal or the like, if the color information is set in the image processing program 730, the entry content checking unit 306 can use the same method as that shown in FIGS. It is possible to easily and rapidly detect these writing instruments and seals to be detected.

  Returning to FIG. 13, the template unit 304 accepts input of information for defining various functions to be realized by the business support system 300, and generates a program for realizing each function based on the information. To do.

  In this embodiment, the template unit 304 is embodied as a business manual registration template 600, a rule definition template 610, and an AI model generation template 620. The business manual registration template 600, the rule definition template 610, and the AI model generation template 620 receive information for defining functions to be realized by the business support system 300 in the spreadsheet software cell. Are used to generate a workflow program 710, a rule engine program 720, and an AI model 740, respectively.

  According to such an aspect, it is possible to easily generate a program for realizing the function of the business support system 300 even without programming knowledge and skills. As a result, for example, based on the requests of people in the world, design, build and maintain the inspection mechanism even when the flow of inspection and the flow of inspection including the level of inspection change frequently. It is possible to reduce the burden on the administrator and the entire organization.

== Flow of processing ==
Next, the flow of processing by the business support system 300 according to the present embodiment will be described with reference to the flowchart of FIG.

  First, the business support system 300 acquires image data of the form 900 (S1000). In this embodiment, the business support system 300 acquires the image data of the form 900 from the scanner 200. However, when the business support system 300 has a scanner function, the image of the form 900 read using this scanner function. Data may be acquired.

  Next, the business support system 300 performs image processing and character recognition processing using artificial intelligence on the image data, identifies the type of the form 900, and reads the entry contents in each entry column (S1010).

  Then, the business support system 300 displays the entry contents of the form 900 on the business terminal 100 or the output device 360 (S1020), and accepts a correction when there is an erroneous recognition (S1030).

  Thereafter, the business support system 300 uses the rule definition template 610 to set the form 900 entry rules (such as financial institution names) and the form 900 entry rules (color, saturation, brightness, etc.) set in the image processing program 730. In step S1040, the information on the form 900 is checked to see if it matches the entry rule (S1040).

  In this way, by comparing the contents of the form 900 with the color information, it is possible to easily detect the use of writing instruments and seals that are prohibited from being used, such as pencils, ink penetration marks, correction liquids, and the like. it can.

  The business support system 300 displays the inspection result on the business terminal 100 or the output device 360 (S1050).

  The business support system 300 activates each of the above processes ((S1000) to (S1050)) and subsequent processes according to a workflow set in advance.

  As described above, the business support system 300 according to the present embodiment has been described. However, according to the business support system 300 according to the present embodiment, the use of a pencil, an ink penetrating mark, a correction liquid, or the like that is prohibited on the form 900 is prohibited. It is possible to easily detect that a type of writing instrument or seal is being used. And the burden of the inspection work of the form 900 can be reduced.

  The embodiments described above are for facilitating the understanding of the present invention, and are not intended to limit the present invention. The present invention can be changed and improved without departing from the gist thereof, and equivalents thereof are also included in the present invention.

  Further, in the present invention, the discrimination processing target such as the use of a seal, writing instrument, and correction fluid based on the color information is limited to the handwritten form. It can also be an invention.

100 business terminal 200 scanner 300 business support system 301 image data acquisition unit 302 entry content specifying unit 303 workflow control unit 304 template unit 305 confirmation result receiving unit 306 entry content checking unit 310 CPU
320 Memory 330 Communication device 340 Storage device 350 Input device 360 Output device 370 Recording medium reading device 500 Network 600 Business manual registration template 610 Rule definition template 611 Outline setting template 612 Detailed setting template 613 Rule master table 620 AI model generation template 630 Form Layout 700 Business support system control program 710 Workflow program 720 Rule engine program 730 Image processing program 740 AI model 800 Recording medium 900 Form

Claims (12)

An image data acquisition unit for acquiring image data of a completed form;
An image content and character recognition process is performed on the image data, and an entry content specifying unit for specifying the entry content of the form,
An entry content inspection unit that checks the entry content with a predetermined entry rule and checks whether the entry content conforms to the entry rule;
A business support system. The business support system according to claim 1,
In addition to prescribing the execution order of the work to be performed using the form, and at the predetermined timing according to the execution order to allow the operator to check the entry content specified by the entry content specifying unit A workflow control unit for presenting the entry to the worker;
A business support system further comprising The business support system according to claim 2,
A confirmation result receiving unit for receiving a confirmation result by the worker for the entry content;
Further comprising
The work flow control unit causes the entry content check unit to check whether or not the entry content conforms to the entry rule after the confirmation result acceptance unit accepts the confirmation result. The business support system according to claim 3,
The confirmation result accepting unit accepts, as the confirmation result, the content of the correction made by the worker on the entry content,
The workflow control unit causes the entry content check unit to check whether or not the entry content reflecting the content of the correction conforms to the entry rule.
Business support system. The business support system according to any one of claims 1 to 4,
The entry rule includes color information that specifies the color that appears in the image data when a specific type of writing instrument or a specific type of seal that is prohibited from being used in the form is used,
The entry content check unit detects that the specific type of writing instrument or the specific type of seal has been used by detecting a color area corresponding to the color information from each entry column of the form. Support system. The business support system according to claim 5,
The work content support system, wherein the entry content check unit outputs a predetermined message when it is detected that the specific type of writing instrument or the specific type of seal is used. The business support system according to claim 5 or 6,
When the entry content check unit detects that the number of pixels in the color area corresponding to the color information in each entry field exceeds a threshold, the specific type of writing instrument or the specific type of seal is A business support system that detects usage. A business support system according to any one of claims 5 to 7,
The writing instrument of the specific type includes at least one of a pen having a graphite-containing core, a pen using a transparent ink, a correction liquid, and a correction tape,
The specific type of seal includes a business support system including an ink penetrating seal. The business support system according to any one of claims 5 to 8,
The color information is specified by any of RGB (Red Green Blue) expression, HSV (Hue Saturation Value) expression, or HSB (Hue Saturation Brightness) expression.
Business support system. A business support system according to any one of claims 1 to 9,
Further comprising a template unit that accepts input of information for defining a function to be realized by the business support system and generates a program to be executed by the business support system to realize the function based on the information; Support system. A business support method using a business support system that supports business performed using a form,
The business support system is
Get the image data of the completed form,
Perform image processing and character recognition processing on the image data, specify the entry content of the form,
A business support method of checking the entry content with a predetermined entry rule and checking whether the entry content conforms to the entry rule. To a computer that supports work done using forms,
The procedure to get the image data of the completed form,
A procedure for performing image processing and character recognition processing on the image data, and specifying the entry content of the form,
A procedure for checking the entry content against a predetermined entry rule and checking whether the entry content meets the entry rule;
A program for running