Information generation system, workflow system, information generation program, and workflow program

Abstract

This invention provides an information generation system, a workflow system, an information generation program, and a workflow program that facilitate the creation of templates for generating specific information from images in documents. [Solution] The information generation system includes a template generation unit that generates a template in S107 for generating specific information including information extracted from an invoice image, and a registration status change unit that, if a template in a provisional registration state is stored in the template database, changes the status of this template to a fully registered state in accordance with instructions from the user. The template generation unit stores the template it has generated in the template database in a provisional registration state (S108), and if a template in a provisional registration state is stored in the template database, it notifies a specific contact person of an instruction prompting them to change the status of this template to a fully registered state.

Description

Technical Field

[0001] The present invention relates to an information generation system, a workflow system, an information generation program, and a workflow program for generating specific information from an image of a document.

Background Art

[0002] Conventionally, a technique for performing character recognition (character extraction analysis) on a position defined by a template in which the position on a paper surface where characters are entered is predefined has been known (see Patent Document 1). This technique identifies the form ID, which is an identifier of the template definition information, by printing or entering it on the paper surface, and performs OCR (Optical Character Recognition) processing using the corresponding template by identifying the form ID of the document converted into image data.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the conventional technology, there is a problem that it is very laborious to create a template for generating specific information from an image of a document.

[0005] Therefore, an object of the present invention is to provide an information generation system, a workflow system, an information generation program, and a workflow program that can facilitate the creation of a template for generating specific information from an image of a document.

Means for Solving the Problems

[0006] The information generation system of the present invention comprises a document classification unit that classifies documents according to their format, a template-type information generation unit that generates information from an image using a template for generating specific information from an image of a document, a general-purpose information generation unit that generates information from an image in a general manner regardless of the format, and a template generation unit that generates the template. The template-type information generation unit generates information from an image using the template that corresponds to the format identified by the document classification unit when the format of the document corresponding to the image is identified by the document classification unit, and the template that corresponds to the format identified by the document classification unit is managed by a template management unit that manages the templates. The template generation unit generates the template that corresponds to the format identified by the document classification unit when the format of the document corresponding to the image is identified by the document classification unit, and the general-purpose information generation unit generates the information from the image based on the image, the information, and the template managed by the template management unit.

[0007] With this configuration, the information generation system of the present invention generates a template corresponding to the format of a document based on an image of the document, information generated from the image of the document by a general-purpose information generation unit, and templates managed by a template management unit. This eliminates the need for templates corresponding to the format of the document to be created solely by human effort. In other words, the information generation system of the present invention can facilitate the creation of templates for generating specific information from an image of a document.

[0008] In the information generation system of the present invention, the general-purpose information generation unit may generate the information from the image by utilizing an external service of the information generation system that generates the information from the image in a general manner, regardless of the format.

[0009] This configuration eliminates the need for the information generation system of the present invention to pre-generate a learning model for generating specific information from document images by the provider or user of the information generation system, thereby reducing the burden on the provider or user of the information generation system.

[0010] The workflow system of the present invention comprises a workflow generation unit that generates a workflow representing a series of processes by linking any of a plurality of blocks, each representing a specific process, and a workflow execution unit that executes the processes shown in the workflow generated by the workflow generation unit, wherein the workflow generation unit is capable of generating a workflow that realizes a document classification unit that classifies the documents according to the document format, a template-type information generation unit that generates specific information from the images of the documents using a template for generating the information from the images of the documents, a general-purpose information generation unit that generates the information from the images in a general manner regardless of the format, and a template generation unit that generates the template, and the template-type information generation unit generates the template of the documents corresponding to the images - When a mat is identified by the document classification unit, and the template corresponding to the format identified by the document classification unit is managed by the template management unit that manages the templates, the information is generated from this image using the template corresponding to the format identified by the document classification unit from among the templates managed by the template management unit. The template generation unit is characterized in that, when the format of the document corresponding to the image is identified by the document classification unit, and the general-purpose information generation unit generates the information from this image, the template generation unit generates the template corresponding to the format identified by the document classification unit based on this image, this information, and the template managed by the template management unit.

[0011] With this configuration, the workflow system of the present invention can generate a template corresponding to the format of a document based on the image of the document, the information generated from the image of the document by the general-purpose information generation unit, and the templates managed by the template management unit. This eliminates the need for templates corresponding to the format of the document to be created solely by human effort. In other words, the workflow system of the present invention can facilitate the creation of templates for generating specific information from the image of a document.

[0012] The information generation program of the present invention provides a computer with a document classification unit that classifies documents according to their format, a template-type information generation unit that generates information from an image using a template for generating specific information from an image of a document, a general-purpose information generation unit that generates information from an image in a general manner regardless of the format, and a template generation unit that generates the template. The template-type information generation unit generates information from an image using the template that corresponds to the format identified by the document classification unit when the format of the document corresponding to the image is identified by the document classification unit, and the template that corresponds to the format identified by the document classification unit is managed by a template management unit that manages the templates. The template generation unit generates the template that corresponds to the format identified by the document classification unit when the format of the document corresponding to the image is identified by the document classification unit, and the general-purpose information generation unit generates the information from the image based on the image, the information, and the template managed by the template management unit.

[0013] With this configuration, the computer executing the information generation program of the present invention generates a template corresponding to the format of the document based on the image of the document, the information generated from the image of the document by the general-purpose information generation unit, and the templates managed by the template management unit. This eliminates the need for the template corresponding to the format of the document to be created solely by human effort. In other words, the computer executing the information generation program of the present invention can facilitate the creation of templates for generating specific information from the image of the document.

[0014] The workflow program of the present invention enables a computer to implement a workflow that includes a workflow generation unit that generates a workflow representing a series of processes by linking any of a plurality of blocks, each representing a specific process, and a workflow execution unit that executes the processes shown in the workflow generated by the workflow generation unit. The workflow generation unit is capable of generating a workflow that includes a document classification unit that classifies documents according to their format, a template-type information generation unit that generates specific information from images using templates for generating information from images of documents, a general-purpose information generation unit that generates information from images in a general manner regardless of the format, and a template generation unit that generates the templates. The template-type information generation unit generates the documents corresponding to the images. When the format of the document corresponding to the image is identified by the document classification unit, and the template corresponding to the format identified by the document classification unit is managed by the template management unit that manages the templates, the information is generated from the image using the template corresponding to the format identified by the document classification unit from among the templates managed by the template management unit. The template generation unit is characterized in that, when the format of the document corresponding to the image is identified by the document classification unit, and the general-purpose information generation unit generates the information from the image, the template generation unit generates the template corresponding to the format identified by the document classification unit based on the image, the information, and the template managed by the template management unit.

[0015] With this configuration, a computer running the workflow program of the present invention can generate a template corresponding to the format of a document based on the document image, the information generated from the document image by the general-purpose information generation unit, and the templates managed by the template management unit. This eliminates the need for templates corresponding to the document format to be created solely by human effort. In other words, a computer running the workflow program of the present invention can facilitate the creation of templates for generating specific information from document images. [Effects of the Invention]

[0016] The information generation system, workflow system, information generation program, and workflow program of the present invention can facilitate the creation of templates for generating specific information from images of documents. [Brief explanation of the drawing]

[0017] [Figure 1] This is a block diagram of an information processing system according to one embodiment of the present invention. [Figure 2] Figure 1 is a block diagram of the information generation system when it is configured using a single computer. [Figure 3] Figure 1 is a block diagram of the image reading device in the case of an MFP (Multi-Function Printer). [Figure 4] Figure 1 is a block diagram of the user terminal. [Figure 5] Figure 2 shows a flowchart illustrating the operation of the information generation system when generating specific information from an invoice image. [Figure 6] Figure 2 is a flowchart illustrating the operation of the information generation system when generating specific information from an invoice image using a template. [Figure 7] Figure 5 shows a flowchart of the template generation process. [Figure 8]It is a flowchart of the operation of the information generation system shown in FIG. 2 when generating a template from a claim image. [Figure 9] It is a block diagram of a workflow system that realizes the information generation system shown in FIG. 2. [Figure 10] It is a diagram showing an example of a workflow shown on a workflow generation screen displayed on the display unit of the user terminal shown in FIG. 1.

Embodiments for Carrying Out the Invention

[0018] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0019] First, the configuration of the information processing system according to an embodiment of the present invention will be described.

[0020] FIG. 1 is a block diagram of an information processing system 10 according to the present embodiment.

[0021] As shown in FIG. 1, the information processing system 10 includes an information generation system 20 that generates specific information (hereinafter referred to as "specific information") including information extracted from an image of a claim as a document (hereinafter referred to as "claim image"), an image reading device 30 that reads the claim image from an actual claim (hereinafter referred to as "actual claim") constituted by a recording medium such as paper, and a user terminal 40 used by a user of the information processing system 10. The information processing system 10 may include at least one image reading device similar to the image reading device 30 in addition to the image reading device 30. The information processing system 10 may include at least one user terminal similar to the user terminal 40 in addition to the user terminal 40.

[0022] The specific information is information in which an item name and a value are associated for each item.

[0023] The information generation system 20 may be constituted by one computer or may be constituted by a plurality of computers.

[0024] The image reading device 30 is composed of, for example, a dedicated scanner, an MFP (Multifunction Peripheral), and the like.

[0025] The user terminal 40 is comprised of a computer, such as a PC (Personal Computer).

[0026] The information generation system 20 and the image reading device can communicate via a network such as a LAN (Local Area Network) or the Internet, or directly via wired or wireless connections without using a network. Similarly, the information generation system 20 and the user terminal can communicate via a network such as a LAN or the Internet, or directly via wired or wireless connections without using a network.

[0027] Figure 2 is a block diagram of the information generation system 20 when it is configured with a single computer.

[0028] The information generation system 20 shown in Figure 2 comprises an operation unit 21, which is an operation device such as a keyboard or mouse that receives various operations; a display unit 22, which is a display device such as an LCD (Liquid Crystal Display) that displays various information; a communication unit 23, which is a communication device that communicates with external devices via a network such as a LAN or the Internet, or directly via wired or wireless connection without going through a network; a storage unit 24, which is a non-volatile storage device such as a semiconductor memory or HDD (Hard Disk Drive) that stores various types of information; and a control unit 25 that controls the entire information generation system 20.

[0029] The memory unit 24 stores an information generation program 24a for generating specific information. The information generation program 24a may be installed in the information generation system 20 during the manufacturing stage, or it may be additionally installed in the information generation system 20 from an external storage medium such as a USB (Universal Serial Bus) memory, or it may be additionally installed in the information generation system 20 from a network.

[0030] The memory unit 24 can store an invoice classification model 24b, which serves as a learning model for classifying invoices according to their format (hereinafter simply referred to as "format"). The invoice classification model 24b is created by machine learning.

[0031] The storage unit 24 can store a template database 24c that stores templates for generating specific information from an invoice image. The template database 24c manages templates for each format and constitutes the template management unit of the present invention. Each template includes, for each item of information to be extracted from the invoice image, the item name of the information to be extracted from the invoice image, an anchor that serves as the reference for the position from which the information to be extracted from the invoice image is extracted, the position of the area where the anchor exists (hereinafter referred to as the "anchor area") (hereinafter referred to as the "anchor area position"), and the position of the area where the information to be extracted from the invoice image exists (hereinafter referred to as the "extraction area") relative to the anchor (hereinafter referred to as the "extraction area position"). Here, the anchor may be, for example, a specific string of characters or a specific image such as a logo. The template is defined, for example, in XML (Extensible Markup Language) format.

[0032] The memory unit 24 can store an information generation model 24d, which serves as a learning model for generating specific information from invoice images. The information generation model 24d is created by machine learning. The information generation model 24d is a format-independent, general-purpose learning model that can extract information and the location of this information on the invoice image from the invoice image.

[0033] The memory unit 24 can store a dictionary 24e that shows various representations of the item names of the information extracted from the invoice image. For example, the dictionary 24e shows "invoice_no" and "invoice_number" as the item names for the item indicating "invoice number," "invoice_date" as the item name for the item indicating "invoice date," and "total_amount" and "total" as the item names for the item indicating "invoice amount."

[0034] The control unit 25 includes, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory) that stores programs and various data, and a RAM (Random Access Memory) used as a workspace for the CPU of the control unit 25. The CPU of the control unit 25 executes programs stored in the storage unit 24 or the ROM of the control unit 25.

[0035] The control unit 25, by executing the information generation program 24a, realizes an invoice classification unit 25a as a document classification unit that classifies invoices by format, a template-type information generation unit 25b that generates specific information from invoice images using templates, a general-purpose information generation unit 25c that generates specific information from invoice images in a general manner regardless of format, and a template generation unit 25d that generates templates.

[0036] Figure 3 is a block diagram of the image reading device 30 when it is an MFP (Multi-Function Printer).

[0037] The image reading device 30 shown in Figure 3 comprises an operation unit 31, which is an operation device such as a button into which various operations are input; a display unit 32, which is a display device such as an LCD that displays various information; a printer 33, which is a printing device that prints images onto a recording medium such as paper; a scanner 34, which is a reading device that reads images from an original document; a fax communication unit 35, which is a fax device that performs fax communication with an external facsimile device (not shown) via a communication line such as a public telephone line; a communication unit 36, which is a communication device that communicates with an external device via a network such as a LAN or the Internet, or directly via wired or wireless connection without going through a network; a storage unit 37, which is a non-volatile storage device such as a semiconductor memory or HDD that stores various information; and a control unit 38 that controls the entire image reading device 30.

[0038] The control unit 38 includes, for example, a CPU, a ROM that stores programs and various data, and RAM as memory used as a workspace for the CPU of the control unit 38. The CPU of the control unit 38 executes programs stored in the storage unit 37 or the ROM of the control unit 38.

[0039] Figure 4 is a block diagram of the user terminal 40.

[0040] The user terminal 40 shown in Figure 4 comprises an operation unit 41 which is an operation device such as a keyboard or mouse into which various operations are input; a display unit 42 which is a display device such as an LCD that displays various information; a communication unit 43 which is a communication device that communicates with external devices via a network such as a LAN or the Internet, or directly via wired or wireless connection without going through a network; a storage unit 44 which is a non-volatile storage device such as a semiconductor memory or HDD that stores various information; and a control unit 45 that controls the entire user terminal 40.

[0041] The control unit 45 includes, for example, a CPU, a ROM that stores programs and various data, and RAM as memory used as a workspace for the CPU of the control unit 45. The CPU of the control unit 45 executes programs stored in the storage unit 44 or the ROM of the control unit 45.

[0042] Next, the operation of the information processing system 10 will be described.

[0043] First, we will explain the operation of the information generation system 20 when generating specific information from an invoice image.

[0044] Figure 5 is a flowchart showing the operation of the information generation system 20 when generating specific information from an invoice image.

[0045] The user places the actual invoice into the scanner of the image reading device and can instruct the image reading device to generate specific information from the actual invoice, for example, via the operation unit of the image reading device. When the control unit of the image reading device receives an instruction to generate specific information from the actual invoice, it reads an invoice image from the actual invoice placed in the scanner of the image reading device using the scanner and instructs the information generation system 20 via the communication unit of the image reading device to generate specific information from this invoice image. Here, the instruction to generate specific information from the invoice image (hereinafter referred to as "information generation instruction") includes the invoice image that is the target of this information generation instruction. When the control unit 25 of the information generation system 20 receives the information generation instruction via the communication unit 23, it executes the operation shown in Figure 5.

[0046] As shown in Figure 5, the invoice classification unit 25a identifies the invoice format corresponding to the invoice image targeted by the information generation instruction, using the invoice classification model 24b based on this invoice image (S101).

[0047] Next, the template-type information generation unit 25b determines whether or not a template corresponding to the format identified in S101 exists in the template database 24c (S102).

[0048] If the template-type information generation unit 25b determines in S102 that a template corresponding to the format identified in S101 exists in the template database 24c, it uses the template stored in the template database 24c that corresponds to the format identified in S101 to generate specific information from the invoice image that is the target of the information generation instruction (S103). For example, if the template stored in the template database 24c that corresponds to the format identified in S101 defines "invoice_number," "invoice_date," and "total" as the item names for the items indicating "invoice number," "invoice date," and "invoice amount," respectively, the template-type information generation unit 25b extracts the values ​​of each item in S103 and generates specific information by associating the extracted values ​​with the item names defined in the template.

[0049] Figure 6 is a flowchart showing the operation of the information generation system 20 when generating specific information from an invoice image using a template.

[0050] As shown in Figure 6, the template-type information generation unit 25b targets one item from the items included in the target template that has not yet been targeted in the operation shown in Figure 6 (S121).

[0051] Next, the template-type information generation unit 25b detects anchors included in the target template for the current target item in the target invoice image from the anchor regions included in the target template for the current target item (S122).

[0052] Next, the template-type information generation unit 25b identifies the extraction region for the current target item in the target invoice image based on the anchor detected in S122 and the extraction region position included in the target template for the current target item (S123).

[0053] Next, the template-type information generation unit 25b extracts information from the extraction region identified in S123 in the target invoice image (S124).

[0054] Next, the template-type information generation unit 25b associates the information extracted in S124 with the item names included in the target template for the current target item (S125).

[0055] Next, the template-type information generation unit 25b determines whether or not there are any items included in the target template that have not yet been targeted in the operation shown in Figure 6 (S126).

[0056] If the template-type information generation unit 25b determines in S126 that there are items included in the target template that have not yet been targeted in the operation shown in Figure 6, it executes the process in S121.

[0057] If the template-type information generation unit 25b determines in S126 that there are no items in the target template that have not yet been targeted in the operation shown in Figure 6, it terminates the operation shown in Figure 6.

[0058] As shown in Figure 5, after processing in S103, the template-type information generation unit 25b stores the specific information generated in S103 in the storage unit 24, associating it with the format specified in S101 and the invoice image that is the target of the information generation instruction (S104). Therefore, the user can access the information generation system 20 using, for example, a user terminal, and confirm the specific information stored in the storage unit 24 via the display unit of the user terminal, and can also modify this specific information as needed via the operation unit of the user terminal.

[0059] After processing in S104, the template-type information generation unit 25b terminates the operation shown in Figure 5.

[0060] If the general-purpose information generation unit 25c determines in S102 that a template corresponding to the format identified in S101 does not exist in the template database 24c, it generates specific information from the invoice image targeted by the information generation instruction using the information generation model 24d (S105). For example, if the information generation model 24d defines "invoice_no", "invoice_date", and "total_amount" as the item names for the items indicating "invoice number", "invoice date", and "invoice amount", respectively, the general-purpose information generation unit 25c extracts the values ​​for each item in S105 and generates specific information by associating the extracted values ​​with the item names defined in the information generation model 24d.

[0061] After processing in S105, the general-purpose information generation unit 25c stores the specific information generated in S105 in the storage unit 24, associating it with the format specified in S101 and the invoice image that is the target of the information generation instruction (S106). Therefore, a user can, for example, access the information generation system 20 using a user terminal to confirm the specific information stored in the storage unit 24 via the display unit of the user terminal, and can also modify this specific information as needed via the operation unit of the user terminal.

[0062] After the processing in S106, the template generation unit 25d performs a template generation process to generate a template based on the specific information generated in S105 (S107).

[0063] Figure 7 is a flowchart of the template generation process shown in Figure 5.

[0064] As shown in Figure 7, the template generation unit 25d retrieves item names from all templates stored in the template database 24c (S141). For example, if the templates stored in the template database 24c have "invoice_number", "invoice_date", and "total" defined as item names for items indicating "invoice number", "invoice date", and "invoice amount", respectively, the template generation unit 25d retrieves "invoice_number", "invoice_date", and "total" in S141.

[0065] After processing in S141, the template generation unit 25d obtains the item names in the specific information generated in S105 using the information generation model 24d (S142). For example, in S142, the template generation unit 25d obtains "invoice_no", "invoice_date", and "total_amount".

[0066] After processing in S142, the template generation unit 25d identifies, based on dictionary 24e, which of the item names obtained in S142, i.e., the item names in the specific information generated using the information generation model 24d, corresponds to any of the item names obtained in S141, i.e., the item names obtained from all templates (S143). For example, the template generation unit 25d identifies that the item names "invoice_no", "invoice_date", and "total_amount" obtained in S142 correspond to the item names "invoice_number", "invoice_date", and "total", respectively, obtained in S141.

[0067] After processing in S143, the template generation unit 25d identifies the anchor, anchor region position, and extraction region position for each item name identified in S143 from the invoice image targeted by the information generation instruction (S144). Here, the template generation unit 25d may use any string shown in dictionary 24e for the target item name as the anchor if that string exists in the invoice image targeted by the information generation instruction. If none of the strings shown in dictionary 24e for the target item name exists in the invoice image targeted by the information generation instruction, the template generation unit 25d may use any information present in the invoice image, such as a logo, as the anchor. Furthermore, the template generation unit 25d uses a specific region containing the region from which the values ​​of each item were extracted in S105 as the extraction region.

[0068] After processing in S144, the template generation unit 25d generates a template (S145) by converting the item names identified in S143 to the item names obtained in S141, and including the anchor, anchor region position, and extraction region position identified in S144 for each item, and then terminates the template generation process shown in Figure 7.

[0069] As shown in Figure 5, after the template generation process in S107, the template generation unit 25d stores the template generated in S107 in the template database 24c as a template in the format specified in S101 (S108). Therefore, a user can access the information generation system 20 using, for example, a user terminal, and confirm the template stored in the template database 24c via the display unit of the user terminal, and can also modify this template as needed via the operation unit of the user terminal. Note that the template stored in the template database 24c in S108 is in a provisional registration state. When a user confirms this template via the display unit of the user terminal, they can change the status of this template to a fully registered state via the operation unit of the user terminal. If a template in a provisional registration state is stored in the template database 24c, the template generation unit 25d can notify a specific contact prompting them to change the status of this template to a fully registered state.

[0070] After processing in S108, the template generation unit 25d completes the operation shown in Figure 5.

[0071] Next, we will describe the operation of the information generation system 20 when generating a template from an invoice image.

[0072] Figure 8 is a flowchart showing the operation of the information generation system 20 when generating a template from an invoice image.

[0073] The user can place the actual invoice into the scanner of the image reading device and, for example, instruct the image reading device to generate a template from the actual invoice via its operation unit. When the control unit of the image reading device receives an instruction to generate a template from the actual invoice, it reads the invoice image from the actual invoice placed in the scanner of the image reading device using the scanner and instructs the information generation system 20, via the communication unit of the image reading device, to generate a template from this invoice image. Here, the instruction to generate a template from an invoice image (hereinafter referred to as the "template generation instruction") includes the invoice image that is the target of this template generation instruction. When the control unit 25 of the information generation system 20 receives the template generation instruction via the communication unit 23, it executes the operation shown in Figure 8.

[0074] As shown in Figure 8, the invoice classification unit 25a identifies the invoice format corresponding to the invoice image targeted by the template generation instruction, using the invoice classification model 24b based on this invoice image (S161).

[0075] Next, the general-purpose information generation unit 25c generates specific information from the invoice image targeted by the template generation instruction using the information generation model 24d, similar to the process in S105 (S162).

[0076] Next, the template generation unit 25d performs a template generation process to generate a template, similar to the process in S107, based on the specific information generated in S162 (S163).

[0077] Next, the template generation unit 25d stores the template generated in S163 in the storage unit 24 as a template candidate (hereinafter referred to as "template candidate") in the format specified in S161 (S164). Therefore, the user can, for example, access the information generation system 20 using a user terminal to confirm the template candidate stored in the storage unit 24 via the display unit of the user terminal, and can modify this template candidate as needed via the operation unit of the user terminal. When the user confirms this template candidate via the display unit of the user terminal, the user can, for example, instruct the information generation system 20 to adopt this template candidate as a template via the operation unit of the user terminal.

[0078] After processing in S164, the template generation unit 25d determines whether or not it has been instructed to adopt the template candidate saved in S164 as a template, until it determines that it has been instructed to adopt the template candidate saved in S164 as a template (S165).

[0079] If the template generation unit 25d determines in S165 that it has been instructed to adopt the template candidate saved in S164 as a template, it stores the template candidate saved in S164 as a template in the format specified in S161 in the template database 24c (S166), and then terminates the operation shown in Figure 8.

[0080] As described above, the information generation system 20 generates a template corresponding to the invoice format corresponding to the invoice image based on the invoice image, the specific information generated from the invoice image by the general-purpose information generation unit 25c, and the templates managed in the template database 24c (S107 and S163). Therefore, the need for the template corresponding to this invoice format to be created solely by human effort is eliminated. In other words, the information generation system 20 can facilitate the creation of templates for generating specific information from invoice images.

[0081] Even if a template corresponding to the invoice format corresponding to the invoice image targeted by the information generation instruction does not exist in the template database 24c (NO in S102), the information generation system 20 generates specific information from this invoice image using the general-purpose information generation unit 25c (S105). This eliminates the need to manually generate specific information from this invoice image, and as a result, the time required to generate specific information from the invoice image can be shortened.

[0082] The information generation system 20 may also be implemented by a workflow system 50 (see Figure 9) that executes the processes shown in the workflow representing a series of processes. The workflow system 50 may consist of one computer or multiple computers, as shown in Figure 9.

[0083] Figure 9 is a block diagram of the workflow system 50 that realizes the information generation system 20.

[0084] The workflow system 50 shown in Figure 9 comprises an operation unit 51 which is an operation device such as a keyboard or mouse into which various operations are input; a display unit 52 which is a display device such as an LCD that displays various information; a communication unit 53 which is a communication device that communicates with external devices via a network such as a LAN or the Internet, or directly via wired or wireless connection without going through a network; a storage unit 54 which is a non-volatile storage device such as a semiconductor memory or HDD that stores various information; and a control unit 55 that controls the entire workflow system 50.

[0085] The storage unit 54 stores a workflow program 54a for executing the processes shown in the workflow. The workflow program 54a may be installed in the workflow system 50 during the manufacturing stage, or it may be additionally installed in the workflow system 50 from an external storage medium such as a USB memory stick, or it may be additionally installed in the workflow system 50 from a network.

[0086] The memory unit 54 is capable of storing the invoice classification model 24b, the template database 24c, the information generation model 24d, and the dictionary 24e.

[0087] The control unit 55 includes, for example, a CPU, a ROM that stores programs and various data, and RAM as memory used as a workspace for the CPU of the control unit 55. The CPU of the control unit 55 executes programs stored in the storage unit 54 or the ROM of the control unit 55.

[0088] The control unit 55 implements a workflow generation unit 55a that generates a workflow by executing a workflow program 54a, thereby linking arbitrary blocks from a plurality of blocks, each representing a specific process, and a workflow execution unit 55b that executes the processes shown in the workflow generated by the workflow generation unit 55a.

[0089] The workflow system 50 allows workflow generation to be instructed by the user, for example, via the user terminal's operation unit. For example, the workflow generation unit 55a can display a workflow generation screen on the user terminal's display unit, and in response to instructions from the user via the user terminal's operation unit on the workflow generation screen, it can generate workflows that implement the invoice classification unit 25a, the template-type information generation unit 25b, the general-purpose information generation unit 25c, and the template generation unit 25d. In other words, the workflow generation unit 55a can generate workflows that implement the operations shown in Figure 5 and workflows that implement the operations shown in Figure 8.

[0090] Figure 10 shows an example of a workflow displayed on the workflow generation screen shown on the user terminal's display unit.

[0091] The workflow 60 shown in Figure 10 is an example of a workflow that realizes the operation shown in Figure 5. Workflow 60 is generated by linking together block 61 which shows the input of an invoice image from an image reader, block 62 which shows the processing of S101 to S103, block 63 which shows the processing of S104, block 64 which shows the storage unit 24, block 65 which shows the processing of S105, block 66 which shows the processing of S106, block 67 which shows the storage unit 24, block 68 which shows the processing of S107, block 69 which shows the processing of S108, and block 70 which shows the template database 24c.

[0092] The workflow system 50 shown in Figure 9 allows the user to confirm the workflow generated by the workflow generation unit 55a via the display unit of the image reader, and also allows the user to instruct the execution of the workflow generated by the workflow generation unit 55a via the operation unit of the image reader.

[0093] The user can place the actual invoice into the scanner of the image reader and, for example, instruct the image reader to execute the workflow shown in Figure 5, for example, via the operation unit of the image reader. When the control unit of the image reader receives an instruction to execute this workflow, it reads an invoice image from the actual invoice placed in the scanner of the image reader and transmits the instruction to execute this workflow, along with the read invoice image, to the workflow system 50 via the communication unit of the image reader. When the workflow execution unit 55b of the workflow system 50 receives this instruction to execute the workflow via the communication unit 23, it executes the operations shown in Figure 5.

[0094] Similarly, the user can place the actual invoice into the scanner of the image reader and, for example, instruct the image reader to execute the workflow shown in Figure 8, for example, via the operation unit of the image reader. When the control unit of the image reader is instructed to execute this workflow, it reads the invoice image from the actual invoice placed in the scanner of the image reader and transmits the instruction to execute this workflow, along with the read invoice image, to the workflow system 50 via the communication unit of the image reader. When the workflow execution unit 55b of the workflow system 50 receives this instruction to execute the workflow via the communication unit 23, it executes the operations shown in Figure 8.

[0095] In the above description, the general-purpose information generation unit 25c generates specific information from the invoice image using the information generation model 24d (S105 and S162). However, the general-purpose information generation unit 25c may generate specific information from the invoice image without using the information generation model 24d. For example, the general-purpose information generation unit 25c may generate specific information from the invoice image using an external service of the information generation system 20, such as a cloud service that generates specific information from the invoice image in a general manner regardless of format. When the general-purpose information generation unit 25c generates specific information from the invoice image using an external service of the information generation system 20, the information generation model 24d does not need to be generated in advance by the provider or user of the information generation system 20, thus reducing the burden on the provider or user of the information generation system 20. Furthermore, in the information generation system 20, the template generation unit 25d generates a template based on the specific information generated from the invoice image by the general-purpose information generation unit 25c. Therefore, even if the general-purpose information generation unit 25c uses an external service to generate specific information from the invoice image, and it is necessary to pay a usage fee to the external service provider depending on the use of the external service, the number of times the general-purpose information generation unit 25c uses an external service to generate specific information from the invoice image can be reduced, and as a result, the usage fee paid to the external service provider can be reduced.

[0096] In the above description, the template generation unit 25d identifies, based on dictionary 24e, the item names in the specific information generated by the general-purpose information generation unit 25c that correspond to any of the item names obtained from all templates stored in the template database 24c (S143). However, the template generation unit 25d may identify, without using dictionary 24e, the item names in the specific information generated by the general-purpose information generation unit 25c that correspond to any of the item names obtained from all templates stored in the template database 24c. For example, the template generation unit 25d may identify, using a machine learning model, the item names in the specific information generated by the general-purpose information generation unit 25c that correspond to any of the item names obtained from all templates stored in the template database 24c.

[0097] In the above explanation, an invoice has been used as an example of a document. However, the present invention is also applicable to documents other than invoices. For example, the present invention is applicable to various forms such as quotations, receipts, and delivery slips. [Explanation of Symbols]

[0098] 20 Information Generation System 24a Information generation program 24c Template Database (Template Management Department) 25a Bill classification section (document classification section) 25b Template-type information generation unit 25c General-purpose information generation unit 25d Template generation unit 50 Workflow Systems 54a Workflow Program 55a Workflow Generation Unit 55b Workflow Execution Unit

Claims

1. A template generation unit generates a template for generating specific information that includes information extracted from the image of a document, If the template management unit that manages the aforementioned template has the aforementioned template stored in a provisional registration state, the registration status change unit changes the status of the template to a fully registered state in accordance with instructions from the user. Equipped with, The template generation unit stores the template it has generated in the template management unit in a provisional registration state. The information generation system is characterized in that, when the template generation unit has the template stored in the template management unit in a provisional registration state, it notifies a specific contact person of an instruction prompting them to change the status of the template to a fully registered state.

2. The information generation system according to claim 1, characterized in that the template generation unit notifies the specific contact person of an instruction prompting them to change the status of the template to a fully registered status after the template has been stored in the template management unit in a provisional registration state.

3. The information generation system according to claim 1, characterized in that when the user confirms the template via the display unit of the user terminal, the registration status change unit changes the status of the template to the registered status in response to an operation via the operation unit of the user terminal.

4. The system includes an information generation unit that generates the specific information from the image without using the aforementioned template, The information generation system according to claim 1, characterized in that the template generation unit generates a new template, which is a new template, based on the image, the specific information generated from the image by the information generation unit, and the templates managed by the template management unit.

5. The information generation system according to claim 4, characterized in that the information generation unit generates the specific information from the image using an external service of the information generation system.

6. The aforementioned specific information is such that each item has an item name and a value associated with it. The template includes, for each item of the information, the item name of the information and the location of the region where the information exists. The information generation system according to claim 4, characterized in that the template generation unit generates a new template by including, for each item, the item name in the specific information generated from the image by the information generation unit, which has been converted to the item name included in the template managed by the template management unit.

7. The information generation system according to claim 6, characterized in that the template generation unit generates a new template by converting the name of the item in the specific information generated from the image by the information generation unit with respect to a first item which is one of the items into the name of the item included in the first template which is one of the templates managed by the template management unit, and by converting the name of the item in the specific information generated from the image by the information generation unit with respect to a second item which is one of the items but different from the first item into the name of the item included in the second template which is one of the templates managed by the template management unit but different from the first template, with respect to a second item which is one of the items but different from the first item.

8. The information generation system according to claim 6, characterized in that the template generation unit generates a new template by including, for each item, the item name in the specific information generated from the image by the information generation unit, which has been converted to the item name included in the template managed by the template management unit based on a dictionary.

9. The template includes anchors for each item of the information that serve as reference points for the position from which the information is extracted from the image. The location of the region included in the template is the location relative to the anchor. The information generation system according to claim 8, characterized in that the template generation unit uses the string indicated in the dictionary as the anchor when the string exists in the image for the item name.

10. The information generation system according to claim 6, characterized in that the template generation unit generates a new template by including, for each item, the item name in the specific information generated from the image by the information generation unit, which has been converted using a learning model to the item name included in the template managed by the template management unit.

11. A workflow generation unit generates a workflow representing a series of processes by linking together any of several blocks, each representing a specific process. A workflow execution unit that executes the processes shown in the workflow generated by the workflow generation unit. Equipped with, The workflow generation unit, A template generation unit generates a template for generating specific information that includes information extracted from the image of a document, If the template management unit that manages the aforementioned template has the aforementioned template stored in a provisional registration state, the registration status change unit changes the status of the template to a fully registered state in accordance with instructions from the user. It is possible to generate the aforementioned workflow that realizes this, The template generation unit stores the template it has generated in the template management unit in a provisional registration state. The workflow system is characterized in that, when the template generation unit has the template stored in the template management unit in a provisional registration state, it notifies a specific contact person of an instruction prompting them to change the status of the template to a fully registered state.

12. A template generation unit generates a template for generating specific information that includes information extracted from the image of a document, If the template management unit that manages the aforementioned template has the aforementioned template stored in a provisional registration state, the registration status change unit changes the status of the template to a fully registered state in accordance with instructions from the user. To make this a reality on a computer, The template generation unit stores the template it has generated in the template management unit in a provisional registration state. The template generation unit is characterized by notifying a specific contact person of an instruction prompting them to change the status of the template to a registered status when the template management unit has the template stored in a provisional registration state.

13. A workflow generation unit generates a workflow representing a series of processes by linking together any of several blocks, each representing a specific process. A workflow execution unit that executes the processes shown in the workflow generated by the workflow generation unit. To make this a reality on a computer, The workflow generation unit, A template generation unit generates a template for generating specific information that includes information extracted from the image of a document, If the template management unit that manages the aforementioned template has the aforementioned template stored in a provisional registration state, the registration status change unit changes the status of the template to a fully registered state in accordance with instructions from the user. It is possible to generate the aforementioned workflow that realizes this, The template generation unit stores the template it has generated in the template management unit in a provisional registration state. The workflow program is characterized in that, when the template generation unit has the template stored in the template management unit in a provisional registration state, it notifies a specific contact person of an instruction prompting them to change the status of the template to a fully registered state.

Images