JP7421866B2 - Learning information manufacturing device, information processing device, learning information manufacturing method, information processing method, and program - Google Patents

Description

The present invention relates to an information processing apparatus and the like that process information regarding a bill using a bill image.

Traditionally, it was set up between a computer system server and a client, and was used to read documents such as invoices as images, and perform a variety of processes that required not only accounting processing but also extraction, classification, aggregation, etc., without requiring manpower. There is a technology for a business processing system that makes it possible to perform tasks without having to do so (see Patent Document 1).

Such a business processing system is set up in a server and a client of a computer system, and each is equipped with an image reading means and a memory. This system assigns a continuous business code number to each document that is generated sequentially in the process of business processing, and uses an item code storage means and standard item terminology code to perform each process on the data read as image data by an image reading means. Each of the data storage means, random code data storage means, and extraction, classification, or aggregation processing means is provided.

Japanese Patent Application Publication No. 2014-228914

However, in the conventional technology, OCR software is simply used to obtain character strings from invoice images (hereinafter referred to as invoice images). It was not possible to obtain and store information easily and with high precision. Note that the information to be managed in the bill image includes, for example, company name, company address, account information, and the like.

The learning information production device of the first invention includes a bill image receiving section that receives a bill image, a bill image output section that outputs the bill image received by the bill image receiving section, and a bill image output section. For the bill image output by a billing-related information acquisition unit that acquires billing-related information that is a character string written in each of two or more areas; and two or more sets that are a pair of area specifying information and billing-related information corresponding to the area specifying information. This is a learning information manufacturing device that includes a learning information configuration section that configures learning information using information, and a learning information storage section that accumulates learning information.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily acquired.

Further, the learning information manufacturing device of the second invention is different from the first invention in that the billing-related information acquisition unit receives manually input billing-related information.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily and accurately acquired.

Further, in the learning information production device of the third invention, in contrast to the first invention, the billing-related information acquisition unit performs character recognition processing on the bill image received by the bill image receiving unit, and Using a character recognition means for acquiring one billing-related information, a billing-related information receiving means for receiving manually inputted second billing-related information, and the first billing-related information and the second billing-related information, a determining means for determining the billing-related information to be accumulated, and the learning information configuration unit configures the learning information using one or more sets of the area specifying information and the billing-related information determined by the determining means. This is a learning information production device.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily and accurately acquired.

Further, in the learning information manufacturing device of the fourth invention, in contrast to the first invention, the billing-related information acquisition unit receives two or more pieces of second billing-related information manually input by each of the two or more users. The learning information configuration unit includes a billing-related information receiving unit and a determining unit that uses two or more pieces of second billing-related information to determine the billing-related information to be accumulated. This is a learning information manufacturing device that configures learning information using one or more sets of determined billing-related information.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily and extremely accurately acquired.

Further, in the learning information production device of the fifth invention, in contrast to the fourth invention, the learning information configuration unit includes two or more pieces of information that are a set of area specifying information and billing related information corresponding to the area specifying information. The learning information manufacturing device includes billing source information configuring means for configuring billing source information having set information, and the learning information includes billing source information.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily acquired.

Further, in the learning information production device of the sixth invention, in contrast to any one of the first to fifth inventions, the learning information configuration unit is configured to combine area specifying information and billing related information corresponding to the area specifying information. The learning information manufacturing apparatus includes a learning means for generating a learning device by a machine learning algorithm using two or more sets of set information, and the learning information includes the learning device.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily acquired.

Furthermore, in the learning information production device of the seventh invention, for any one of the first to sixth inventions, two or more types of billing-related information include a billing organization name including the billing organization name. The learning information production device includes source organization information, one or more pieces of payee information regarding a payee, billing amount information regarding a billed amount, and payment due date information regarding a payment due date.

With this configuration, information to be managed can be easily and accurately acquired from the bill image, and learning information that can be stored can be easily acquired.

Further, the information processing device of the eighth invention includes a learning information storage section in which learning information accumulated by the learning information production device is stored, a bill image receiving section that receives a bill image, and a bill image receiving section. a bill information acquisition unit that acquires bill information having two or more types of bill-related information using reception information that is information about received bill images and learning information in a learning information storage unit; The information processing device includes a bill information output unit that outputs a bill information output unit.

With this configuration, information to be managed can be easily and accurately acquired from bill images and stored.

Further, in the information processing device of the ninth invention, in contrast to the eighth invention, the learning information includes invoice source information, is information regarding the invoice image received by the invoice image reception unit, and has an account number. The invoice information acquisition section searches for billing source information using the reception information received by the reception information receiving section as a key, and searches for billing source information corresponding to the received information. This is an information processing device that acquires bill information, which is all or part of the information.

With this configuration, information to be managed can be easily and accurately acquired and stored from bill images using learning information.

Furthermore, in the information processing device of the tenth invention, in contrast to the ninth invention, the bill information acquisition section searches for bill source information using the reception information received by the reception information reception section as a key, and searches for the reception information. a determination means for determining whether or not invoice source information corresponding to the reception information exists; A bill image output unit that outputs the bill image received by the bill image reception unit, and a bill image output unit that outputs the bill image received by the bill image reception unit, and a determination unit that outputs bill source information corresponding to the reception information. If it is determined that the invoice does not exist, the invoice image output unit outputs area identification information that specifies areas of two or more types of billing-related information related to the billing information for each billing-related information. an area specifying information reception unit that receives the area specifying information; a billing related information acquisition unit that obtains the billing related information that is a character string written in each of two or more areas specified by the area specifying information; The billing information acquisition unit further includes a learning information configuration unit that configures learning information using two or more sets of information that are paired with corresponding billing-related information, and a learning information storage unit that accumulates learning information. is an information processing device that acquires bill information having two or more types of bill-related information acquired by a bill-related information acquisition unit.

With this configuration, information to be managed from bill images can be easily and accurately acquired and stored using learning information, and learning information can be evolved.

Further, in the information processing device of the eleventh invention, in contrast to the eighth invention, the learning information includes a learning device, the reception information is the bill image itself received by the bill image reception unit, The bill information acquisition unit is an information processing device that applies the bill image received by the bill image reception unit to a learning device using a machine learning algorithm, and acquires bill information having two or more types of bill-related information. It is.

With this configuration, information to be managed can be easily and accurately acquired and stored from bill images using learning information.

Further, the information processing device of the twelfth invention, for any one of the eighth to eleventh inventions, further includes a modification receiving section that receives a modification to the bill information outputted by the bill information output section; The information processing apparatus includes a bill information storage section that stores modified bill information based on the modifications received by the reception section.

With this configuration, information to be managed can be easily and accurately acquired and stored from bill images using learning information.

According to the learning information production device according to the present invention, it is possible to easily and accurately acquire and store information to be managed from bill images.

Conceptual diagram of information system A in Embodiment 1 Block diagram of the information system A A block diagram of the learning information production device 1 that constitutes the information system A First, a flowchart explaining an example of the operation of the learning information production device 1. Flowchart explaining an example of same group information etc. acquisition process Also, a flowchart explaining an example of learning information configuration processing. Next, a flowchart explaining an example of the operation of the second terminal 3 Diagram showing an example of the same invoice image Diagram showing an example of the same cutting instruction screen Diagram showing an example of the cutout location of the same area Diagram showing an example of the same screen Diagram showing an example of the same screen Diagram showing an example of the same screen Diagram showing the billing source information management table Block diagram of information system B in embodiment 2 A block diagram of the information processing device 4 that constitutes the information system B A flowchart explaining an example of the operation of the information processing device 4 Diagram showing an example of the same screen Diagram showing an example of the same screen Diagram showing an example of the same input screen An overview diagram of the computer system in the above embodiment Block diagram of the computer system

Hereinafter, embodiments of a learning information production device, an information processing device, etc. will be described with reference to the drawings. Note that in the embodiments, constituent elements with the same reference numerals perform similar operations, and therefore, repeated explanation may be omitted.

(Embodiment 1)

In this embodiment, an information system including a learning information production device that acquires and accumulates learning information used to efficiently acquire billing-related information from received bill images will be described. More specifically, in this embodiment, information indicating areas of billing-related information in an invoice image is accepted, and area identification information that specifies each area is associated with information on character strings in each area. An information system including a learning information production device that acquires and accumulates learning information using two or more sets of information will be described.

Note that in this embodiment, the billing-related information, which is a character string within the area, may be received manually from one or more users, may be obtained by character recognition technology, or may be received by one or more users by manual input, or may be obtained by character recognition technology. The information may be acquired by a combination of manual input from the user and acquisition using character recognition technology. In addition, when receiving manual input from two or more users, or when using both manual input from users and acquisition using character recognition technology, it is preferable to determine the billing-related information to be used through appropriate processing. .

Further, in this embodiment, the learning information is information including one or more of billing source information, which will be described later, and a learning device, which will be described later.

FIG. 1 is a conceptual diagram of an information system A in this embodiment. The information system A includes a learning information production device 1, one or more first terminals 2, and one or more second terminals 3. The learning information production device 1 is, for example, a so-called server. The server may be, for example, a so-called cloud server, an ASP server, etc., and its type, installation location, etc. do not matter. The first terminal 2 and the second terminal 3 are, for example, a so-called personal computer, a smartphone, a tablet terminal, etc., and their types do not matter.

FIG. 2 is a block diagram of information system A in this embodiment. FIG. 3 is a block diagram of the learning information production device 1 that constitutes the information system A.

The learning information manufacturing device 1 includes a storage section 11, a reception section 12, a processing section 13, and an output section 14. The storage unit 11 includes a learning information storage unit 111. The reception unit 12 includes, for example, a bill image reception unit 112 and an area identification information reception unit 113. The processing unit 13 includes, for example, a billing-related information acquisition unit 131, a learning information configuration unit 132, and a learning information storage unit 133. The billing-related information acquisition unit 131 includes, for example, a character recognition unit 1311, a billing-related information receiving unit 1312, and a determining unit 1313. The learning information configuration unit 132 includes, for example, a billing source information configuration unit 1321 and a learning unit 1322. The output unit 14 includes, for example, a bill image output unit 141.

The first terminal 2 includes a first storage section 21 , a first reception section 22 , a first processing section 23 , a first transmission section 24 , a first reception section 25 , and a first output section 26 .

The second terminal 3 includes a second storage section 31 , a second reception section 32 , a second processing section 33 , a second transmission section 34 , a second reception section 35 , and a second output section 36 .

A storage unit 11 that constitutes the learning information manufacturing device 1 stores various types of information. The various types of information are, for example, learning information described later.

The learning information storage unit 111 stores learning information. The learning information is, for example, one or more types of information of billing source information, which will be described later, and a learning device, which will be described later.

The reception unit 12 receives information, instructions, and the like. The information, instructions, etc. are, for example, a bill image, which will be described later, and area specifying information, which will be described later. The reception unit 12 normally receives information, instructions, etc. from the first terminal 2 or the second terminal 3. However, the reception unit 12 may also accept information, instructions, etc. from the user. In such a case, any means for inputting information, instructions, etc. may be used, such as a touch panel, keyboard, mouse, or menu screen. In such a case, the reception unit 12 may be realized by a device driver for input means such as a touch panel or a keyboard, control software for a menu screen, or the like.

The bill image receiving unit 112 receives bill images. The bill image is image data of a bill. The bill image may be, for example, a PDF file, a GIF file, or a JPEG file, but its format and data structure are not limited. The reception here usually means reception from the first terminal 2. Note that the first terminal 2 is a terminal used by a first user who desires to manage bills. Further, the bill image may be accepted by accepting information input from an input device such as a keyboard, mouse, or touch panel, or by accepting information read from a recording medium such as an optical disk, a magnetic disk, or a semiconductor memory. Further, the bill image receiving unit 112 may perform a conversion process on the received bill image and obtain the converted bill image. It may be said that the bill image after such conversion is also the bill image accepted by the bill image receiving unit 112. For example, the bill image receiving unit 112 may convert the received PDF bill image into a bill image in a format such as jpeg or gif.

The region specifying information receiving unit 113 receives one or more pieces of region specifying information. It is suitable for the bill image output by the bill image output unit 141 to receive area specifying information that specifies each area of two or more types of bill related information for each bill related information.

It is preferable that the area specifying information received by the area specifying information receiving unit 113 be associated with a type identifier. The area specifying information receiving unit 113 acquires, for example, one or more pairs of area specifying information and type identifier. The type identifier is information that specifies the type of billing-related information. The type identifier is, for example, "company name", "address", "billing amount", "payee", and "payment date".

Here, the area specifying information is information that specifies the area of a portion within the bill image. The area identification information includes, for example, the upper left coordinate value and lower right coordinate value of the area in the invoice image, the upper left coordinate value, width and height of the area in the invoice image, and the rectangular area in the invoice image. such as the center of gravity. The data structure of the area identification information does not matter.

Moreover, billing related information is information related to billing. The billing-related information is information displayed in the bill image. Billing-related information includes, for example, organization name information that specifies the name of the billing organization, address information that specifies the address of the organization, bill amount information that specifies the bill amount, payee information that specifies the payment destination, and payment date. This is one or more types of payment date information. Preferably, the two or more types of billing-related information are organization name information, address information, billing amount information, payee information, and payment date information. In addition, it is preferable that the two or more types of billing-related information include billing organization information regarding the billing organization including the billing organization name, one or more payment destination information, billing amount information, and payment date information. be. The billing organization information includes, for example, organization name information and address information. Note that the payee information is, for example, information on the account of the transferee. The payee information includes, for example, bank name, branch name, account type, account number, and account name. The payee information may be, for example, a credit card number, name, and expiration date.

The billing-related information is, for example, fixed billing-related information that is fixed information, or variable billing-related information that is variable information. Fixed billing-related information is information that does not change dynamically, and is, for example, billing-related information paired with a type identifier "company name" or "address." The variable billing-related information is usually information that differs depending on the bill image, and is, for example, billing-related information paired with the type identifiers "billing amount," "payee," and "payment date."

The reception by the area specifying information reception unit 113 usually means reception from the second terminal 3. Note that the second terminal 3 is a terminal of a second user of an organization that manages the learning information production device 1. The second user is, for example, a user who provides a bill management service.

The area specifying information receiving unit 113 may accept information input from an input device such as a keyboard, mouse, or touch panel, or may accept information read from a recording medium such as an optical disk, a magnetic disk, or a semiconductor memory.

The processing unit 13 performs various processes. The various processes are, for example, processes performed by the billing-related information acquisition unit 131, the learning information configuration unit 132, the learning information storage unit 133, and the like. The various processes are, for example, processes that configure a cutting instruction screen. The cutting instruction screen is a screen that includes a receipt image. The cutting instruction screen is a screen for instructing the user to input two or more pieces of area specifying information for the receipt image. The cut instruction screen may be said to be a screen for cutting out two or more pieces of billing-related information from the receipt image. Further, the cut-out instruction screen may be a screen for specifying two or more areas of billing-related information on the receipt image.

The billing-related information acquisition unit 131 acquires billing-related information that is a character string written in each of two or more areas specified by the area specifying information. Here, acquisition means acquisition of information manually input by the user or automatic acquisition. The user here is the second user of the second terminal 3.

Obtaining billing-related information usually means acquiring one or more pairs of billing-related information and type identifier.

The billing-related information acquisition unit 131 receives, for example, one or more pieces of manually input billing-related information. The reception here is, for example, reception from the second terminal 3. However, the acceptance may be, for example, acceptance of manual input by the user.

For example, the billing-related information acquisition unit 131 performs character recognition processing on the bill image to acquire one or more pieces of billing-related information. Note that character recognition processing is performed by character recognition means 1311, which will be described later.

For example, the billing-related information acquisition unit 131 performs character recognition processing on each area of one or more area specifying information, and acquires one or more first billing-related information. Note that the character recognition process is a well-known technique, so a detailed explanation will be omitted. Further, the billing-related information acquisition unit 131 usually acquires first billing-related information for each area specifying information. That is, the billing-related information acquisition unit 131 usually acquires two or more pieces of first billing-related information. Note that the first billing-related information is billing-related information automatically acquired through character recognition processing. The billing-related information input by the user is appropriately referred to as second billing-related information to distinguish it from the first billing-related information.

It is preferable that the billing-related information acquisition unit 131 acquires one or more pieces of first billing-related information in association with the type identifier.

The character recognition unit 1311 performs character recognition processing on the bill image received by the bill image receiving unit 112, and acquires first bill-related information.

It is preferable that the character recognition unit 1311 performs character recognition processing on each area of one or more area specifying information to obtain one or more pieces of first billing-related information.

Billing related information receiving means 1312 receives manually inputted second billing related information. The manually inputted second billing-related information may be billing-related information that is a modified version of the first billing-related information. It is preferable that the billing-related information receiving means 1312 receives one or more pieces of first billing-related information associated with a type identifier. Here, manual input is usually input by the user of the second terminal 3. Moreover, the reception here usually means reception from the second terminal 3.

It is preferable that the billing-related information receiving means 1312 receives two or more pieces of second billing-related information manually input by two or more users. The two or more second billing-related information manually input by each of the two or more users is information received from each of the two or more second terminals 3.

The determining means 1313 determines billing related information to be accumulated.

The determining means 1313 determines the billing-related information to be accumulated, for example, using the first billing-related information and the second billing-related information. The determining means 1313, for example, compares the first billing-related information and the second billing-related information, and determines one of the pieces of information as the billing-related information to be accumulated.

For example, if the first billing-related information and the second billing-related information match regarding a predetermined type of billing-related information, the determining means 1313 determines the billing-related information to be stored. Here, the predetermined type of billing-related information is, for example, one or more types of information among organization name information, address information, and payee information.

For example, when the first billing-related information and one or more second billing-related information are the same, the determining means 1313 selects the first billing-related information or the second billing-related information from the accumulated billing-related information. Obtain as information. If the first billing-related information and the second billing-related information are different, the determining means 1313 outputs confirmation information for confirmation to the user, for example. The confirmation information is, for example, "Please confirm whether the input information is correct." Here, output usually means transmission to the second terminal 3.

The determining means 1313 determines the billing-related information to be accumulated, for example, using two or more pieces of second billing-related information. For example, if two or more pieces of second billing-related information match for a predetermined type of billing-related information, the determining means 1313 determines the second billing-related information as the billing-related information to be accumulated. For example, the determining unit 1313 determines the second billing-related information with a high matching ratio among the three or more pieces of second billing-related information as the billing-related information to be accumulated. For example, if the degree of matching is low enough to match a predetermined condition, confirmation information for confirmation to the user is output. Note that the predetermined type of billing-related information is, for example, one or more types of information among organization name information, address information, and payee information. If all the two or more pieces of second billing related information are not the same, the determining means 1313 outputs confirmation information for confirmation to the user, for example. Here, output usually means transmission to the second terminal 3.

Note that the determination of the billing-related information may be the acquisition of the billing-related information, the acquisition of a link to the billing-related information, the acquisition of the ID of the billing-related information, or the like.

The learning information configuration unit 132 configures learning information using one or more set information. The set information is information on a set of area specifying information and billing related information corresponding to the area specifying information. The billing-related information corresponding to the area specifying information is billing-related information that is a character string written in the area specified by the area specifying information. The set information typically also includes a type identifier. The learning information is information used to support input of bill information, which will be described later. The learning information is, for example, one or more types of information of billing source information, which will be described later, and a learning device, which will be described later.

The learning information configuration unit 132 typically configures learning information using two or more sets of information.

The learning information configuration unit 132 configures learning information using one or more sets of information, which are information on sets of area specifying information and billing-related information determined by the determining means 1313.

The bill source information configuration means 1321 configures bill source information having two or more sets of information, each of which is a pair of area specifying information and billing related information corresponding to the area specifying information. It is preferable that the billing source information also includes a billing image, for example. The bill source information may include, for example, a partial image corresponding to each bill related information. It goes without saying that the data structure of group information, billing source information, etc. does not matter.

The learning unit 1322 generates a learning device by a machine learning algorithm using two or more sets of information, which are a set of area specifying information and billing related information corresponding to the area specifying information. Note that the learning device may also be called a classifier.

The learning unit 1322 typically generates a learning device using a machine learning algorithm using the bill image and two or more sets of information.

It goes without saying that the algorithm used in machine learning does not matter. For machine learning, for example, deep learning, SVM, SVR, random forest, decision tree, etc. can be used. In addition, in machine learning, in order to configure a learning device, for example, input information groups and information to be output are passed to machine learning modules (e.g. fasttext, tiny_svm, various randomForest functions, etc.) as arguments. If you give it, you will get a learning device. Furthermore, when making predictions in machine learning, predicted information can be obtained by giving a learning device and a group of input information as arguments to a machine learning module. Note that the predicted information may be said to be information as a result of classification.

Note that the input information group is, for example, "bill image, two or more area specifying information," or "two or more partial images, two or more area specifying information," or "bill image, type. ``Two or more sets of an identifier and area specifying information,'' or ``Two or more sets of partial images, type identifier, and area specifying information.'' The partial image is an image of a part of the bill image, and is an image of the area specified by the area specifying information. Further, the information to be outputted is "billing related information" or "type identifier and billing related information".

The learning information storage unit 133 stores learning information. The learning information storage unit 133 stores the learning information configured by the learning information configuration unit 132. The learning information is stored in, for example, the learning information storage unit 111, but may be an external device.

The output unit 14 outputs various information. The various types of information include, for example, a bill image and a cutting instruction screen. Moreover, the output here usually means transmission to the second terminal 3. However, output refers to, for example, displaying on a display, projecting using a projector, printing on a printer, outputting sound, storing on a recording medium, passing processing results to other processing devices or programs, etc. It may also be a concept that includes

The bill image output unit 141 outputs the bill image received by the bill image reception unit 112. The output here usually means transmission to the second terminal 3.

The first storage unit 21 that constitutes the first terminal 2 stores various types of information. The various types of information include, for example, a first user identifier that identifies the first user, a first terminal identifier that identifies the first terminal 2, and the like. Note that the first user identifier may be the same as the first terminal identifier.

The first reception unit 22 receives various information, instructions, and the like. The various information and instructions include, for example, a bill image and confirmation information.

The first processing unit 23 performs various processes. The various processes include, for example, converting the information received by the first reception unit 22 into a data structure for transmission, and converting the information received by the first reception unit 25 into a data structure to display it.

The first transmitter 24 transmits various information, instructions, etc. to an external device. The various types of information include, for example, bill images. The external devices are, for example, the learning information production device 1 and the information processing device 4 described below.

The first receiving unit 25 receives various types of information from external devices. The external devices are, for example, the learning information production device 1 and the information processing device 4 described below.

The first output unit 26 outputs information received by the first reception unit 25, information acquired by the first processing unit 23, information received by the first reception unit 22, and the like. Here, output refers to display on a display, projection using a projector, printing on a printer, sound output, transmission to an external device, storage on a recording medium, processing to other processing devices or other programs, etc. This is a concept that includes the delivery of results.

The second storage unit 31 that constitutes the second terminal 3 stores various types of information. The various types of information include, for example, a second user identifier that identifies the second user, a second terminal identifier that identifies the second terminal 3, and the like. Note that the second user identifier may be the same as the second terminal identifier.

The second reception unit 32 receives various information, instructions, and the like. The various information and instructions include, for example, area specifying information and billing-related information. Any means for inputting various information and instructions may be used, such as a touch panel, keyboard, mouse, or menu screen. The second reception unit 32 may be realized by a device driver for input means such as a touch panel or a keyboard, control software for a menu screen, or the like.

The second processing unit 33 performs various processes. The various processes include, for example, converting the information received by the second reception unit 32 into a data structure for transmission, and converting the information received by the second reception unit 35 into a data structure for display.

The second transmitter 34 transmits various information, instructions, etc. to an external device. The various types of information include, for example, area specifying information and billing-related information. The external devices are, for example, the learning information production device 1 and the information processing device 4 described below.

The second receiving unit 35 receives various information from external devices. The various types of information include, for example, a cutting instruction screen and confirmation information. The external devices are, for example, the learning information production device 1 and the information processing device 4 described below.

The second output unit 36 outputs information received by the second reception unit 35, information acquired by the second processing unit 33, information received by the second reception unit 32, and the like. Here, output refers to display on a display, projection using a projector, printing on a printer, sound output, transmission to an external device, storage on a recording medium, processing to other processing devices or other programs, etc. This is a concept that includes the delivery of results.

The storage section 11, the learning information storage section 111, the first storage section 21, and the second storage section 31 are preferably non-volatile recording media, but can also be realized using volatile recording media.

The process by which information is stored in the storage unit 11 or the like does not matter. For example, information may be stored in the storage unit 11 or the like via a recording medium, or information transmitted via a communication line or the like may be stored in the storage unit 11 or the like. Alternatively, information input via an input device may be stored in the storage unit 11 or the like.

The reception unit 12, the bill image reception unit 112, the area identification information reception unit 113, the first reception unit 25, and the second reception unit 35 are usually realized by wireless or wired communication means, but they do not receive broadcasts. It may be realized by means.

Processing unit 13, billing related information acquisition unit 131, learning information configuring unit 132, learning information storage unit 133, character recognition unit 1311, billing related information receiving unit 1312, determining unit 1313, bill source information configuring unit 1321, learning unit 1322 , the first processing section 23, and the second processing section 33 can usually be realized by an MPU, a memory, or the like. The processing procedures of the processing unit 13 and the like are usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

The output unit 14, the bill image output unit 141, the first transmitting unit 24, and the second transmitting unit 34 are usually realized by wireless or wired communication means, but may also be realized by broadcasting means.

The first reception unit 22 and the second reception unit 32 may be realized by a device driver for input means such as a touch panel or a keyboard, control software for a menu screen, or the like.

The first output section 26 and the second output section 36 may or may not include output devices such as a display and a speaker. The first output unit 26 and the second output unit 36 may be realized by output device driver software, output device driver software and the output device, or the like.

Next, an example of the operation of the information system A will be explained. First, an example of the operation of the learning information manufacturing device 1 will be described using the flowchart of FIG. 4.

(Step S401) The bill image reception unit 112 determines whether or not a bill image has been received. If the bill image is received, the process goes to step S402, and if the bill image is not received, the process goes to step S410.

(Step S402) The processing unit 13 stores the bill image received in step S401 in the storage unit 11. The processing unit 13 stores, in the storage unit 11, the bill image received in step S401, for example, in pairs with the first user identifier received in pairs with the bill image.

(Step S403) The processing unit 13 configures a cutting instruction screen including the bill image and the like received in step S401.

(Step S404) The bill image output unit 141 transmits the cutting instruction screen configured in step S403 to one or more second terminals 3.

(Step S405) The area specifying information reception unit 113 determines whether area specifying information and the like have been received from one or more second terminals 3. If area specifying information, etc. is received from all the one or more second terminals 3 that sent the cutting instruction screen, the process goes to step S406, and area specifying information, etc. is received from all the one or more second terminals 3 that sent the cutting instruction screen. If not received, the process returns to step S405. Note that the received area specifying information and the like is, for example, a set of information including a pair of area specifying information and a type identifier. Further, the received area specifying information and the like are, for example, a plurality of sets of information including area specifying information, type identifier, and billing related information.

(Step S406) The billing related information acquisition unit 131 performs group information etc. acquisition processing. The group information etc. acquisition process is a process of acquiring a set of group information to be accumulated. The group information etc. acquisition process will be explained using the flowchart of FIG. Note that the set information is information on a set of area specifying information and billing-related information.

(Step S407) The billing-related information acquisition unit 131 stores the set of group information acquired in step S406 in the storage unit 11.

(Step S408) The processing unit 13 uses the set of group information acquired in step S406 to acquire bill information to be accumulated. Bill information includes two or more types of bill-related information. The bill information is, for example, a set of two or more pieces of bill-related information associated with a type identifier.

(Step S409) The processing unit 13 stores the bill information acquired in step S408 in pairs with the first user identifier accepted in pairs with the bill image and/or with the bill image received in step S401. It is stored in section 11. Return to step S401.

(Step S410) The learning information configuration unit 132 determines whether to perform processing for configuring learning information. If the process of forming the learning information is to be performed, the process goes to step S411, and if the process of forming the learning information is not to be performed, the process returns to step S401. Note that, for example, when the learning information configuration unit 132 determines that the number of accumulated sets of set information in step S407 is so large that it satisfies a predetermined condition (for example, the number of accumulated sets of set information is greater than or equal to a threshold value), When a predetermined time has arrived, when an instruction from a user has been received, etc., it is determined that processing for configuring learning information is to be performed.

(Step S411) The learning information configuration unit 132 performs a process of configuring learning information using the set of set information stored in the storage unit 11. Return to step S401. Note that the learning information configuration process will be explained using the flowchart of FIG. 6.

Note that in the flowchart of FIG. 4, if a learning device is not configured, steps S410 and S411 do not exist.

Further, in the flowchart of FIG. 4, the process is terminated by turning off the power or by an interrupt to terminate the process.

Next, an example of the group information etc. acquisition process in step S406 will be explained using the flowchart of FIG.

(Step S501) The billing-related information acquisition unit 131 assigns 1 to the counter i.

(Step S502) The billing-related information acquisition unit 131 determines whether or not the i-th type identifier or the i-th area identification information exists in the area identification information etc. received by the area identification information reception unit 113. . If the i-th type identifier etc. exists, the process goes to step S503, and if it does not exist, the process goes to step S509.

(Step S503) The billing-related information acquisition unit 131 determines whether or not to perform character recognition processing on the area specified by the i-th area specifying information. If character recognition processing is to be performed, the process proceeds to step S504, and if character recognition processing is not to be performed, the process proceeds to step S506. Note that, for example, the billing-related information acquisition unit 131 determines to perform the character recognition process when the i-th type identifier is a predetermined type identifier (for example, a type identifier corresponding to variable billing-related information). , it is determined that the character recognition process is not performed when the i-th type identifier is another type identifier (for example, a type identifier corresponding to fixed billing related information).

(Step S504) The character recognition means 1311 acquires the i-th area specifying information.

(Step S505) The character recognition unit 1311 performs character recognition processing on the partial image of the area specified by the i-th area specifying information, and obtains a character string.

(Step S506) The determining unit 1313 determines billing-related information to be adopted from one or more billing-related information candidates. Note that the one or more bill-related information candidates are, for example, a bill paired with the one or more i-th area specific information or the i-th type identifier received by the area specifying information reception unit 113 in step S405. This is related information or one or more character strings among the character strings acquired in step S505. Furthermore, since the method for determining the billing-related information adopted by the determining means 1313 has been described above, the description thereof will be omitted here.

(Step S507) The billing related information acquisition unit 131 acquires the i-th group information. The i-th group information is information that associates the i-th area specifying information with the i-th billing related information. It is preferable that the i-th group information is information that associates the i-th type identifier, the i-th area specifying information, and the i-th billing-related information. The i-th set of information is, for example, information including the i-th type identifier, the i-th area specifying information, and the i-th billing-related information. The i-th group information is, for example, information including the i-th type identifier, the i-th area specifying information, the i-th billing-related information, and the i-th partial image.

Here, the information that associates a plurality of pieces of information may be information that includes a plurality of pieces of information, or information that includes link information for each piece of information for acquiring a plurality of pieces of information. The information that associates multiple pieces of information may be information for acquiring multiple pieces of information, and its data structure does not matter.

(Step S508) The billing-related information acquisition unit 131 increments the counter i by 1. Return to step S502.

(Step S509) The billing-related information acquisition unit 131 uses the two or more sets of information acquired in step S507 to configure set information and the like to be accumulated, and stores it in the storage unit 11. Return to upper level processing. Note that the set information etc. to be accumulated are usually a bill image and two or more set information. However, the group information etc. to be accumulated may be, for example, two or more group information without a bill image.

Next, an example of the learning information configuration process in step S411 will be explained using the flowchart of FIG. 6.

(Step S601) The learning information configuration unit 132 assigns 1 to the counter i.

(Step S602) The learning information configuration unit 132 determines whether the i-th bill image to be learned exists in the storage unit 11. If the i-th bill image exists, the process goes to step S603; if the i-th bill image does not exist, the process goes to step S612. Note that if no bill images are stored in the storage unit 11, it is determined whether or not there is a set of i-th set of information corresponding to the i-th bill image.

(Step S603) The invoice source information configuration means 1321 acquires the i-th set of information into the storage unit 11. The i-th set of information is two or more sets of information, and is information corresponding to the i-th bill image to be learned.

(Step S604) The learning information configuration unit 132 assigns 1 to the counter j.

(Step S605) The learning information configuration unit 132 determines whether the j-th group information exists in the i-th set of information acquired in step S603. If the j-th group information exists, the process goes to step S606, and if the j-th group information does not exist, the process goes to step S608.

(Step S606) The bill source information configuration means 1321 acquires the j-th partial image corresponding to the j-th set of information. The learning information configuration unit 132 acquires, for example, a partial image included in the j-th set of information. Further, the learning information configuration unit 132, for example, cuts out the image of the area specified by the area specifying information included in the j-th set of information from the bill image paired with the j-th set of information to obtain a partial image. .

(Step S607) The learning information configuration unit 132 increments the counter j by 1. Return to step S605.

(Step S608) The billing source information configuration unit 1321 configures a vector to be passed to the machine learning function using the group information acquired in step S603, or the group information acquired in step S603 and the partial image acquired in step S606. do. Note that the vector may also be referred to as a feature amount vector.

(Step S609) The billing source information configuration means 1321 configures billing source information using the vector configured in step S608.

(Step S610) The bill source information configuration means 1321 temporarily stores the bill source information configured in step S609. Note that the temporary storage destination is, for example, the storage unit 11, but is not limited thereto.

(Step S611) The learning information configuration unit 132 increments the counter i by 1. Return to step S602.

(Step S612) The learning means 1322 gives the set of vectors configured in step S608 as an argument to a module that implements a machine learning algorithm, executes the module, and obtains a learning device.

(Step S613) The learning means 1322 stores the learning device acquired in step S612 in the learning information storage unit 111. Return to upper level processing.

In addition, in the flowchart of FIG. 6, one learning device was accumulated. However, in the flowchart of FIG. 6, learning devices may be generated and stored for each type identifier. In such a case, the learning means 1322 provides the set of vectors configured in step S608 for each type identifier as an argument to a module that implements a machine learning algorithm, executes the module, and obtains a learning device. Furthermore, in such a case, the learning unit 1322 may generate a learning device by a machine learning algorithm using, for example, positive example billing source information and negative example billing source information for each type identifier. suitable.

Next, the operation of the first terminal 2 will be explained. The first reception unit 22 of the first terminal 2 receives the bill image. Next, the first processing unit 23 converts the bill image received by the first reception unit 22 into a bill image having a data structure to be transmitted. Next, the first transmitting unit 24 transmits the bill image configured by the first processing unit 23 to the learning information manufacturing device 1. Next, the first receiving unit 25 receives information indicating that the bill image and the like have been registered. Next, the first processing unit 23 creates a data structure that outputs information indicating that the invoice image and the like have been registered. The first output unit 26 outputs information indicating that the bill image etc. acquired by the first processing unit 23 has been registered. Note that the bill image and the like are, for example, a bill image and bill information. Further, the first receiving unit 25 receives an inquiry to confirm whether the bill information is correct, the first output unit 26 outputs the inquiry, the first receiving unit 22 receives the confirmation information, and the first sending unit 25 receives the inquiry to confirm whether the bill information is correct. The unit 24 may transmit the confirmation information to the learning information manufacturing device 1.

Next, an example of the operation of the second terminal 3 will be explained using the flowchart of FIG.

(Step S701) The second receiving unit 35 determines whether or not a cutting instruction screen has been received from the learning information manufacturing device 1. If the cutting instruction screen is received, the process goes to step S702, and if the cutting instruction screen is not received, the process returns to step S701.

(Step S702) The second processing unit 33 constructs a cutting instruction screen to be output from the cutting instruction screen received in step S701. Next, the second output unit 36 outputs a cutting instruction screen.

(Step S703) The second processing unit 33 assigns 1 to the counter i.

(Step S704) The second processing unit 33 determines whether the i-th type identifier exists. If the i-th type identifier exists, the process goes to step S705, and if the i-th type identifier does not exist, the process goes to step S708. Note that the i-th type identifier is the i-th area specifying information for the cutting instruction screen.

(Step S705) The second reception unit 32 determines whether or not the i-th area specifying information, etc. corresponding to the i-th type identifier has been received. If the i-th region specifying information, etc. is received, the process goes to step S706, and if the i-th area specifying information, etc. is not received, the process returns to step S705. The i-th area specifying information, etc. is, for example, the i-th area specifying information and the i-th billing related information, or only the i-th area specifying information. Further, the i-th area specifying information and the like may include billing-related information obtained by the second processing unit 33 performing character recognition processing on the partial image specified by the i-th area specifying information. Further, the i-th area specifying information, etc. may include billing-related information input by the second user. Furthermore, the billing-related information that constitutes the area specifying information and the like may include a mixture of billing-related information obtained through character recognition processing and billing-related information input by the second user. Furthermore, the billing-related information constituting the area specifying information and the like may include billing-related information that has been corrected and accepted by the second user with respect to the character string obtained by performing character recognition processing.

(Step S706) The second processing unit 33 temporarily accumulates the i-th area specifying information etc. accepted in step S705.

(Step S707) The second processing unit 33 increments the counter i by 1. Return to step S704.

(Step S708) The second transmitting unit 34 transmits the set of area specifying information etc. accumulated in step S706 to the learning information manufacturing device 1. Return to step S701.

A specific example of the operation of the information system A in this embodiment will be described below. A conceptual diagram of information system A is shown in FIG.

It is now assumed that the first user inputs the bill image shown in FIG. 8 into the first terminal 2. Then, the first reception unit 22 of the first terminal 2 receives the bill image. Next, the first processing unit 23 converts the bill image received by the first reception unit 22 into a bill image having a data structure to be transmitted. Next, the first transmitting unit 24 transmits the bill image configured by the first processing unit 23 to the learning information manufacturing device 1.

Next, the bill image receiving unit 112 of the learning information manufacturing device 1 receives the bill image shown in FIG. 8 . Then, the processing unit 13 stores the received bill image in the storage unit 11. It is assumed that the processing unit 13 has accumulated the received bill image under a file name with a unique ID "ID892.pdf."

Next, the processing unit 13 configures a cutting instruction screen including the bill image and the like received in step S401.

Next, the bill image output unit 141 transmits the configured cutting instruction screen to one or more second terminals 3 managed in advance.

Next, the second receiving unit 35 of each of the one or more second terminals 3 receives the cutting instruction screen from the learning information production device 1. Next, the second processing unit 33 constructs a cutting instruction screen to be output from the received cutting instruction screen. Next, the second output unit 36 outputs a cutting instruction screen. An example of such a cutting instruction screen is shown in FIG. In FIG. 9, the instruction is to cut out only the "company name".

Note that before the second receiving unit 35 receives the cutting instruction screen shown in FIG. 9, it receives the bill type selection screen shown in FIG. 18, which will be described later, and the second output unit 36 outputs the cutting instruction screen. Also good. Then, after the user of the second terminal 3 selects the type (after instructing one of the plurality of type selection buttons), the cutting instruction screen of FIG. 9 may be output to the second terminal 3. . In addition, after the user of the second terminal 3 selects the type, only if the type is one or more predetermined types (for example, other than "Bill (domestic)" or "Other"), as shown in FIG. A cutting instruction screen may be output to the second terminal 3. In such a case, for example, type information specifying the type is transmitted from the second terminal 3 to the learning information manufacturing device 1, the learning information manufacturing device 1 receives the type information, and displays the cutting instruction screen according to the type information. Transmission control may be performed such that the data is transmitted to two terminals 3 or not. Note that if the cutting instruction screen is not sent to the second terminal 3, for example, the user of the second terminal 3 manually inputs the necessary information to the second terminal 3, and the necessary information is transmitted from the second terminal 3. It may be transmitted to the learning information manufacturing device 1, received by the learning information manufacturing device 1, and stored. The second user uses an input means such as a mouse to specify the "company name" area according to the instructions in FIG. 9 . Here, the cutting instruction screen consists of five screens, and the second output unit 36 displays, in addition to the billing company's "company name," the company's "address," "billing amount," "payee," and "payment date." Suppose that screens with instructions for specifying (cutting) each area of the image are sequentially output. It is assumed that the second user inputs area specifying information that specifies areas corresponding to five type identifiers: "company name," "address," "billing amount," "payee," and "payment date." In other words, as shown in FIG. Assume that you have input it.

Then, the second receiving unit 32 receives area specifying information corresponding to each type of identifier ("company name", "address", "billed amount", "payee", and "payment date").

Next, the second terminal 3 outputs a screen for inputting five pieces of billing-related information: "company name," "address," "billing amount," "payee," and "payment date." Such screens are, for example, the screens shown in FIGS. 11, 12, and 13. Further, such a screen is, for example, a screen transmitted from the learning information production device 1 together with the cutting instruction screen.

Next, the second user inputs "company name", "billed amount", "payment date", etc. on each screen of FIGS. 11, 12, and 13 while looking at the bill image. Here, character strings may be input into fields corresponding to "company name," "billing amount," "payment date," etc. by character recognition processing on the bill image.

Then, the second receiving unit 32 receives billing-related information corresponding to each type of identifier ("company name", "address", "billing amount", "payee", and "payment date"). Note that the second reception unit 32 may also receive type information.

Next, the second processing unit 33 forms a set of sets (5 sets of information) of the type identifier, area specifying information, and billing related information. Note that the second processing unit 33 may configure information that also includes type information.

Then, the second transmitting unit 34 transmits a set of configured type identifiers, area specifying information, and billing related information (a set of area specifying information, etc.) to the learning information manufacturing device 1. Note that the second transmitter 34 may transmit information that also includes type information to the learning information manufacturing device 1.

Next, in the learning information manufacturing device 1, the area specifying information receiving unit 113 receives area specifying information and the like from one or more second terminals 3. Here, the area specifying information, etc. is a set of sets (5 sets of information) of the above-mentioned type identifier, area specifying information, and billing related information. Note that when receiving area specifying information and the like from two or more second terminals 3, the received billing-related information is a candidate for billing-related information to be accumulated.

Next, the determining unit 1313 of the billing-related information acquisition unit 131 determines billing-related information to be adopted from one or more billing-related information candidates for each type identifier. Note that the algorithm example for determining the billing-related information to be adopted has been described above, so detailed description thereof will be omitted here.

Next, the billing-related information acquisition unit 131 configures billing source information using the billing-related information and area specifying information for each type identifier. Note that the billing source information is, for example, "Company name: Sansan Co., Ltd. | Address: Miyamae-cho, Shibuya-ku, Tokyo, 150-0001... | Billing amount: 1 | Payment address: Resona Bank..., Mizuho Bank ..., Sumitomo Mitsui Banking Corporation... | Payment date: October 31, 2017 | Area specific information: (1) (x ₁₁ , y ₁₁ ) (x ₁₂ , y ₁₂ ) (2) (x ₂₁ , y ₂₁ ) (x ₂₂ , y ₂₂ ) (3) (x ₃₁ , y ₃₁ ) (x ₃₂ , y ₃₂ ) (4) (x ₄₁ , y ₄₁ ) (x ₄₂ , y ₄₂ ) (5) (x ₅₁ , y ₅₁ )(x ₅₂ , y ₅₂ )|Bill image: ID892.pdf”.

Next, the billing-related information acquisition unit 131 stores the configured billing source information in the learning information storage unit 111. An example of such accumulated billing source information is the record "ID=892" in FIG. 14. Note that FIG. 14 is a billing source information management table. The billing source information management table is a table for managing billing source information included in learning information, and is stored in the learning information storage section 111. Further, the billing-related information acquisition unit 131 may store the billing source information in association with the received type information. Furthermore, the billing source information may be considered to have type information.

It is assumed that the above processing is repeated and a large number of pieces of billing source information are stored in the billing source information management table.

Then, using the fact that the number of billing sources information that is equal to or greater than the threshold value is accumulated in the billing source information management table as a trigger, the learning information manufacturing device 1 performs a learning process as follows.

In other words, the learning information configuration unit 132 stores a large number of sets of information (here, "(1) company name", "(2) address", "(3) billed amount", "(4) A learning device is generated by a machine learning algorithm (for example, deep learning) using the actual image data of ``(5) payment destination,'' ``(5) payment date,'' ``area specifying information,'' and ``bill image.'' This learning device is a learning device for inputting image data (invoice image) and outputting "(1) company name," "(2) address," and "(4) payee." Note that the learning information configuration unit 132 stores information such as "(1) company name," "(2) address," "(4) payee," and "area specific information" among the many sets of information in the billing source information management table of FIG. 14. A learning device may be generated by a machine learning algorithm using the actual image data of ``Bill Image''.

Next, the learning means 1322 stores the generated learning device in the learning information storage unit 111.

Through the above processing, the learning information manufacturing device 1 was able to store the billing source information and the learning device in the learning information storage unit 111. Note that the learning information manufacturing device 1 may store only one of the billing source information and the learning device in the learning information storage unit 111.

As described above, according to this embodiment, learning information can be easily acquired from a bill image. Note that the learning information is information that becomes the basis for easily and accurately acquiring and accumulating information to be managed from bill images.

Note that the processing in this embodiment may be realized by software. Then, this software may be distributed by software download or the like. Furthermore, this software may be recorded on a recording medium such as a CD-ROM and distributed. Note that this also applies to other embodiments in this specification. Note that the software that implements the learning information manufacturing device 1 in this embodiment is the following program. In other words, this program configures a computer to include a bill image receiving section that receives a bill image, a bill image output section that outputs the bill image received by the bill image receiving section, and a bill image output section that outputs the bill image received by the bill image receiving section. an area specifying information receiving unit that receives, for each billing related information, area specifying information specifying areas of two or more types of billing related information related to the billing for the output bill image; a billing-related information acquisition unit that acquires billing-related information that is a character string written in each of two or more areas; a pair of the area specifying information and billing-related information corresponding to the area specifying information; This is a program for functioning as a learning information composition unit that configures learning information using set information of , and a learning information storage unit that accumulates the learning information.

(Embodiment 2)

In this embodiment, using the learning information produced by the learning information production device 1 described in Embodiment 1, bill information having bill source information corresponding to reception information regarding a received bill image is acquired. An information system including an information processing device that outputs information will be described. Note that the reception information is, for example, billing-related information (eg, account number).

Furthermore, in this embodiment, an information system including an information processing device that performs the learning process described in Embodiment 1 when there is no invoice source information corresponding to reception information including an account number will be described. do.

Furthermore, in this embodiment, the reception information is a bill image, the bill image is image recognized, a similar bill image is acquired, bill information corresponding to the bill image is acquired, and the bill image is output. An information system including an information processing device will be described.

Furthermore, in the present embodiment, an information system including an information processing device that receives input of corrections from a second user to output bill information and acquires bill information to be stored will be described.

The conceptual diagram of information system B is the same as FIG. 1 except for the symbols. The information system B includes an information processing device 4, one or more first terminals 2, and one or more second terminals 3. The information processing device 4 is, for example, a so-called server. The server may be, for example, a so-called cloud server, an ASP server, etc., and its type, installation location, etc. do not matter.

FIG. 15 is a block diagram of information system B in this embodiment. FIG. 16 is a block diagram of the information processing device 4 that constitutes the information system B. As shown in FIG.

The information processing device 4 includes a storage section 41, a reception section 42, a processing section 43, and an output section 44. The storage unit 41 includes a learning information storage unit 111. The reception unit 42 includes a bill image reception unit 112, an area identification information reception unit 113, a reception information reception unit 421, and a modification reception unit 422. The processing unit 43 includes a bill information acquisition unit 431, a bill information storage unit 432, a bill related information acquisition unit 131, a learning information configuration unit 132, and a learning information storage unit 133. The bill information acquisition unit 431 includes a determination unit 4311 and a bill information acquisition unit 4312. The output unit 44 includes a bill image output unit 141 and a bill information output unit 441.

A storage unit 41 that constitutes the information processing device 4 stores various types of information. The various types of information are, for example, learning information. As described above, the learning information is, for example, one or more types of billing source information and one or more types of learning device information.

The reception unit 42 receives information, instructions, and the like. The information and instructions include, for example, a bill image, area specifying information, reception information, and correction instructions. The modification instruction is an instruction to modify bill information. A correction instruction can be simply called a correction.

The reception unit 42 normally receives information, instructions, etc. from the first terminal 2 or the second terminal 3. However, the reception unit 42 may also accept information, instructions, etc. from the user. In such a case, any means for inputting information, instructions, etc. may be used, such as a touch panel, keyboard, mouse, or menu screen. In such a case, the reception unit 42 may be realized by a device driver for input means such as a touch panel or a keyboard, control software for a menu screen, or the like.

The reception information reception unit 421 receives reception information.
The reception information is information regarding the bill image received by the bill image reception unit 112, and is information including an account number, for example. Note that the reception here is usually from the second terminal 3. However, the reception information may be received from an input device such as a keyboard, a mouse, or a touch panel, or information read from a recording medium such as an optical disk, a magnetic disk, or a semiconductor memory.

The modification accepting unit 422 accepts modifications to the bill information output by the bill information output unit 441. The reception here is also normally from the second terminal 3. However, the modification may be accepted by accepting information input from an input device such as a keyboard, mouse, or touch panel, or by accepting information read from a recording medium such as an optical disk, a magnetic disk, or a semiconductor memory.

The processing unit 43 performs various processes. The various processes are, for example, processes performed by the bill information acquisition unit 431, the bill information storage unit 432, the bill related information acquisition unit 131, the learning information configuration unit 132, the learning information storage unit 133, and the like. The various processes are, for example, processes that configure an input screen. The input screen is a screen for the second user to input or modify billing-related information constituting bill information while viewing the bill image. The input screen includes a bill image. Further, various types of processing include, for example, processing that configures a cutting instruction screen.

The bill information acquisition unit 431 has one or more types of bill-related information. The bill information acquisition unit 431 acquires bill information. There are various possible methods for the bill information acquisition unit 431 to acquire bill related information or bill information. Note that the bill information includes two or more types of bill-related information. The bill information includes, for example, organization name information (eg, billing company name), address information (eg, billing company address), billing amount information, payee information, and payment date information.

The bill information acquisition unit 431 acquires bill information using, for example, the reception information received by the reception information reception unit 421. The bill information acquisition unit 431 searches for bill source information using, for example, the reception information received by the reception information reception unit 421 as a key, and retrieves bill information that has all or part of the bill source information corresponding to the reception information. get.

The bill information acquisition unit 431 acquires bill information having two or more types of bill related information acquired by the bill related information acquisition unit 131.

The bill information acquisition unit 431 uses, for example, reception information that is information regarding the bill image received by the bill image reception unit 112 and learning information in the learning information storage unit 111 to obtain two or more types of bill-related information. Get billing information with .

For example, the bill information acquisition unit 431 searches the storage unit 41 for a bill image similar to the bill image received by the bill image reception unit 112. Then, the bill information acquisition unit 431 acquires, for example, two or more types of bill-related information paired with the similar bill image, and configures bill information having the two or more types of bill-related information.

Note that there may be various methods of searching for a bill image similar to the bill image received by the bill image receiving unit 112. The bill information acquisition unit 431 searches, for example, from the storage unit 41 that has the highest degree of similarity to the bill image and whose degree of similarity is equal to or higher than a threshold value. For example, the bill information acquisition unit 431 vectorizes the bill image received by the bill image reception unit 112, calculates the distance between the vector and the vector acquired from the bill image in the storage unit 41, and calculates the distance. Calculate the degree of similarity used as a parameter.

In addition, the bill information acquisition unit 431 applies the bill image received by the bill image reception unit 112 to a learning device using a machine learning algorithm, and generates bill information having two or more types of bill-related information. may be obtained. In such a case, it can be said that the bill information acquisition unit 431 has searched for a bill image similar to the bill image received by the bill image reception unit 112.

It goes without saying that the algorithm used in machine learning does not matter. For machine learning, for example, deep learning, SVM, SVR, random forest, decision tree, etc. can be used. In addition, in machine learning, when making a prediction (here, when obtaining bill information), if you give a learning device and a group of input information as arguments to the machine learning module, the predicted information (here, when obtaining bill information), billing information). Note that the predicted information may be said to be information as a result of classification.

The bill information acquisition unit 431 performs character recognition on the bill image received by the bill image reception unit 112, and obtains area information of one or more rectangular areas (for example, The upper left coordinates (x ₁ , y ₁ ) and the lower right coordinates (x ₂ , y ₂ )) are acquired, and the one or more pieces of area information and one or more parts that are images of the area indicated by each of the one or more pieces of area information are obtained. Two or more types of information, an image and a bill image, are input to a machine learning module, and a learning device in the learning information storage unit 111 is used to perform predictive processing using a machine learning algorithm. Bill information having the above billing-related information may be acquired.

It is preferable that the bill information acquisition unit 431 determines the bill-related information to be accumulated by the same process as the determining unit 1313 described above. When there is billing-related information transmitted from two or more second terminals 3, the billing information acquisition unit 431 determines the billing-related information to be accumulated, for example, when all the billing-related information matches. (unanimous method). Further, when there is billing-related information transmitted from three or more second terminals 3, the billing information acquisition unit 431 determines the billing-related information to be accumulated, for example, by selecting a large number of matching billing-related information. (majority voting method). Furthermore, the unanimous decision method or the majority decision method may be adopted only for variable billing related information.

The determining means 4311 searches the billing source information stored in the learning information storage section 111 using the reception information received by the reception information receiving section 421 as a key, and determines whether billing source information corresponding to the reception information exists. Determine whether or not. Note that the bill source information corresponding to the reception information is, for example, bill source information that has information that matches the reception information, or bill source information that is paired with information that matches the reception information. The bill source information corresponding to the reception information is, for example, bill source information having an account number that matches the account number that is the reception information.

The bill information acquisition unit 4312 acquires bill information including all or part of the bill source information when the determining unit 4311 determines that bill source information corresponding to the reception information exists.

The bill information storage section 432 accumulates bill information that has been corrected by the correction received by the correction reception section 422. The bill information storage section 432 may accumulate the bill information acquired by the bill information acquisition means 4312.

The bill information storage unit 432 performs image recognition processing on the bill image received by the bill image reception unit 112, and stores the acquired bill information (for example, company name, address, bill amount information, payee information, The bill information may be configured by partially or completely correcting the information (including payment date information) received by the correction accepting unit 422, and the configured bill information may be stored.

Note that approval from the first terminal 2 may be required before the bill information storage unit 432 stores bill information. That is, the bill information storage unit 432 may accumulate the bill information when approval is received from the first terminal 2. Note that approval may also be referred to as confirmation information.

The output unit 44 outputs various information. The various types of information include, for example, a bill image, bill information, an input screen, and a cutting instruction screen. Moreover, the output here usually means transmission to the second terminal 3. However, output refers to, for example, displaying on a display, projecting using a projector, printing on a printer, outputting sound, storing on a recording medium, passing processing results to other processing devices or programs, etc. It may also be a concept that includes

The bill information output unit 441 outputs the bill information acquired by the bill information acquisition unit 431. Here, the output means, for example, storage in the storage unit 41. Further, output means, for example, transmission to the second terminal 3.

Next, an example of the operation of the information processing device 4 that constitutes the information system B will be described using the flowchart of FIG. 17. Note that in the flowchart of FIG. 17, descriptions of the same steps as in the flowchart of FIG. 4 will be omitted.

(Step S1701) The bill image reception unit 112 determines whether or not a bill image has been received. If the bill image has been received, the process advances to step S1702; if the bill image has not been received, the process advances to step S410.

(Step S1702) The processing unit 43 stores the bill image received in step S1701 in the storage unit 11. The processing unit 43 stores the bill image received in step S1701 in the storage unit 11, for example, by pairing it with the first user identifier received in pair with the bill image.

(Step S1703) The bill information acquisition unit 431 searches the storage unit 41 for a bill image similar to the bill image received in step S1701.

(Step S1704) The bill information acquisition unit 431 determines whether a similar bill image could be retrieved in step S1703. If the search is successful, the process goes to step S1705; if the search is not successful, the process goes to step S1706.

(Step S1705) The bill information acquisition unit 431 acquires one or more types of bill-related information paired with the searched bill image from the learning information storage unit 111. Note that it is preferable that the one or more types of billing-related information acquired here are fixed billing-related information. Fixed billing related information is, for example, organization name information, address information, and payee information.

(Step S1706) The processing unit 43 configures an input screen that is a screen that includes the bill image received in step S1701. Note that it is preferable that the input screen includes one or more types of billing-related information. Further, the one or more types of billing-related information is, for example, information acquired by the billing information acquisition unit 431 in step S1705. Further, the one or more types of billing-related information is, for example, information acquired by the billing-related information acquisition unit 131 through character recognition processing on a bill image. Note that character recognition processing is omitted in the flowchart of FIG. 17.

(Step S1707) The output unit 44 transmits the input screen configured in step S1706 to one or more second terminals 3.

(Step S1708) The reception unit 42 determines whether one or more pieces of variable billing related information have been received from all one or more second terminals 3. If one or more variable billing related information is received from all one or more second terminals 3, the process goes to step S1709, and if one or more variable billing related information is not received, the process goes to step S1711. Note that the variable billing related information is, for example, billing amount information and payment date information. Further, here, the reception unit 42 may also receive fixed billing related information.

(Step S1709) The bill information acquisition unit 431 configures bill information having two or more pieces of bill-related information. Note that it is preferable that the bill information acquisition unit 431 determines the bill-related information to be accumulated by the same process as the determining unit 1313 described above, and configures the bill information using the bill-related information. .

(Step S1710) The bill information storage unit 432 stores the bill information configured in step S1709 in the storage unit 41 in pairs with the bill image received in step S1701. Return to step S1701.

(Step S1711) The reception unit 42 determines whether reception information (for example, account number) has been received from the second terminal 3. If reception information is received, the process goes to step S1712; if reception information is not received, the process returns to step S1711.

(Step S1712) The bill information acquisition unit 431 searches for bill source information using the reception information received in step S1711 as a key.

(Step S1713) The bill information acquisition unit 431 determines whether the bill source information could be retrieved in step S1712. If the search is successful, the process goes to step S1714; if the search is not successful, the process goes to step S403.

(Step S1714) The bill information acquisition unit 431 acquires one or more pieces of fixed billing related information corresponding to the reception information received in step S1711. Note that here, the bill information acquisition unit 431 may also acquire variable billing related information. The bill information acquisition unit 431 also acquires fixed billing related information or variable billing related information included in the newest (recently accumulated) learning information among the learning information corresponding to the reception information received in step S1711. It is preferable to obtain

(Step S1715) The bill information acquisition unit 431 performs character recognition processing on the bill image received in step S1701, and acquires variable billing related information. Note that the bill information acquisition unit 431 may acquire variable bill related information by the process described in step S1714 instead of the character recognition process.

(Step S1716) The processing unit 43 configures an input screen. This process is similar to step S1706.

(Step S1717) The output unit 44 transmits the input screen configured in step S1716 to one or more second terminals 3. This process is similar to step S1707.

(Step S1718) The receiving unit 42 determines whether one or more pieces of variable billing related information have been received from all one or more second terminals 3. If one or more variable billing related information is received from all one or more second terminals 3, the process goes to step S1719, and if one or more variable billing related information is not received, the process returns to step S1718. Note that the variable billing related information is, for example, billing amount information and payment date information. Further, here, the reception unit 42 may also receive fixed billing related information.

(Step S1719) The bill information acquisition unit 431 configures bill information having two or more pieces of bill-related information. Note that it is preferable that the bill information acquisition unit 431 determines the bill-related information to be accumulated by the same process as the determining unit 1313 described above, and configures the bill information using the bill-related information. . Furthermore, the two or more pieces of billing-related information that constitute the billing information may include, for example, the billing-related information received in step S1718 as is.

(Step S1720) The bill information storage unit 432 stores the bill information configured in step S1719 in the storage unit 41 in pairs with the bill image received in step S1701. Return to step S1701.

Note that in the flowchart of FIG. 17, there are a first process from step S1703 to step S1710 and a second process from step S1711 to step S1720 as the process to obtain and accumulate bill information. , the information processing device 4 may perform only one of the first process and the second process.

Further, in the flowchart of FIG. 17, the process is terminated by turning off the power or by an interrupt to terminate the process.

A specific example of the operation of the information system B in this embodiment will be described below. A conceptual diagram of information system B is shown in FIG.

The learning information storage unit 111 currently stores a billing source information management table having the structure shown in FIG. The learning information storage unit 111 also stores learning devices acquired by machine learning.

Here, it is assumed that the first user inputs the bill image shown in FIG. 8 into the first terminal 2. Then, the first reception unit 22 of the first terminal 2 receives the bill image. Next, the first processing unit 23 converts the bill image received by the first reception unit 22 into a bill image having a data structure to be transmitted. Next, the first transmitting unit 24 transmits the bill image configured by the first processing unit 23 to the information processing device 4.

Next, the bill image reception unit 112 of the information processing device 4 receives the bill image shown in FIG. Then, the processing unit 43 stores the received bill image in the storage unit 11.

Next, the bill information acquisition unit 431 searches the storage unit 41 for a bill image similar to the received bill image in FIG. However, assume here that there is no bill image whose similarity is within the threshold value in the storage unit 41.

Next, the processing unit 13 configures a screen to be transmitted to one or more second terminals 3. Then, the output unit 44 transmits the screen to one or more second terminals 3.

Next, the second receiving unit 35 of the second terminal 3 receives the screen. Next, the second processing unit 33 configures a screen to be output from the received screen. Then, the second output unit 36 outputs the configured screen. An example of such a screen is shown in FIG.

Next, it is assumed that the second user views the bill image, selects "Bill (Domestic)" 1801, and presses the "Confirm" button. Then, the second reception unit 32 receives information identifying the “bill (domestic)” 1801, and the second transmission unit 34 transmits this information to the information processing device 4.

Next, the receiving unit 42 of the information processing device 4 receives information identifying the “bill (domestic)” 1801. The processing unit 43 is a screen corresponding to such information, and constitutes a screen for inputting reception information. Next, the output unit 44 transmits this screen to the second terminal 3.

In addition, when selecting the type of bill in Figure 18, for example, only ``Bill (domestic)'' is accepted among ``Bill (domestic)'', ``Bill (overseas)'', ``Direct debit'', and ``Other''. A screen for inputting information is configured. In addition, when selecting the type of bill in Figure 18, for example, enter reception information only in the case of "Other" among "Bill (domestic)", "Bill (overseas)", "Direct debit", and "Other". It is also possible to perform processing such that no screen is configured for this purpose. That is, the processing unit 43 may determine whether the bill is of a predetermined type, and configure a screen for inputting reception information only in the case of a predetermined bill type. Further, the processing unit 43 may determine whether or not the bill is of a predetermined type, and may not configure the screen for inputting reception information only in the case of a predetermined bill type. .

Next, the second receiving unit 35 of the second terminal 3 receives a screen for inputting reception information. The second processing unit 33 then configures a screen for output. Next, the second output unit 36 outputs a screen for inputting reception information. Note that this screen includes a bill image and a field for inputting reception information.

Next, it is assumed that the second user of the second terminal 3 inputs the company name "Sansan Co., Ltd." and the account number "1407848" on the output screen. Note that here, the company name and account number are examples of reception information. However, the reception information may be just the account number.

Next, the second reception unit 32 of the second terminal 3 receives reception information. Next, the second processing unit 33 converts the received reception information into a data structure for transmission. Next, the second transmitter 34 transmits the reception information to the information processing device 4.

Next, the reception unit 42 of the information processing device 4 receives reception information from the second terminal 3.

Next, the bill information acquisition unit 431 searches the bill source information in the storage unit 41 using the received reception information as a key. Here, it is assumed that the bill information acquisition unit 431 was unable to search for bill source information. In other words, it is assumed that the account number "1407848" etc. does not exist in the storage unit 41.

Next, the processing unit 43 configures a screen when there is no invoice source information corresponding to the reception information. Then, the output unit 44 transmits the screen to the second terminal 3.

Next, the second receiving unit 35 of the second terminal 3 receives the screen. Then, the second processing unit 33 configures a screen to be output from the received screen. Next, the second output unit 36 outputs the screen. Note that an example of such a screen is shown in FIG. 19. Further, in FIG. 19, if the target transfer account (account number) does not exist, "Not applicable" is displayed in the transfer account display field 1901.

Next, the processing described in the first embodiment, from step S403 to step S409 in the flowchart of FIG. 17, is executed.

On the other hand, for example, if the learning information storage section 111 stores the invoice source information management table shown in FIG. 14, the invoice information acquisition section 431 uses the received reception information as a key to Assume that as a result of searching the billing source information of 41, the billing source information of "ID=892" in FIG. 14 was retrieved.

Then, the bill information acquisition unit 431 acquires three pieces of fixed bill-related information from among the attribute values of the record "ID=892" in FIG. The three fixed billing related information are (1) Company name "Sansan Co., Ltd.", (2) Address "Jingumae, Shibuya-ku, Tokyo 150-0001...", (4) Payment address "Resona Bank..."・、Mizuho Bank...、Sumitomo Mitsui Banking..."

Next, the bill information acquisition unit 431 performs character recognition processing on the received bill image and acquires variable bill related information. Here, the variable billing related information includes the billing amount "1" and the payment date "October 31, 2017."

Next, the processing unit 43 configures an input screen for the second user to modify billing-related information in order to obtain billing information.

Next, the output unit 44 transmits the configured input screen to one or more second terminals 3.

Next, the second receiving unit 35 of one or more second terminals 3 receives the input screen. The second processing unit 33 then configures an input screen to be displayed using the received input screen. Next, the second output unit 36 displays the configured input screen. An example of such an input screen is shown in FIG. In FIG. 20, the company name is displayed in the field 2001, the postal code composing the address is displayed in the field 2002, and the payee is displayed in the field 2003.

The second user then uses the screen in FIG. 20, the screens in FIGS. 11 to 13, etc. to check or modify the bill information displayed on the second terminal 3, and to request changes in the bill amount, etc. Enter related information. Next, the second receiving unit 32 of the second terminal 3 receives bill information including variable billing related information. Next, the second processing unit 33 configures bill information in a data structure to be transmitted. Then, the second transmitting unit 34 transmits the bill information to the information processing device 4.

Next, the reception unit 42 of the information processing device 4 receives bill information including one or more variable billing related information from all the one or more second terminals 3.

Next, the bill information acquisition unit 431 uses the bill information received from one or more second terminals 3 to configure bill information to be accumulated.

Next, the bill information storage unit 432 stores the configured bill information in the storage unit 41 in pairs with the received bill image.

Through the above processing, the information processing device 4 was able to efficiently accumulate learning information with high accuracy.

As described above, according to the present embodiment, information to be managed can be easily and accurately acquired and stored from bill images.

Further, according to the present embodiment, information to be managed can be easily and accurately acquired and stored from invoice images using learning information, and learning information can be evolved.

Note that the processing in this embodiment may be realized by software. Then, this software may be distributed by software download or the like. Furthermore, this software may be recorded on a recording medium such as a CD-ROM and distributed. Note that this also applies to other embodiments in this specification. Note that the software that implements the information processing device 4 in this embodiment is the following program. In other words, this program connects a computer that can access the learning information storage section in which the learning information accumulated by the learning information production device 1 is stored to the invoice image reception section that receives the invoice image and the invoice image reception section. a bill information acquisition unit that acquires bill information having two or more types of bill-related information using reception information that is information about the received bill image and learning information in the learning information storage unit; This is a program for functioning as a bill information output unit that outputs bill information.

Further, FIG. 21 shows the external appearance of a computer that executes the programs described in this specification and realizes the information processing apparatus 4 and the like of the various embodiments described above. The embodiments described above may be implemented in computer hardware and computer programs executed thereon. FIG. 21 is an overview diagram of this computer system 300, and FIG. 22 is a block diagram of the system 300.

In FIG. 21, computer system 300 includes a computer 301 including a CD-ROM drive, a keyboard 302, a mouse 303, and a monitor 304.

In FIG. 22, the computer 301 includes, in addition to a CD-ROM drive 3012, an MPU 3013, a bus 3014 connected to the CD-ROM drive 3012, a ROM 3015 for storing programs such as a boot-up program, and an MPU 3013. It is connected and includes a RAM 3016 for temporarily storing application program instructions and providing temporary storage space, and a hard disk 3017 for storing application programs, system programs, and data. Although not shown here, the computer 301 may further include a network card that provides connection to a LAN.

A program that causes the computer system 300 to execute the functions of the information processing device 4 and the like in the embodiment described above may be stored in the CD-ROM 3101, inserted into the CD-ROM drive 3012, and further transferred to the hard disk 3017. . Alternatively, the program may be transmitted to the computer 301 via a network (not shown) and stored on the hard disk 3017. The program is loaded into RAM 3016 during execution. The program may be loaded directly from CD-ROM 3101 or the network.

The program does not necessarily need to include an operating system (OS) or a third party program that causes the computer 301 to execute the functions of the information processing device 4 of the above-described embodiment. The program need only contain those parts of the instructions that call the appropriate functions (modules) in a controlled manner to achieve the desired results. How computer system 300 operates is well known and will not be described in detail.

Further, the number of computers that execute the above program may be one or more. That is, centralized processing or distributed processing may be performed.

Furthermore, in each of the above embodiments, it goes without saying that two or more communication means existing in one device may be physically realized by one medium.

Furthermore, in each of the above embodiments, each process may be realized by being centrally processed by a single device, or may be realized by being distributed by a plurality of devices.

It goes without saying that the present invention is not limited to the above-described embodiments, and that various modifications can be made, and these are also included within the scope of the present invention.

As described above, the learning information manufacturing device according to the present invention has the effect of being able to easily and accurately acquire and store information to be managed from invoice images, and can be used as a learning information manufacturing device, etc. Useful.

1 Learning information production device 2 First terminal 3 Second terminal 4 Information processing device 11, 41 Storage section 12, 42 Reception section 13, 43 Processing section 14, 44 Output section 21 First storage section 22 First reception section 23 First Processing section 24 First transmitting section 25 First receiving section 26 First output section 31 Second storage section 32 Second receiving section 33 Second processing section 34 Second transmitting section 35 Second receiving section 36 Second output section 111 Learning information Storage unit 112 Bill image reception unit 113 Area identification information reception unit 131 Billing related information acquisition unit 132 Learning information composition unit 133 Learning information storage unit 141 Bill image output unit 421 Reception information reception unit 422 Correction reception unit 431 Bill information acquisition Section 432 Bill information storage section 441 Bill information output section 1311 Character recognition means 1312 Bill-related information reception means 1313 Determination means 1321 Bill source information composition means 1322 Learning means 4311 Judgment means 4312 Bill information acquisition means

Claims (14)

a bill image reception unit that accepts bill images;
a bill image output unit that outputs the bill image received by the bill image reception unit;
For the bill image output by the bill image output unit, area specifying information that specifies areas of two or more types of billing-related information related to billing is associated with a type identifier that specifies the type of billing-related information. an area specific information reception department that accepts each billing-related information;
a billing-related information acquisition unit that acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
a learning information configuration unit that configures learning information using two or more sets of information that are a set of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information;
and a learning information storage unit that stores the learning information,
The billing-related information acquisition unit includes:
character recognition means for performing character recognition processing on the bill image received by the bill image receiving unit and acquiring first bill-related information;
billing-related information receiving means for receiving manually input second billing-related information;
determining means for determining the first billing-related information or the second billing-related information as the billing-related information to be accumulated when the first billing-related information and the second billing-related information match; Equipped with
The learning information configuration section includes:
The learning information manufacturing device configures the learning information using one or more sets of the area specifying information and the billing related information determined by the determining means. a bill image reception unit that accepts bill images;
a bill image output unit that outputs the bill image received by the bill image reception unit;
For the bill image output by the bill image output unit, area specifying information that specifies areas of two or more types of billing-related information related to billing is associated with a type identifier that specifies the type of billing-related information. an area specific information reception department that accepts each billing-related information;
a billing-related information acquisition unit that acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
a learning information configuration unit that configures learning information using two or more sets of information that are a set of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information;
and a learning information storage unit that stores the learning information,
The billing-related information acquisition unit includes:
billing-related information receiving means for receiving two or more pieces of second billing-related information manually input by each of the two or more users;
determining means for determining the second billing-related information as the billing-related information to be accumulated when the degree of coincidence of the two or more second billing-related information satisfies a condition;
The learning information configuration section includes:
A learning information manufacturing device that configures learning information using one or more sets of the area specifying information and billing-related information determined by the determining means. The learning information configuration section includes:
comprising billing source information configuring means for configuring billing source information having two or more sets of information, each of which is a pair of the area specifying information and billing related information corresponding to the area specifying information;
The learning information manufacturing device according to claim 2 , wherein the learning information includes the billing source information. The learning information configuration section includes:
comprising a learning means for generating a learning device by a machine learning algorithm using two or more sets of information that are a set of the area specifying information and billing related information corresponding to the area specifying information,
The learning information manufacturing device according to any one of claims 1 to 3 , wherein the learning information includes the learning device. The two or more types of billing-related information include billing organization information regarding the billing organization including the billing organization name, one or more payee information regarding the payment destination, billing amount information regarding the billing amount, and payment due date information regarding the payment date. The learning information production device according to any one of claims 1 to 4 , comprising: Learning configured using two or more sets of information, which are a set of billing-related information related to a billing, a type identifier that specifies the type of the billing-related information, and area specifying information that specifies the area of the billing-related information. a learning information storage section in which information is stored;
a bill image reception unit that accepts bill images;
A bill that obtains bill information having two or more types of bill-related information using reception information that is information regarding a bill image received by the bill image reception unit and learning information in the learning information storage unit. Information acquisition department;
and a bill information output unit that outputs the bill information,
The learning information includes billing source information,
The bill information acquisition unit includes:
Using the invoice image received by the invoice image reception unit, the learning information storage unit is searched, and one or more pieces of fixed information that are part of the invoice source information corresponding to the reception information are searched. An information processing device that acquires fixed billing related information of , and obtains bill information including the one or more fixed billing related information . The bill information acquisition unit includes:
determining means for determining whether or not invoice source information corresponding to the invoice image exists;
If the determining means determines that bill source information corresponding to the bill image exists, the one or more pieces of fixed billing related information are acquired, and the billing information including the one or more fixed billing related information is obtained. and bill information acquisition means,
a bill image output unit that outputs the bill image received by the bill image reception unit;
If the determining means determines that there is no billing source information corresponding to the billing image , the billing image output unit outputs two or more types of billing information related to billing. an area specifying information reception unit that accepts area specifying information specifying an information area for each billing-related information;
a billing-related information acquisition unit that acquires billing-related information that is a character string written in each of two or more areas specified by the area specifying information;
a learning information configuration unit that configures learning information using two or more sets of information that are a set of the area specifying information and billing related information corresponding to the area specifying information;
further comprising a learning information storage unit that stores the learning information,
The bill information acquisition unit includes:
The information processing apparatus according to claim 6 , wherein the billing-related information acquisition unit acquires bill information having two or more types of billing-related information. a modification reception unit that accepts modifications to the bill information output by the bill information output unit;
8. The information processing apparatus according to claim 6, further comprising a bill information storage section that stores bill information modified by the modification received by the modification reception section. A learning information manufacturing method realized by a bill image receiving section, a bill image outputting section, an area specifying information receiving section, a billing related information acquiring section, a learning information composition section, and a learning information storage section. hand,
a bill image receiving step in which the bill image receiving unit receives a bill image;
a bill image output step in which the bill image output unit outputs the bill image accepted in the bill image receiving step;
The area specifying information receiving unit adds area specifying information that specifies areas of two or more types of billing related information related to billing to the bill image output in the bill image output step. an area specifying information receiving step of accepting each piece of billing-related information in association with a type identifier that specifies the type;
a billing-related information acquisition step in which the billing-related information acquisition unit acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
A learning information configuration step in which the learning information configuration unit configures learning information using two or more sets of information, which are a pair of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information. and,
The learning information storage unit includes a learning information storage step for storing the learning information ,
In the billing related information acquisition step,
a character recognition substep of performing character recognition processing on the bill image received by the bill image receiving unit to obtain first bill-related information;
a billing-related information reception substep of receiving manually input second billing-related information;
a determining substep of determining the first billing-related information or the second billing-related information as the billing-related information to be accumulated when the first billing-related information and the second billing-related information match; and
In the learning information configuration step,
A method for producing learning information, wherein the learning information is configured using one or more sets of the area specifying information and the billing-related information determined in the determining substep . A learning information manufacturing method realized by a bill image receiving section, a bill image outputting section, an area specifying information receiving section, a billing related information acquiring section, a learning information composition section, and a learning information storage section. hand,
a bill image receiving step in which the bill image receiving unit receives a bill image;
a bill image output step in which the bill image output unit outputs the bill image accepted in the bill image receiving step;
The area specifying information receiving unit sends area specifying information that specifies areas of two or more types of billing related information related to billing to the bill image output in the bill image output step. an area specifying information receiving step of accepting each piece of billing-related information in association with a type identifier that specifies the type;
a billing-related information acquisition step in which the billing-related information acquisition unit acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
a learning information configuration step in which the learning information configuration unit configures learning information using two or more sets of information that are a set of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information; and,
The learning information storage unit includes a learning information storage step for storing the learning information,
The billing related information acquisition step includes:
a billing-related information reception substep of receiving two or more pieces of second billing-related information manually input by each of the two or more users;
a determining substep of determining the second billing-related information as billing-related information to be accumulated when the degree of coincidence of the two or more pieces of second billing-related information satisfies a condition;
In the learning information configuration step,
A method for producing learning information, wherein the learning information is configured using one or more sets of the area specifying information and the billing-related information determined in the determining substep. Learning configured using two or more sets of information, which are a set of billing-related information related to a billing, a type identifier that specifies the type of the billing-related information, and area specifying information that specifies the area of the billing-related information. An information processing method realized by a learning information storage section in which information is stored, a bill image reception section, a bill information acquisition section, and a bill information output section,
a bill image receiving step in which the bill image receiving unit receives a bill image;
The bill information acquisition unit obtains two or more types of bill-related information using reception information that is information regarding the bill image received in the bill image reception step and learning information in the learning information storage unit. a bill information acquisition step of acquiring bill information having the
The bill information output unit includes a bill information output step of outputting the bill information ,
The learning information includes billing source information,
In the bill information acquisition step,
Using the invoice image received by the invoice image reception unit, the learning information storage unit is searched, and one or more pieces of fixed information that are part of the invoice source information corresponding to the reception information are searched. An information processing method comprising: acquiring fixed billing related information of the fixed billing related information, and obtaining bill information including the one or more fixed billing related information . computer,
a bill image reception unit that accepts bill images;
a bill image output unit that outputs the bill image received by the bill image reception unit;
For the bill image output by the bill image output unit, area specifying information that specifies areas of two or more types of billing-related information related to billing is associated with a type identifier that specifies the type of billing-related information. an area specific information reception department that accepts each billing-related information;
a billing-related information acquisition unit that acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
a learning information configuration unit that configures learning information using two or more sets of information that are a set of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information;
A program for functioning as a learning information storage unit that stores the learning information ,
The billing-related information acquisition unit includes:
character recognition means for performing character recognition processing on the bill image received by the bill image receiving unit and acquiring first bill-related information;
billing-related information receiving means for receiving manually input second billing-related information;
determining means for determining the first billing-related information or the second billing-related information as the billing-related information to be accumulated when the first billing-related information and the second billing-related information match; Equipped with
The learning information configuration section includes:
A program for causing the computer to function as configuring the learning information using one or more sets of the area specifying information and the billing-related information determined by the determining means. computer,
a bill image reception unit that accepts bill images;
a bill image output unit that outputs the bill image received by the bill image reception unit;
For the bill image output by the bill image output unit, area specifying information that specifies areas of two or more types of billing-related information related to billing is associated with a type identifier that specifies the type of billing-related information. an area specific information reception department that accepts each billing-related information;
a billing-related information acquisition unit that acquires billing-related information, which is a character string written in each of two or more areas specified by the area specifying information, in association with a type identifier;
a learning information configuration unit that configures learning information using two or more sets of information that are a set of the area specifying information, the type identifier, and billing related information corresponding to the area specifying information;
A program for functioning as a learning information storage unit that stores the learning information,
The billing-related information acquisition unit includes:
billing-related information receiving means for receiving two or more pieces of second billing-related information manually input by each of the two or more users;
determining means for determining the second billing-related information as the billing-related information to be accumulated when the degree of coincidence of the two or more second billing-related information satisfies a condition;
The learning information configuration section includes:
A program for causing the computer to function as configuring learning information using one or more sets of the area specifying information and billing related information determined by the determining means. Learning configured using two or more sets of information, which are a set of billing-related information related to a billing, a type identifier that specifies the type of the billing-related information, and area specifying information that specifies the area of the billing-related information. A computer that can access the learning information storage section where information is stored.
a bill image reception unit that accepts bill images;
A bill that obtains bill information having two or more types of bill-related information using reception information that is information regarding a bill image received by the bill image reception unit and learning information in the learning information storage unit. Information acquisition department;
A program for functioning as a bill information output unit that outputs the bill information ,
The learning information includes billing source information,
The bill information acquisition unit includes:
Using the invoice image received by the invoice image reception unit, the learning information storage unit is searched, and one or more pieces of fixed information that are part of the invoice source information corresponding to the reception information are searched. A program for causing the computer to function as a program for acquiring fixed billing related information of 1 and acquiring bill information including the one or more fixed billing related information .

Images