JP6389913B1 - Receipt data system, receipt prize server, and program - Google Patents

Abstract

A system, server, program, and the like that can collect purchase data extensively by providing a receipt prize service.
A receipt prize server that communicates with a user terminal and provides a receipt prize service, a receipt image acquisition unit for acquiring a receipt image obtained by photographing a paper receipt, and a receipt image data for reading the receipt image data from the receipt image A receipt image data reading unit, a prize information management unit for managing information related to a prize based at least on the receipt image data, and a purchase information management for managing information related to purchase based on at least the receipt image data Part.
[Selection] Figure 2

Description

  The present invention relates to a receipt data system, method, and the like that cooperate with provision of a service using a receipt. In particular, the present invention relates to a system, a server, and a program that can collect purchase data extensively by providing a receipt prize service.

  The diversification of consumption makes it difficult to predict people's consumption behavior, and purchase data in units of shopping receipts (purchase history) that is not limited to POS data and ID-POS data is required.

  For example, an electronic receipt system is introduced in a store, and the user can use his / her “electronic receipt application” to manage his / her own shopping receipt (purchase history) data on his / her smartphone. In addition, an attempt was made to receive provision of personal information such as age and purchase history data from the user. However, the introduction of an electronic receipt system is costly and stores are limited. At present, purchase data that reflects a huge amount of paper receipts issued every day has not been obtained.

  There is also known a “household account book application” that allows a user to take a paper receipt with a smartphone camera or the like, and to easily create a household account book by OCR (Optical Character Recognition) processing of the receipt image. The business operator of the “household account book application” can collect data read by the OCR with the user's consent. However, since the contents of paper receipts (product names, service names, etc.) and description methods have various patterns, and OCR is misread, it is not always possible to obtain effective purchase data. Even using advanced OCR technology, it was considered difficult to fully link to a unified product and service master.

  Conventionally, receipt image data registered by the consumer is received from the input person terminal, the checker terminal used by the checker displays the answer data of the consumer and the image data of the receipt, A marketing data collection system has been proposed in which the consistency of response data is determined using a receipt image to check consistency and convert “data including the reason for purchase for each product / service type” into marketing data (Japanese Patent Laid-Open No. 2014). -49058 gazette: Patent Document 1).

JP 2014-49058 A

  If an operation such as “input answer” is required to perform “individual” marketing instead of “mass”, workers tend to be biased and there is a fear that sufficient data cannot be obtained. Moreover, in order to obtain objective data, it is originally desirable to collect the description content of the paper receipt itself as useful data so as to be suitable for various types of analysis.

  In view of the above, an object of the present invention is to provide a method, a system, and a program that allow a user to easily participate and obtain an appropriate consideration so that the user can be satisfied and provide their own data. . An object of the present invention is to provide a server system in which providing a receipt-related service leads to collection of useful purchase data. An object of the present invention is to provide a system that can collect and provide a wide range of purchase data that has not been available.

  One aspect of the present invention for solving the above problem is a receipt prize server that communicates with a user terminal and provides a receipt prize service, and a receipt image acquisition unit for acquiring a receipt image obtained by photographing a paper receipt; A receipt image data reading unit for reading receipt image data from a receipt image, a prize information management unit for managing information related to a prize based on at least the receipt image data, and purchasing based on at least the receipt image data A purchase information management unit for managing related information.

  In the receipt prize server according to the present invention, provision of a receipt prize service and collection of purchase data are performed in conjunction with each other. Since the receipt of the receipt is an application condition for the sweepstakes, it is easy for the user to provide the receipt by convincing the content of the prizes. Receipts can be collected randomly from users who apply for sweepstakes, and data without store (distribution) or regional restrictions can be collected.

  Preferably, the receipt image data includes at least the product name read from the receipt image, and the purchase information management unit determines whether or not the read product name matches the receipt notation product name in the product master database. When it is determined not to do so, a screen for reading the barcode of the product with the read product name is transmitted to the user terminal.

  Further, the receipt prize server includes a user information management unit for managing information for each user, and points are calculated in the user information management unit when a barcode is read by a user operation.

  By doing in this way, a useful database can be generated by reading a barcode by an easy user operation. Necessary user operations can be promoted by giving incentives to the user.

  Alternatively, the receipt image data preferably includes at least a product name read from the receipt image, and the purchase information management unit determines whether or not the read product name matches the receipt notation product name in the product master database. When it is determined that the product names are not exact but match, the read product name is corrected to be determined to match.

  By doing in this way, even if there is misreading of OCR, a database including an accurate receipt notation product name can be generated, and a useful analysis based on the receipt notation product name can be performed.

  Another aspect of the present invention is a method for providing a receipt prize service suitable for collecting purchase information, the method comprising: obtaining a receipt image obtained by photographing a paper receipt; and reading receipt image data from the receipt image; Determining whether or not to apply for the prize based on the receipt image data, and accepting the application when it is possible, and acquiring the purchase information from at least the receipt image data received the application, The obtaining step includes determining whether or not a product is specified from the receipt image data with reference to a product master database.

  In this way, it is possible to perform processing in accordance with the specification or non-specification of the product, and the database can be updated in real time for seasonal products with a fast cycle.

  Further, when a merchandise master database is generated for each retailer, and the receipt image data includes at least a telephone number, and the retailer is identified based on the telephone number, the merchandise master database of the identified retailer is Referenced. This makes it possible to specify the product more accurately.

  Another aspect of the present invention is a program that causes a computer of a server device to execute the above method.

  According to the present invention, a receipt prize service is provided in the form of a web application, and a wide range of receipt-related data without bias can be collected because the user can easily participate due to the validity and entertainment of the contents of the prize. By providing the consideration, the user can provide the data with peace of mind and consent.

  Further, according to the present invention, data can be sequentially cleaned by various databases, and effective purchase information can be generated regardless of the diversity of paper receipts and the uncertainty of OCR. According to the present invention, various analyzes such as a distribution unit and a merchandise unit can be performed from analysis of a single receipt, and extremely useful big data can be constructed.

FIG. 1 schematically shows a receipt database system according to the present invention. FIG. 2 schematically shows a receipt prize server and a database server according to the present invention. FIG. 3 schematically shows various data related to a receipt obtained through a user terminal according to an embodiment of the present invention. FIG. 4 schematically shows a data structure of a product master database according to one embodiment of the present invention. FIG. 5 schematically shows the data structure of a user database according to an embodiment of the present invention. FIG. 6 schematically shows the data structure of a store master database according to one embodiment of the present invention. FIG. 7 schematically shows a data structure of a receipt (purchase) database according to one embodiment of the present invention. FIG. 8 schematically shows the data structure of a prize database according to one embodiment of the present invention. FIG. 9 is a diagram for explaining the correction of the read product name to the receipt written product name according to the present invention. FIG. 10 is a process flow diagram of one embodiment according to the present invention.

  Various features of the present invention will now be described with reference to the drawings, together with preferred embodiments not intended to limit the invention. The drawings are simplified and schematic for illustrative purposes.

  Referring to FIG. 1, a receipt database system 100 according to the present invention includes a receipt prize server 10, a database (DB) server 20 that manages a database (DB) device, a business personal computer (PC) 30, a receipt prize server 10, and the Internet. And a plurality of user terminals (41, 42, 43, etc.) that communicate via the communication network 1.

  The receipt prize server 10, the DB server 20, and the PC 30 are mounted with a central processing unit (CPU), RAM, ROM, hard disk, etc., execute a programming language under the control of an appropriate operating system (OS), and perform various processes. Functional means for performing are provided. The program stored in the receipt prize server 10 may be constructed by HTML, JavaScript (registered trademark), a native program (object code), or the like, and preferably has a unique function (camera device, such as the user terminal 41, 42, 43). It is constructed to call the control etc.). In the DB server 20, a database management system (DBMS, RDBMS) is operating. The PC 30 can be used as needed by an operator or the like for proper operation of the receipt database system 100.

  The user terminals 41, 42, 43, and the like may be high-functional terminals such as smartphones or tablet-type high-functional terminals equipped with known camera devices and browser software.

  FIG. 2 shows a block diagram of the receipt prize server 10 and the DB server 20 according to the present invention.

  The receipt prize server 10 receives various information and requests from the network 1 via the network interface (NIC) 11. The receipt prize server 10 includes a receipt image acquisition unit 12, a data reading unit 13, a prize information management unit 14, a purchase information management unit 15, and a user information management unit 16.

  The DB server 20 includes a product master database 21, a user database 22, a store (distribution) master database 23, a receipt (purchase) database 24, and a prize database 25.

  The receipt image acquisition unit 12 activates the camera device of the user terminal so that the user captures a paper receipt at the user terminal (41, 42, or 43), and receives the receipt image captured by the user's shooting button tap or the like. It consists of electronic equipment, electronic circuits and programs for acquisition.

  The data reading unit 13 includes a known OCR engine, and includes an electronic device, an electronic circuit, and a program for obtaining text data (receipt image data) by performing OCR processing on a photographed receipt image.

  The prize information management unit 14 determines whether or not to apply for the prize based on the receipt image data acquired by the OCR, and registers and manages data relating to the prize application in the prize database 25 or the like. And a program.

  The purchase information management unit 15 appropriately cleans the receipt image data based on an existing database, or prompts user input of data necessary for cleaning as necessary, and the cleaned data and / or before cleaning. The electronic data, the electronic circuit, and the program for registering and managing the data in the receipt DB 24 and the like.

  The user information management unit 16 includes an electronic device, an electronic circuit, and a program for registering and managing the ID, passcode, attribute information, and the like for each user accessing the server 10 in the user DB 22.

  The receipt prize server 10 further includes a data processing unit 17. The data processing unit 17 can register and search data in the RDB 20, create search results, etc., add data to each DB, generate search results, and delete data. , Enabling processing such as data update. Each data acquired and created by each unit of the receipt prize server 10 is stored in each DB via the data processing unit 17 and the storage interface 18.

  FIG. 3 schematically shows various data related to a receipt, which is acquired via a user terminal according to an embodiment of the present invention.

  Preferably, the user agrees to provide a receipt, downloads a predetermined application program in advance to each user terminal (41, 42, 43, etc.), and is given a user ID and a passcode. The user activates the downloaded application and accesses the server 10, and the server 10 acquires a user ID (b) and transmits a screen corresponding to the user (receipt shooting screen, prize application history, winning result, etc.). be able to.

  When the user takes a paper receipt according to the screen displayed on the terminal, a receipt image (b) is acquired by the receipt image acquisition unit 12. The data reading unit 13 performs OCR processing on the receipt image and reads various kinds of read data (telephone number (c), date and time, minutes and seconds (d), total amount (e), number of items (f), product name (h)), unit price (Li), quantity (nu), etc.) can be acquired. The acquired data need not be limited to these.

  For the read product name (H), referring to FIG. 9A, for example, when a product name consisting of 11 characters is read, the data reading unit 13 refers to an OCR conversion table (not shown), and Based on a set of frequently registered OCR misread characters registered in the table (for example, “Pa”, “B”, etc.), a plurality of candidates for reading results may be held. The illustrated example shows that the first character has a two-character candidate and the sixth character has a three-character candidate.

  The receipt ID (g) may be an arbitrary character string generated for each receipt image (b), and may be generated when the receipt image (b) is received. Alternatively, the receipt ID (g) is based on the read data (for example, telephone number (c), date and time, minutes and seconds (d), total amount (e), number of items (f), etc.) The receipt image (b) can be generated so as to be uniquely identifiable.

  The reading barcode (LE) and the product image (W) are operated by the user holding the terminal over the barcode of the purchased product or tapping the shooting button according to a screen (not shown) transmitted from the server 10. Can be obtained by doing.

  FIG. 4 schematically shows the data structure of the product master DB 21. The product master DB 21 is generated in advance for each store (distribution) based on a set of purchased products (services) and their receipts. In the product master DB 21, a product name (receipt written product name) described in the receipt is associated with a unique code. Here, when the product has a barcode attached to itself (or its package), the unique code refers to the barcode (JAN code, JAN symbol, etc.). Registration of data in the product master DB 21 may be performed manually by the operator via the PC 30. Or, an operator can take a paper receipt using a device such as a smartphone, read the receipt image OCR, and if there is an error in the reading result, manually correct it, and read and register the product barcode with a barcode reader. May be. Alternatively, a receipt image or product image taken by an arbitrary registrant, an OCR-read text data, and a barcode read data are received using an arbitrary application program, etc., and the operator visually checks the received data. After confirming and correcting any misreading, it may be registered in the product master DB 21.

  When a product or the like does not have a JAN code, the unique code indicates an arbitrary character string that can identify the product or the like. Since products or services such as fresh food, meat, fish, and side dishes generally do not have a JAN code, such products and services are registered in the DB as a combination of a general name and an identification code as listed in, for example, a Japanese dictionary. Can be done.

  FIG. 5 schematically shows the data structure of the user DB 22. For each user (ID), the user DB 22 can store a pass code, user attributes (such as age, gender, family composition, etc.), calculated points, and prize information (including winning amount, winning gift code, etc.).

  FIG. 6 schematically shows the data structure of the store master DB 23 according to the present invention. There are generally patterns for each distribution (retailer) in the name of the purchase object (receipt written product name) described on the paper receipt, and the product master DB 21 can be generated for each store (distribution) as described above. The store master DB 23 is generated so as to be able to specify the receipt issuing store (distribution) from the receipt image data so as to refer to the product master DB 21 of an appropriate store (distribution). In general, store (distribution) names on paper receipts may be designed and successful OCR may be difficult, but store phone numbers are often a string of numbers in a basic font and are successful. Easy to read. In the store master DB 23, a telephone number, a store, and a retailer (distribution) that develops the store are linked. Further, the store address (region) and the business type can be linked.

  FIG. 7 schematically shows the data structure of the receipt (purchase) DB 24 according to the present invention. FIG. 7A shows a data structure 24a when a product (unique code) is specified based on the receipt image data.

  The purchase information management unit 15 refers to the store master DB 23 based on the reading telephone number (c) of the receipt image data. When the store (distribution) is specified, the purchase information management unit 15 refers to the product master DB 21 corresponding to the specified store. can do. If not specified, the entire product master DB 21 can be referred to. It is determined whether or not the read product name (h) matches any receipt-written product name in the product master DB 21.

  For example, the degree of coincidence between the read product name in FIG. 9A and the product name in the receipt master DB 21 in FIG. 9B is calculated. In the figure, the shapes of squares and circles indicate matching and non-matching of the characters, and the dotted-line squares and the dashed-dotted line squares are characters that are equated with the solid-line squares by the ODR conversion master. The coincidence degree (or disagreement degree) of the receipt-notation product name having the highest coincidence degree (lowest inconsistency degree) is compared with a predetermined threshold value. The degree of coincidence (or degree of inconsistency) is calculated by, for example, the Levenshtein distance. The Levenshtein distance is a comparison between two character strings and quantifies the procedure required to transform one character string into the other. For example, when “cost” of “insertion”, “deletion”, and “replacement” is set to “1”, in the illustrated example, the third character and the ninth character ( If the 12th character (blank part) is inserted and the product name matches the receipt notation product name, the cost (distance) is “3”. Conversion between characters (first character and sixth character) based on an OCR conversion table storing a set of characters that are easily misread by OCR may not be included in the cost. When the allowable cost is 30% or less with respect to the number of characters of the read product name (determined as coincidence), in the illustrated example, the cost (distance) is equal to the allowable cost of 11 × 0.3≈3 (rounded down). Therefore, it is determined that the read product name matches the receipt notation product name.

  Match or mismatch is not limited to the Levenshtein distance, and may be determined based on an arbitrary measure. The threshold values for coincidence determination and non-coincidence determination are not limited to the above example, and may be determined arbitrarily or empirically as long as desired accuracy is obtained.

  If “matching determination” is not complete matching as described above, the read merchandise name (H) is corrected to the receipt notation merchandise name in the merchandise master DB 21 and stored in the receipt DB 24. Accordingly, the receipt DB 24a includes receipt image data (d, li, no, g, i, etc.) and data specified based on the receipt image data (unique code, receipt notation product name, distribution (store), region, etc.). Can be included.

  FIG. 7B shows the data structure of the receipt DB 24b when no product (unique code) or store is specified based on the receipt image data. The read receipt image data and, when a barcode is read by a user operation, the read barcode (LE) are stored. As shown in FIG. 7C, the receipt DB 24c can further hold a receipt image and a product image for each receipt ID. The operator of the system 100 conducts an appropriate investigation such as visually confirming the receipt image (b) for the data for which the product is not specified, so that the unique code and the receipt notation product are obtained from the read barcode (l) and the read product name (h). A set of names can be specified, or distribution etc. can be specified from a telephone number (c). By generating the receipt DB 24a thus cleaned, a useful database can be obtained. Furthermore, the merchandise master DB 21 and the store master DB 23 can be updated based on the cleaned data.

  FIG. 8 schematically shows the data structure of the sweepstakes DB 25. The prize DB 25 stores an application period for each prize ID, and information relating to the distribution (store phone number list, etc.) for a specific distribution, a winning number used for the lottery, and the like. Further, the receipt ID, the total amount, the user ID, and the like that accepted the application can be stored.

  The winning number may be an n-digit number string that is arbitrarily determined so that a lottery is performed by matching or non-matching with the last n digits of the total amount. The n-digit number sequence may be generated by a random number generator. When the last three digits of the total amount are subject to lottery, the winning probability is 1/1000.

  FIG. 10 shows a processing flow of one embodiment in the receipt prize server 10. A user ID is acquired by access from the user terminal (S601), and a top screen (not shown) including a predetermined user interface is transmitted to the user terminal. The user interface may be, for example, a button for displaying a winning result screen, a button for starting a camera for taking a receipt, a button for starting a predetermined SNS, or the like. When the user taps the shooting button on the terminal, a receipt image is acquired (S602). If the receipt image is subjected to OCR processing and cannot be read sufficiently due to shooting conditions or the like (S603), repeated shooting may be instructed. When the receipt image data is acquired (S603), it can be determined whether the receipt satisfies the basic condition (S604). For example, if the same (overlapping) receipt image is transmitted multiple times, such receipts clearly do not satisfy the prize application conditions, and it is not appropriate to repeatedly store them in the receipt DB 24 as purchase data. When the determination is made, an error is notified to the user terminal (S605), and the process can be terminated.

  To check for duplicates, for example, all or part of the receipt image data telephone number (c), receipt issue date / minute / second (d), total amount (e), and number of items (f) has already been applied. If it is considered that the receipt matches, it can be determined that it overlaps. “Considered to be a match” may be an exact match or a partial match as long as it meets the objective of eliminating duplicates.

  Subsequently, it is determined whether or not a prize can be applied (S606). For example, based on the date (d) of the receipt image data, it is determined whether or not the application period condition is satisfied, and whether the receipt is a conditional distribution (store) based on the telephone number (c) of the receipt image data. Is done. If the condition is satisfied, the total read amount of the receipt image data to be selected is displayed on the user terminal (S608). In response to the confirmation input or correction input by the user, the total amount or the like is stored in the sweepstakes DB 25 (S609). By making the user confirm, even if there is an OCR misread for the lottery target number string, the lottery can be performed without reducing the user satisfaction. Note that the lottery object is not limited to the total amount, and may be any number string, character string, or the like that can be read from the receipt image. If the application conditions for the sweepstakes are not satisfied, the user can be notified of this (S607).

  Further, processing for storing purchase data is performed. Appropriate sweepstakes settings (application period, etc.) can be valid purchase data that meet the sweepstakes requirements. In addition, data that does not satisfy the application conditions for the sweepstakes may be valid as purchase data. For this reason, the appropriate product master DB 21 is referred to (S610), and it is determined whether or not the product can be specified (S611). When specified, it is stored in the receipt DB 24 as purchase data (S615). If not specified, a barcode reader screen for prompting a user operation and / or a shooting screen for shooting a product is transmitted (S612). A barcode or a product image can be acquired by the user holding the camera of the terminal over the barcode and / or tapping a shooting button (S613). The user points are calculated according to the presence / absence of the barcode / product image acquired by the user operation, and to some extent, and stored in the user DB 22 (S614). Preferably, the points can be exchanged for money or goods, and a system in which points can be acquired by a simple user operation regardless of whether or not the prizes are won can make the application easier to attract users. As described above, purchase data corresponding to the specification or non-specification of the product is stored in the receipt DB 24 (S615), and if there are additional receipts, the processing of S602 to S615 can be repeated.

  According to the receipt DB 24 according to the present invention, purchase data for each receipt can be acquired using the receipt ID as a key, and can be used for market basket analysis or the like. Personal marketing is possible using the user ID and user attributes as keys. It is possible to grasp in real time across distribution (retailers) etc. what kind of products (services) are often purchased now using receipt receipt product names and read receipt product names as keys. In addition, the consumption trend of a specific product can be grasped in detail using the unique code as a key.

  Furthermore, by using a category DB (not shown) for associating a unique code with a category, analysis using the category as a key is also possible.

  The purchase data in the receipt DB 24 is provided in the form of, for example, a CSV file, and even if it is a manufacturer or an advertising agency, the purchase data can be acquired without requiring the introduction of a specific system (POS system or the like). it can. According to the present invention, it is possible to contribute to the construction of a new marketing platform.

  Those skilled in the art will appreciate that many different modifications are possible without departing from the spirit and aspects of the invention. Accordingly, it goes without saying that the embodiments of the present invention are merely examples, and do not limit the scope of the present invention.

DESCRIPTION OF SYMBOLS 10 Receipt prize Web / Ap server 11 Network interface 12 Receipt image acquisition part 13 Data reading part 14 Prize information management part 15 Purchasing information management part 16 User information management part 17 Data processing part 18 Storage interface 20 Database server 21 Goods master DB
22 User DB
23 Store Master DB
24 Receipt DB
25 Sweepstakes DB

Claims (6)

A receipt prize server that communicates with a user terminal and provides a receipt prize service,
A receipt image acquisition unit for acquiring a receipt image obtained by photographing a paper receipt;
A receipt image data reading unit for reading receipt image data from the receipt image;
A prize information management unit for managing information related to a prize based on at least the receipt image data;
See containing and a purchase information management unit for managing information related to the purchase on the basis of at least the receipt image data,
The receipt image data includes at least a product name read from the receipt image;
In the purchasing information management unit, it is determined whether or not the read product name matches the receipt notation product name in the product master database, and when it is determined that they do not match, the product name of the read product name A receipt prize server, wherein a screen for reading a barcode is transmitted to the user terminal . And a user information management unit for managing information for each user,
The receipt prize server according to claim 1 , wherein when the barcode is read by a user operation, points are calculated in the user information management unit. Prior SL purchase information management unit, the reading merchandise name, it is determined whether or not matching the receipt representation trade name of the product master database, if not a complete match but is determined to match, was the reading The receipt prize server according to claim 1, wherein the product is corrected to a receipt notation product name determined to match the product name. A method of providing a receipt prize service suitable for collecting purchase information,
Acquiring a receipt image obtained by photographing a paper receipt from a user terminal ;
Reading receipt image data from the receipt image;
Determining whether or not to apply for a prize based on the receipt image data, and accepting the application when it is possible;
Obtaining purchase information from the receipt image data received at least the application,
The receipt image data includes at least a product name read from the receipt image;
The step of obtaining the purchase information includes determining whether or not the read product name matches a receipt notation product name in a product master database. Transmitting to the user terminal a screen for reading a barcode of a name product . The product master database is generated for each retailer,
The receipt image data includes at least a telephone number;
5. The method of claim 4 , wherein when a retailer is identified based on the telephone number, a product master database of the identified retailer is referenced. A program for causing a computer of a server device to execute the method according to claim 4 or 5 .

Images