POS system and POS system operation method

Abstract

This enables efficient detection and appropriate suppression of suspected shoplifting incidents. [Solution] The POS system 100 includes a sensor 106 that observes predetermined events related to customers in a store, a processor 104 that performs a process of applying the observation results obtained from the sensor 106 to an algorithm 112 for detecting suspicious behavior to detect suspicious behavior by the customer, and a process of notifying at least one of the store staff or the customer of the detection result or a predetermined notification content corresponding to the result.

Description

Technical Field

[0001] The present invention relates to a POS system and a method for operating a POS system, and more particularly to a technology that enables efficient detection and appropriate suppression of events suspected of shoplifting.

Background Art

[0002] The number of apprehended shoplifting cases, which had been on a long-term decreasing trend, has recently turned upward due to the influence of social and economic fluctuations, etc. In particular, in retail stores such as supermarkets and convenience stores, shoplifting damage is likely to increase due to the high frequency of customer visits and the large number of product items. Such shoplifting can lead to management problems in the retail store depending on the magnitude and frequency of the damage.

[0003] Therefore, the following technologies have been proposed as conventional technologies focusing on the viewpoint of preventing shoplifting in such POS systems. For example, when a specific operator is absent or when no input operation has been performed for a certain period of time, restrictions are imposed on the input operation in the POS terminal, making it impossible for anyone other than the specific operator to perform an input operation, and a security device for a POS terminal (see Patent Document 1) that prevents fraud and theft has been proposed.

[0004] This device is characterized by comprising a detection means that outputs a detection signal when an operation by the data input unit for outputting various data to the POS main device is interrupted in a POS terminal having a data input unit, a restriction means that imposes a restriction on the input operation in the POS terminal when a detection signal is output from the detection means, and a warning means that gives a warning based on the detection signal by the detection means.

[0005] Also, a shoplifting prevention system (see Patent Document 2) that can check for shoplifting behavior with respect to products sold at each store has been proposed.

[0006] This system is a shoplifting prevention system comprising: products with machine-readable unique ID information attached; a wireless communication device for reading the ID information non-contactually; and an information processing device for receiving the ID information from the wireless communication device. The wireless communication device is installed in a location where the ID information attached to each product can be read non-contactually, for each product that is paid for at the register and each product that is taken out of the store. The information processing device is characterized by comprising: a product information database in which ID information is registered in advance; means for receiving the ID information from each wireless communication device; registration means for registering the ID information received from the wireless communication device installed at the register in the product information database; and determination means for determining whether the received ID information, when received from the wireless communication device installed at the exit of the store, is already registered in the product information database and has already been received from the wireless communication device installed at the register.

[0007] Furthermore, systems have been proposed to prevent shoplifting and protect customer purchase information (see Patent Document 3).

[0008] This system is a product settlement system that performs settlement processing for a product having a wireless tag having a capacitor, a rectifier circuit, a memory, and a first antenna, wherein the wireless tag is configured to charge the capacitor as a DC current by the rectifier circuit when it receives radio waves from the outside and an induced current generated in the first antenna is received, and provisional settlement information can be stored in the memory, and which communicates data related to the product stored in the wireless tag, wherein the product settlement system comprises a wireless tag reader / writer, a second antenna connected to the wireless tag reader / writer, and an alarm device, wherein the wireless tag reader / writer communicates the data with the wireless tag via the second antenna, the product settlement system activates the alarm device when it recognizes that the provisional settlement information is in an unsettled state, and reduces the sensitivity of the wireless tag when it recognizes that the provisional settlement information has been settled.

[0009] Furthermore, automated product accounting systems and methods (see Patent Document 4) have been proposed that do not use barcodes or RFID (electronic tags) to identify and specify purchased products, and that perform accounting processing simply and reliably.

[0010] This technology comprises an entry detection device installed at the entrance of a store to detect when a shopper enters, a smartphone placed in the shopping basket used by the shopper who has entered the store, an exit detection device installed at the exit of the store, and a group of automated product payment servers that constitute a cloud operating the automated product payment system between the entry detection device, the smartphone, and the exit detection device. The automated product payment servers are equipped with at least a CPU and a GPGPU, and the smartphone detects whether an item has been placed in the shopping basket from top to bottom or removed from bottom to top. The device is characterized by having: a product detection unit that detects whether an item has been taken out and identifies the item taken out by AI image recognition using the GPGPU; a purchase count unit that counts the taking out of the item; a purchase details sheet creation unit that displays a details sheet of the purchased items and a payment button; a payment detection unit that detects when the payment button related to the purchase details created by the purchase details sheet creation unit is pressed; a payment completion detection unit that detects whether or not payment has been completed; and an unpaid signal transmission unit that transmits an unpaid radio wave signal to the exit detection device if payment completion has not been detected by the payment completion detection unit. [Prior art documents] [Patent Documents]

[0011] [Patent Document 1] Japanese Patent Publication No. 2001-307225 [Patent Document 2] Japanese Patent Publication No. 2004-240767 [Patent Document 3] Japanese Patent Publication No. 2023-140985 [Overview of the project] [Problems that the invention aims to solve]

[0012] All of the conventional technologies are expected to have a reasonable effect, such as efficiently detecting shoplifting and missed scans. On the other hand, the perspective of supporting subsequent follow-up actions based on the results of shoplifting detection and estimation is still insufficient. It is not enough to simply detect shoplifting mechanically; from the perspective of store operations and customer service, it is important to respond appropriately to customers' psychology, such as guiding them to refrain from shoplifting, to return items to the shelves even if they have shoplifted, or to make it easier for them to regulate their own behavior in the future.

[0013] Therefore, the objective of the present invention is to provide a technology that enables the efficient detection and appropriate suppression of events suspected of shoplifting. [Means for solving the problem]

[0014] The POS system of the present invention, which solves the above problems, is characterized by comprising: a sensor that observes predetermined events relating to customers in a store; a processor that performs the following: applying the observation results obtained from the sensor to an algorithm for detecting suspicious behavior to detect suspicious behavior by the customer; and notifying at least one of the store staff or the customer of the detection result or predetermined notification content corresponding to the result. This makes it possible to efficiently detect and appropriately suppress events that may be suspected of shoplifting.

[0015] Furthermore, in this POS system, the processor may, in response to the notification or in response to an action taken by a staff member who has received the notification, execute a process to stop or restrict subsequent procedures concerning the customer to only accepting staff input. This makes it possible to deal with customers who have engaged in suspicious behavior suspected of shoplifting with a moderate response that evokes a general malfunction, rather than an explicit accusation or warning.

[0016] Furthermore, in this POS system, the processor may, when detecting suspicious behavior, determine the time interval between scanning operations for each product to be registered by the customer based on the observation results of the sensor, and if the time interval does not meet the criteria, it may be detected as suspicious behavior by the customer. This makes it possible to efficiently detect any actions taken between scanning operations that are intended to shoplift.

[0017] Furthermore, in this POS system, the processor may, when detecting suspicious behavior, identify the number of items and the unit price of each item registered by the customer based on the observation results of the sensor, and if the number of items is above a certain standard and the unit price is below a certain standard, it may be detected as suspicious behavior by the customer. This would make it possible to efficiently detect suspicious behavior for shoplifting, such as registering only the price for one item even though it is a set of multiple items.

[0018] Furthermore, in this POS system, the processor may, when detecting suspicious behavior, acquire video footage of the customer during the scanning operation of the product to be registered by the customer from the sensor, input the video footage into the learning model, which is the algorithm, and determine whether the customer's movements or facial expressions constitute suspicious behavior. This makes it possible to detect suspicious behavior based on the customer's movements and facial expressions with high accuracy and efficiency.

[0019] Furthermore, in this POS system, when detecting suspicious behavior, the processor may acquire an image of the bottom of the shopping cart used by the customer from the camera sensor, input the image into the learning model algorithm, and determine whether or not there are large items at the bottom of the cart. This makes it possible to accurately detect suspicious behavior corresponding to shoplifting, such as large items that customers tend to place at the bottom of their carts, even if they are not registered at the register, based on image analysis. In this case, the POS system can appropriately refer to the data of the customer's product registration at the register.

[0020] Furthermore, in this POS system, the processor may acquire images of the underside of the cart from the underside of the checkout counter through which the cart passes, or from a camera installed on the underside of the cart. This makes it possible to accurately acquire images of the products placed on the underside of the cart.

[0021] Furthermore, in this POS system, the processor may, when processing the notification, further perform a process to highlight the output content related to the individual product that also exists as a large product among the products to be registered, or the large product, on at least one of the screens and receipts for at least one of the staff or the customer, based on the information of the product, which is the result of the customer's scanning operation for the product to be registered. This would make it possible to clearly indicate to the staff or the customer, for example, that a large product placed in the customer's cart is not registered (only the individual product is registered). In this case, it is assumed that the POS system can refer to the results of a learning model or the like that has determined whether or not a large product is placed at the bottom of the cart, or that the staff can visually check whether or not a large product is placed at the bottom of the cart.

[0022] Also, in this POS system, when processing the notification, the processor may further execute a process of highlighting output content regarding high-value products among the products to be registered, based on the information of the product, which is the result of the scanning operation on the products to be registered by the customer, on at least one of the screen and the receipt for at least one of the staff or the customer. According to this, for example, it becomes possible to clearly suggest to the staff or the customer a situation where a high-value product placed in the customer's cart or shopping basket is not registered (or a high-value product registered by the customer is not placed in the cart or shopping basket). In this case, it is assumed that the POS system can refer to the result of determining the presence or absence of high-value products in the cart or shopping basket using a learning model or the like, or the staff can visually check the presence or absence of high-value products in the cart or shopping basket.

[0023] Also, in this POS system, when sending the notification, the processor may notify the customer of dummy information indicating the occurrence of an error in the device, and notify the staff of the result of detecting the suspicious behavior of the customer or a predetermined notification content corresponding to the result. According to this, while sending a gentle notification to the customer with suspicious behavior, it becomes possible to prompt the staff to take countermeasures against the suspicious behavior of the customer.

[0024] Also, in this POS system, the processor may execute the notification at any timing from the scanning operation of the product to be registered by the customer to the settlement operation of the product, and in each of the notifications, output information on the registered points of the products scanned up to that timing and information prompting the customer to check for registration omissions. According to this, for example, it becomes possible to at least explicitly show the number of products registered so far to the customer at timings such as every certain period of time or every registration of a certain number of products. Performing such explicit indication may correspond to giving a reasonable warning to malicious customers who intentionally avoid the scanning operation, so effects such as suppressing suspicious actions associated with shoplifting can be expected.

[0025] Also, in this POS system, when the processor scans a discount sticker attached to the product during the scanning operation of the product to be registered by the customer, the processor may issue the notification in a predetermined specific form. According to this, for example, for products that are not eligible for discounts, by attaching a discount sticker attached to other products, it is possible to give a reasonable warning by specifically indicating that the product has been registered as a discount-eligible product for suspicious actions such as disguising a discounted product.

[0026] Also, in this POS system, the processor may output different sounds for at least any one of the types or timings of any of the operations from the scanning operation of the product to be registered by the customer to the settlement operation of the product and for each cash register targeted by the operation, as the notification. According to this, regarding various operations by the customer, it will be clearly known by sound that the POS system has recognized them and that it has spread to the surroundings.

[0027] Furthermore, in this POS system, the processor may, upon notification, acquire an image from the sensor that includes the customer whose suspicious behavior has been detected, mask at least the face portion of the person's image in the image, and then publish the image on a predetermined website that is accessible to at least the customer. This makes it possible to publish an image showing the suspicious behavior in a way that does not allow for the identification of the individual customer who performed the suspicious behavior. This means that the customer will be able to see the image and realize that the store has suspicions about their actions.

[0028] Furthermore, in this POS system, the processor may acquire images from the sensor of at least one of multiple parts of the customer or multiple parts of the cash register where the customer is working, as images to be published on the website, and publish the acquired images on the website. This suggests that customers who are performing suspicious actions while registering products or making payments may be photographed from various angles, and the resulting images may be published.

[0029] Furthermore, in this POS system, the processor may, upon issuing the notification, publish information on a designated website suggesting the effectiveness of anti-shoplifting measures at the store, based on the content or frequency of the detected suspicious behavior. This would create a situation where customers are more likely to recognize that stores prone to shoplifting actually have sufficient anti-shoplifting measures in place, with appropriate monitoring and response, meaning that the stores are less likely to be targeted. This could have the effect of deterring shoplifters from committing theft at those stores.

[0030] Furthermore, in this POS system, the processor may, when issuing the notification, obtain information from a predetermined system regarding the extent of shoplifting damage at the store or the results of countermeasures against such shoplifting damage, and, depending on the magnitude of the shoplifting damage or the results of countermeasures indicated by the obtained information, publish information on a predetermined website that suggests the adequacy of shoplifting countermeasures at the store. This would create a situation where, for example, in stores where shoplifting is actually frequent or where such shoplifting is dealt with strictly, customers are more likely to recognize that sufficient shoplifting countermeasures are implemented and that surveillance and other measures are carried out appropriately, that is, that the store is difficult to shoplift from. This could lead to a deterrent effect against shoplifting at the store.

[0031] Furthermore, in this POS system, the processor may further perform the following processes: acquiring identification information and issuance time information of a receipt issued at the register in conjunction with the payment of goods registered by the customer from a predetermined system; granting exit rights from a predetermined area in the store to receipts whose elapsed time from the issuance time is within a certain limit; having the sensor read the identification information of the receipt presented by the customer attempting to leave the predetermined area, and determining whether the receipt corresponding to the identification information has exit rights; and detecting the customer's exit as suspicious behavior based on the result of the determination. According to this, for example, at a gate that a customer passes through when exiting from a payment area in a store, the receipt presented by the customer can be used as a temporary medium for exit rights, which can help deter suspicious behavior.

[0032] Furthermore, in this POS system, the processor may further execute a process to detect when a customer cancels an item registered at the register, and to output a message prompting the customer to transfer the item to a designated storage device. When detecting suspicious behavior, the processor may also detect whether the cancellation operation is suspicious behavior by comparing the number of items subject to the cancellation operation with the number of items stored in the storage device. This makes it possible to allow cancellation procedures, which were previously difficult to implement in self-service registers, while ensuring sufficient countermeasures against shoplifting. Customers can easily cancel registered items themselves without being unnecessarily suspected of shoplifting. From the store's perspective, it becomes possible to allow cancellation procedures while operating a self-service register, effectively deterring shoplifting and other fraudulent activities, and reducing problems such as sales losses due to shoplifting.

[0033] Furthermore, the present invention, which solves the above problems, is characterized in that the POS system is equipped with a sensor that observes predetermined events concerning customers in a store, applies the results of the observations obtained from the sensor to an algorithm for detecting suspicious behavior, and performs the following processes: detecting the suspicious behavior of the customer; and notifying at least one of the store staff or the customer of the detection result or a predetermined notification content corresponding to the result. This makes it possible to efficiently detect and appropriately suppress events that may be suspected of shoplifting. [Effects of the Invention]

[0034] According to the present invention, it becomes possible to efficiently detect and appropriately suppress incidents that may be suspected of shoplifting. [Brief explanation of the drawing]

[0035] [Figure 1] This figure shows an example of a network configuration including a POS system in this embodiment. [Figure 2] This figure shows an example of the hardware configuration of the POS system in this embodiment. [Figure 3]This figure shows an example of the configuration of the product management database in this embodiment. [Figure 4] This figure shows an example of the configuration of the POS data database in this embodiment. [Figure 5] This figure shows an example of the configuration of the receipt management database in this embodiment. [Figure 6] This figure shows an example of the configuration of the observation results database in this embodiment. [Figure 7] This figure shows an example of the configuration of the customer management database in this embodiment. [Figure 8] This figure shows an example of the configuration of the cash register management database in this embodiment. [Figure 9] This figure shows an example of the configuration of the performance information database in this embodiment. [Figure 10] This figure shows an example of the configuration of the cancellation management DB in this embodiment. [Figure 11] This figure shows an example of the configuration of the reference table in this embodiment. [Figure 12] This figure shows an example of the configuration of the word-of-mouth table in this embodiment. [Figure 13] This figure shows an example of the configuration of the destination table in this embodiment. [Figure 14] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 15] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 16] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 17] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 18] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 19] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 20] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 21] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 22] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 23] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 24] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 25] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 26] This figure shows an example of a flowchart for the operation method of the POS system in this embodiment. [Figure 27] This figure shows an example of the output in this embodiment. [Figure 28] This figure shows an example of the output in this embodiment. [Figure 29] This figure shows an example of the output in this embodiment. [Figure 30] This figure shows an example of the output in this embodiment. [Figure 31] This figure shows an example of the output in this embodiment. [Figure 32] This figure shows an example of the output in this embodiment. [Figure 33] This figure shows an example of the output in this embodiment. [Figure 34] This figure shows an example of the output in this embodiment. [Figure 35] This figure shows an example of the output in this embodiment. [Figure 36] This figure shows an example of the output in this embodiment. [Figure 37] This figure shows an example of the output in this embodiment. [Modes for carrying out the invention]

[0036] <<Example of overall structure>> Embodiments of the present invention will be described in detail below with reference to the drawings. Figure 1 is a diagram showing an example of a network configuration including the POS system 100 of this embodiment. The POS system 100 shown in Figure 1 is a system that works in cooperation with the store's POS register 10, shopping cart 20, or user terminal 30 or cancellation box 40 to enable efficient detection and appropriate suppression of suspected shoplifting incidents in the store.

[0037] In this embodiment, a configuration in which the POS system 100 is a separate device from the POS register 10 is shown as an example, but the two may be configured as an integrated device. Alternatively, the cart 20 may be equipped with some or all of the functions and configurations of the POS register 10. In any case, the implementation configuration is not limited to whether the POS system 100, POS register 10, cart 20, and user terminal 30 or cancel box 40 are integrated or separate, or whether they share functions and configurations and cooperate with each other.

[0038] The POS system 100 described above is connected to the POS register 10, cart 20, user terminal 30, and cancel box 40 via an appropriate network 1, such as a wired or wireless LAN (Local Area Network), or short-range wireless communication using Bluetooth or infrared. Naturally, the protocol used in this communication network 1 has a communication speed at least equal to or greater than the operating speed of the POS register 10, cart 20, user terminal 30, and cancel box 40.

[0039] The POS register 10 is equipped with at least a scanner 11, a printer 12, a display 13, a speaker 14, and a light 15. The scanner 11 is a barcode reader that reads the barcode attached to the product to be registered. However, in order to support processing by cashless payment methods (electronic money, etc.), it may also be equipped with the function of reading / writing two-dimensional codes and IC chips. The printer 12 is a means of printing receipts and invoices on paper when product registration and payment are completed at the POS register 10. The display 13 is a display device that shows information to the operator in accordance with various procedures such as product registration and payment at the POS register 10. The speaker 14 is an audio device that outputs information to the operator by sound in accordance with the above procedures. The light 15 is a light-emitting device that outputs information to the operator by light in accordance with the above procedures.

[0040] Furthermore, the cart 20 is a means of transport used by customers shopping at the aforementioned store for storing and transporting goods. This cart 20 has at least two storage compartments, one above the other. The upper storage compartment opens upwards, similar to a shopping basket, allowing customers to directly place items inside. The lower storage compartment, on the other hand, is the lower storage compartment 21, where large items packaged appropriately in cardboard boxes or similar materials are placed. Examples of such large items include cases of beer containing multiple cans, large quantities of detergent, or large quantities of kitchen towels. Of course, these are just examples and not limiting.

[0041] In addition, the user terminal 30 is a form terminal used by a store attendant or a customer. As the implementation form of the user terminal 30, products such as smartphones, tablet terminals, and small notebook computers may be applicable. Store attendants and customers can recognize various information such as notifications from the POS system 100 through output devices such as the printer 12, display 13, speaker 14, and light 15 of the POS register 10, but such information may also be confirmed on the user terminal 30. In that case, it is assumed that the address of the user terminal 30 used by the attendant or customer on the network 1 (for example, an email address or an IP address, etc.) is managed by the POS register 10 and can be appropriately used as the destination information at the time of notification or the like.

[0042] In addition, when a customer performs a registration cancellation operation, that is, a cancellation operation, on the registered products at the POS register 10, the cancellation box 40 becomes a storage device for putting the products to be cancelled. This cancellation box 40 is installed, for example, at the place where the attendant is present or at the service counter. This cancellation box 40 is connected to the POS system 100 via the network 1, and appropriate sensors such as the box camera 1063 are installed. Such sensors execute processes such as photographing the products put into the storage section of the cancellation box 40 and reading the barcodes attached to the products, and upload the results, that is, the photographed data, barcode reading values, etc. to the POS system 100 at regular intervals or each time sensing is performed.

[0043] <<Configuration Example of POS System>> Here, the hardware configuration of the POS system 100 will be described based on FIG. 2. The POS system 100 of this embodiment includes at least an auxiliary storage device 101, a main storage device 103, a processor 104, a communication device 105, and a camera 106. Among these, the auxiliary storage device 101 is composed of an appropriate non-volatile memory element such as an SSD (Solid State Drive), a hard disk drive, or a ROM (Read Only Memory).

[0044] The non-volatile memory elements in the auxiliary storage device 101 store the OS 110 (Operating System) for basic control of the POS system 100, as well as programs 111 for implementing necessary functions on the OS 110, a learning model 112, a product management DB 120, a POS data DB 121, a receipt management DB 122, an observation results DB 123, a customer management DB 124, a cash register management DB 125, a performance information DB 126, a cancellation management DB 127, a reference table 128, a word-of-mouth table 129, and a publication destination table 130. Details of these databases and tables will be described later.

[0045] Furthermore, the main memory 103 is the read destination during the execution of the program 111 described above, and becomes the actual entity on which the functions of the POS system 100 are implemented. The main memory 103 is composed of volatile memory elements such as RAM (Random Access Memory), but it may also be composed of non-volatile memory elements.

[0046] Furthermore, the processor 104 is an arithmetic unit that reads the above-mentioned program 111, which is stored in the auxiliary storage device 101, into the main memory device 103 and executes it, performs overall control of the device itself, and performs various judgments, calculations, and control processes necessary for the POS system of the present invention.

[0047] Furthermore, the communication device 105 is a network interface card that handles communication processing with the POS register 10, cart 20, user terminal 30, cancel box 40, etc., via the network 1. This communication device 105 will be implemented according to the protocol of network 1.

[0048] Furthermore, camera 106 is a type of sensor in the POS system 100 of this embodiment for detecting suspicious customer behavior. Camera 106 includes a cash register camera 1061, a cart camera 1062, and a box camera 1063.

[0049] Of these, the register camera 1061 is a digital video camera that captures customers and their scanning actions near the location where the scanner 11 of the POS register 10 is installed. The cart camera 1062 is a digital video camera that primarily captures the lower storage compartment 21 of the cart 20. The box camera 1063 is a digital video camera that captures the storage compartment of the cancellation box 40 as its target area, and captures the items to be canceled that have been placed in that compartment. It is even preferable if the box camera 1063 also has a barcode reading function for the products.

[0050] These cash register cameras 1061, cart cameras 1062, and box cameras 1063 are connected to the communication device 105 of the POS system 100 via, for example, a wireless LAN (network 1), and upload captured data, etc., to the POS system 100 at regular intervals or each time a photo is taken.

[0051] [Product management DB] Next, the database and tables used by the POS system 100 of this embodiment will be described. Figure 3 shows an example of the product management DB 120 in this embodiment. This product management DB 120 is a database that stores data for each product sold in the store.

[0052] As shown in Figure 3, this product management DB120 is a data set where each value such as the product's genre, product name, form, size, and price is associated with the product ID as the key. Of these, "form" indicates whether it is sold individually or as part of a set. "Size" indicates the package size of the product, and is divided into values ​​from the smallest "S" to the largest "LL". Products with a size of "L" or larger can be classified as large products.

[0053] [POS Data Database] Next, the POS data DB121 in this embodiment will be described. Figure 4 shows an example of the POS data DB121 in this embodiment. This POS data DB121 is a database that stores product sales data acquired and managed by the store's POS register 10.

[0054] As shown in Figure 4, the POS data DB121 is a data set in which the date and time and register identification information are used as keys, and various values ​​such as the cart 20 used by the customer, discount stickers, the ID of the product registered and paid for at that register, product name, quantity, unit price, and total after discount are associated with each other. Each record in the POS data DB121 mainly contains information recognized from the product barcode during the reading operation by the scanner 11 of the POS register 10.

[0055] [Receipt Management Database] Next, the receipt management DB 122 in this embodiment will be described. Figure 5 shows an example of the receipt management DB 122 in this embodiment. This receipt management DB 122 is a database that stores data related to receipts output by the store's POS register 10.

[0056] As shown in Figure 5, the receipt management DB 122 is a data set in which the receipt ID and exit authority are associated with the receipt issue date and time and the register identification information as keys. Of these, "exit authority" is granted or revoked according to the criteria of the elapsed time since the receipt was issued, as defined in the criteria table 128 described later. Each record in this receipt management DB 122 mainly contains information acquired by the POS register 10 in conjunction with the receipt printing operation by the printer 12 of the POS register 10 and distributed to the POS system 100.

[0057] [Observation Results Database] Next, the observation results DB123 in this embodiment will be described. Figure 6 shows an example of the observation results DB123 in this embodiment. This observation results DB123 is a database that stores image data (still images or videos) of customers, etc., captured by sensors, i.e., cameras 106, installed in the store's POS register 10 or shopping cart 20. Of course, here, the camera 106 is used as an example of a sensor, but sensors that detect the presence of people or objects using infrared or ultrasonic waves, or scanners 11 in the POS register 10 that acquire various information on products to be registered, can also be included in the concept of sensors.

[0058] As shown in Figure 6, the observation results DB123 is a data set in which the date and time of observation and the sensor ID are used as keys, and various values ​​such as the type of sensor, observation location, and observation data are associated with each other. Of these, "observation location" corresponds to the area being photographed by camera 106. The "observation data" is in a format that depends on the type of sensor; for example, if the sensor type is "camera," it will be image data.

[0059] [Customer management DB] Next, the customer management DB 124 in this embodiment will be described. Figure 7 shows an example of the customer management DB 124 in this embodiment. This customer management DB 124 is a database that stores information on each customer of the store.

[0060] As shown in Figure 7, the customer management DB124 is a data set where each value, such as the customer's name, contact information, and purchase history, is associated with the customer ID as the key. Of these, "contact information" is the address (such as an email address or IP address) set for the user terminal 30 used by the customer on network 1. "Purchase history" is historical data about the customer's past purchasing behavior, and mainly includes information shown in POS data distributed from the POS register 10, for example.

[0061] [Cash register management database] Next, the cash register management DB 125 in this embodiment will be described. Figure 8 shows an example of the cash register management DB 125 in this embodiment. This cash register management DB 125 is a database that stores data related to the store's POS register 10.

[0062] As shown in Figure 8, the register management DB 125 is a data set in which various values, such as the identification information of sensors installed at a register, are associated with the register's identification information as the key. Among these, the "installed sensor" is the identification information of the register camera 1061, which is installed near the location of the scanner 11 at the register. Alternatively, it could be the identification information of the cart camera 1062, which is located under the register counter and targets the lower storage compartment 21 of the cart 20. By managing register information with this register management DB 125, it becomes possible to identify which sensors are installed at which registers.

[0063] [Performance Information Database] Next, the performance information DB126 in this embodiment will be described. Figure 9 shows an example of the performance information DB126 in this embodiment. This performance information DB126 is a database that stores data on suspicious customer behavior detected at each store, shoplifting incidents, or the results of shoplifting prevention measures.

[0064] As shown in Figure 9, this performance information DB126 is a data set in which various values ​​are associated with the POS system 100 regarding the store, such as past suspicious behavior detected by the POS system 100, shoplifting incidents, and shoplifting prevention measures, using the store identification information as the key. Of these, "suspicious behavior" includes information such as the date, time, and content of suspicious behavior detected by the POS system 100 regarding customers. "Shoplifting incidents" is information entered, for example, from the store system or the store operator's user terminal 30, and includes information such as the number of shoplifting incidents and the amount of damage. "Shoplifting prevention measures" includes information such as the number, range, and density of sensors (cameras 106) installed in the store. [Cancellation Management Database] Next, the cancellation management DB 127 in this embodiment will be described. Figure 10 shows an example of the cancellation management DB 127 in this embodiment. This cancellation management DB 127 is a database that stores data on events where a customer has performed a cancellation action at the store's POS register 10.

[0065] As shown in Figure 10, the cancellation management DB127 is a data set in which the date and time of the cancellation operation and the register identification information are used as keys, and each value, such as the product ID, product name, quantity, price, and box storage, is associated with the product subject to the cancellation event. Of these, "box storage" indicates whether or not the target product was stored in the cancellation box 40 when the customer performed a cancellation operation at the POS register 10, that is, whether it is "not stored" or "stored". These values ​​will be "stored" if the box camera 1063 of the cancellation box 40 was able to acquire and recognize the image or barcode value of the target product, and "not stored" in all other situations.

[0066] [Reference Table] Next, the reference table 128 in this embodiment will be described. Figure 11 shows an example of the reference table 128 in this embodiment. This reference table 128 is a table that stores information on the criteria that the POS system 100 uses to compare the sensor observation results when detecting suspicious behavior.

[0067] As shown in Figure 11, this criteria table 128 is a data set in which each event and a reference value are associated. The "event" indicates the type of suspicious behavior, and is a value that the POS system 100 has pre-recognized to adopt for each sensor observation result. The "reference value" is the criterion for determining whether the value indicated by the observation result corresponds to suspicious behavior for that event. For example, for event "Suspicious behavior 1 (illegal scan)", a "reference value" is defined that determines that a scan operation includes suspicious behavior when the interval between scan operations for each product to be registered, as shown by the POS data obtained for a certain customer, is "3 seconds or more".

[0068] [Customer Review Table] Next, the word-of-mouth table 129 in this embodiment will be described. Figure 12 shows an example of the word-of-mouth table 129 in this embodiment. This word-of-mouth table 129 is a table in which word-of-mouth content is predefined in order to widely publicize the level of shoplifting prevention measures at a store.

[0069] As shown in Figure 12, this review table 129 is a data set in which each value of an event and a review are associated. The "event" indicates the event that the review is about, for example, "improved shoplifting prevention measures." The "review" is text data describing what should be posted on the website regarding such an event. For example, text data such as "They have installed state-of-the-art cash registers, so checkout is quick and easy. You don't see the staff very often, but it seems like there are cameras installed in various places..." is stored in the record.

[0070] [Publication destination table] Next, the destination table 130 in this embodiment will be described. Figure 13 shows an example of the destination table 130 in this embodiment. This destination table 130 is a table that stores information about the websites that publish the above-mentioned reviews and customer photos.

[0071] As shown in Figure 13, Table 130 is a data set where the values ​​for "Publication Target" and "Publication Destination URL" are associated. "Publication Target" indicates what data will be published, such as customer photos or reviews. "Publication Destination URL" indicates the URL of the website on Network 1 where these "Publication Target" values ​​will be published. More specifically, this website could be a review posting site for stores or products, or a review posting page for stores or facilities on a map service published on Network 1.

[0072] <<Operation Method of POS System: Main Flow>> The operation method of the POS system in the present embodiment will be described below based on the drawings. Various operations corresponding to the POS system operation method described below are realized by a program 111 that is read out and executed by the POS system 100 in a memory or the like. The program 111 is composed of codes for performing various operations described below. However, it is assumed that the POS system 100 can appropriately cooperate with external devices such as the POS register 10, the cart 20, the user terminal 30, and the cancellation box 40 via the network 1 and perform necessary processing.

[0073] FIG. 14 is a diagram showing an example of a flow of the operation method of the POS system in the present embodiment. Here, it is assumed that a certain customer visits a retail store such as a supermarket and is performing a scanning operation on the products to be purchased at the store, that is, operating a self-service type of POS register 10. Note that the POS register 10 may be of a so-called full self-service type or a semi-self-service type.

[0074] First, the processor 104 of the POS system 100 acquires data of a photographed image, which is an observation result, from a camera 106 (the register camera 1061 or the cart camera 1062), which is a sensor, regarding a customer who is in the process of purchasing at the POS register 10 (s1). Further, the processor 104 of the POS system 100 inputs the data of the photographed image obtained in s1 into the learning model 112 and detects suspicious behavior (s2). The learning model 112 is, for example, a learning model of deep learning, and is a model for detecting suspicious behavior obtained by advancing learning using, as teacher data, set data of a photographed image whose subject is a customer and a flag indicating whether or not the customer shown in the photographed image is performing suspicious behavior.

[0075] Furthermore, when detecting suspicious behavior using this learning model 112, for example, as shown in screen G12 of Figure 29, the detection result of whether or not the customer's captured image is suspicious may be displayed on the display 13 of the POS register 10 or on the user terminal 30. This screen G12 includes a group of images G121 showing the customer scanning products at the POS register 10 from multiple angles, and a judgment result G122 indicating the detection result of suspicious behavior for each image constituting the group of images G121 as "OK" or "NG". In the example in Figure 29, of the three images constituting the group of images G121, only the top image does not show detected suspicious behavior, while the remaining middle and bottom images show detected suspicious behavior.

[0076] The processor 104 of the POS system 100 determines whether suspicious behavior has been detected with respect to the customer in question based on the above-mentioned suspicious behavior detection process (s3). If no suspicious behavior is detected as a result of this determination (s3:N), the processor 104 terminates this flow. On the other hand, if suspicious behavior is detected as a result of this determination (s3:Y), the processor 104 notifies, for example, the POS register 10 or the user terminal 30 of the suspicious behavior or the notification content corresponding to that suspicious behavior (s4).

[0077] The notification screen G10 shown in Figure 27 is a screen viewed by the customer and does not explicitly state that suspicious behavior has been detected, but rather suggests that a system error has occurred. On the other hand, the notification screen G11 shown in Figure 28 is a screen viewed by store staff and indicates that suspicious behavior has been detected and that staff should take further action.

[0078] Next, the processor 104 notifies the POS register 10 of a control instruction that only operations by the store staff will be accepted (s5). This instruction includes, for example, instructions to gray out various operation buttons on the customer-side screen of the POS register 10's display 13 so that they cannot be operated, and instructions to display a button that should be operated when suspicious behavior is detected (e.g., a security guard call button) on the staff-side screen.

[0079] Furthermore, processor 104 determines whether the series of transactions has been completed through operations by staff, for example, by pressing buttons such as "End Transaction" or "Interrupt" (s6). If the result of this determination is that the transaction has not been completed (s6:N), processor 104 transitions the process to s5. On the other hand, if the result of this determination is that the transaction has been completed (s6:Y), processor 104 terminates this flow.

[0080] <Flowchart for detecting suspicious behavior (scan operation interval)> Next, we will explain the case in which suspicious behavior is detected based on the interval between scan operations, using the flow shown in Figure 15. In this case, the processor 104 refers to the POS data in the POS data DB 121 and identifies the time interval between scan operations during a customer's shopping opportunity (s10). This identification is performed, for example, by calculating the time difference (average across records) between the date and time information in the POS data.

[0081] Furthermore, the processor 104 compares the time duration identified in s10 with the criterion value "scan operation interval of 3 seconds or more" for "Suspicious behavior 1 (illegal scan)" in the criterion table 128 (s11). Based on the result of the comparison in s11, the processor 104 determines whether the time duration between the scan operations meets the criterion of 3 seconds or more (s12).

[0082] If the result of this determination does not meet the criteria (s12:N), processor 104 terminates this flow. On the other hand, if the result of the above determination does meet the criteria (s12:Y), processor 104 detects suspicious activity in the customer's scanning operation (s13) and terminates this flow.

[0083] <Flowchart for detecting suspicious behavior (number of items and unit price)> Next, we will explain the case in which suspicious behavior is detected based on the number of items and the unit price of each item, using the flow chart in Figure 16. In this case, the processor 104 refers to the POS data in the POS data DB 121 and identifies the number of items and the unit price of each item during a customer's shopping opportunity (s20).

[0084] Furthermore, processor 104 compares the number of items and the unit price identified in s20 against the criteria value "10 or more items, unit price of 300 yen or less" for "Suspicious behavior 2 (illegal scan)" in the criteria table 128 (s21). If the result of this determination does not meet the criteria (s21:N), processor 104 terminates this flow.

[0085] On the other hand, if the above determination results in compliance with the criteria (s21:Y), the processor 104 detects suspicious activity in the customer's scanning operation (s23) and terminates this flow.

[0086] <Flowchart for detecting suspicious behavior (customer's facial expressions, movements)> Next, we will explain the case in which suspicious behavior is detected based on the customer's facial expressions and movements, using the flow chart in Figure 17. In this case, the processor 104 acquires video footage of the customer during the scanning operation from, for example, the cash register camera 1061 among the cameras 106 (s30).

[0087] Furthermore, the processor 104 inputs the customer video obtained in s30 into the learning model 112 and determines whether the customer's movements or facial expressions constitute suspicious behavior (s31). If no suspicious behavior is found as a result of this determination (s32:N), the processor 104 terminates this flow. On the other hand, if suspicious behavior is found as a result of the above determination (s32:Y), the processor 104 masks the subject in the customer video, i.e., the customer's face (s33). The processor 104 then publishes the customer video with the face masked on a website (s34) and terminates this flow. The website is the site published at the URL specified in the publication destination table 130 for "customer's captured images".

[0088] In this case, the processor 104 uploads and publishes the group of captured images G181 (masked) showing the customer's appearance, as shown in screen G18 in Figure 35, for example, to the value of the "Publication URL" defined in the publication destination table 130 for the "Captured Images of Customers" that are to be published.

[0089] <Flowchart for detecting suspicious behavior (large items)> Next, we will explain the case in which suspicious behavior is detected based on large items in the lower storage compartment 21 of the cart 20, using the flowchart in Figure 18. In this case, the processor 104 acquires an image of the lower storage compartment 21 of the cart 20 from, for example, the cart camera 1062 of the cameras 106 (s40).

[0090] Furthermore, regarding the image showing the state in which large items are placed in the lower storage compartment 21 of the cart 20, for example, the screen G13 shown in Figure 30 may be displayed on the display 13 of the POS register 10 or the user terminal 30 to clearly show it to the store staff. In the example of screen G13 in Figure 30, it includes a captured image G131 of the cart 20 and the judgment result G132 of the judgment (s41, s42) described later.

[0091] Furthermore, the processor 104 inputs the image obtained in s40 into the learning model 112 to determine whether or not there are large items in the lower storage compartment 21 of the cart 20 (s41). If the presence of large items is not identified as a result of this determination (s42:N), the processor 104 terminates this flow. On the other hand, if the presence of large items is identified as a result of the above determination (s42:Y), the processor 104 refers to the POS data for the items in the cart 20 in the POS data DB 121 to determine whether or not large items are registered (s43). Large items are identified as such if the "Size" column in the product management DB 120 has a value such as "L" or "LL".

[0092] Furthermore, based on the result of s43, processor 104 determines whether there is a registration of large items in the target customer's shopping opportunity (s44). If the result of this determination is that there is a registration of large items (s44:Y), processor 104 terminates this flow. On the other hand, if the result of the above determination is that there is no registration of large items (s44:N), processor 104 recognizes the existence of suspicious behavior in relation to the customer's shopping opportunity (s45) and terminates this flow.

[0093] <Flowchart for detecting suspicious activity (registration of large items and related individual items)> Next, we will explain the case in which suspicious behavior is detected based on the registration of individual items related to large items, using the flow shown in Figure 19. In this case, the processor 104 refers to the POS data in the POS data DB 121 and identifies whether there is a registration of individual items that also exist in the form of large items for a particular shopping opportunity of a particular customer (s50).

[0094] This identification can be performed, for example, by searching the product management DB120 using the "product ID" indicated by the POS data obtained regarding the above-mentioned shopping opportunity (e.g., for a single bottle of beer). If products with the same product category (e.g., alcohol), a common product name (e.g., beer), and a "set" "format" are found, then it can be identified that there is a registered single product that also exists as a "set" product, which is a "large product."

[0095] Next, the processor 104 determines whether or not individual items are registered based on the above specific result (s51). If the result of this determination is that no individual items are registered (s51:N), the processor 104 terminates this flow. On the other hand, if the result of the above determination is that individual items are registered (s51:Y), the processor 104 prints the individual items, which also exist in the form of large items, in an emphasized format on the screen or receipt (s52), and terminates this flow.

[0096] In this case, the processor 104 instructs the printer 12 of the POS register 10 to perform emphasis such as bolding and shading on the printing of a single item, such as "one bottle of beer," which also includes larger items, as shown in receipt G15 in Figure 32.

[0097] <Notification (Registration of High-Value Products) Flow> Next, regarding notifications based on the detection results of suspicious behavior, the notification based on the registration of high-priced items will be explained using the flow shown in Figure 20. In this case, the processor 104 refers to the POS data in the POS data DB 121 and identifies whether or not high-priced items have been registered for a particular shopping opportunity of a particular customer (s60). These high-priced items can be identified, for example, by comparing the criteria value "item price of 5,000 yen or more" for "suspicious behavior 4 (high-priced items)" in the criteria table 128 with the amount of each item indicated in the POS data.

[0098] Next, the processor 104, based on the result of s60, determines whether or not high-value items are registered (s61). If the result of this determination is that no high-value items are registered (s61:N), the processor 104 terminates this flow. On the other hand, if the result of the above determination is that high-value items are registered (s61:Y), the processor 104 prints the high-value items in an emphasized format on the screen or receipt (s62) and terminates this flow.

[0099] In this case, the processor 104 instructs the printer 12 of the POS register 10 to highlight the high-priced item, "Sirloin Steak 250g," by using bolding and shading, as shown in receipt G14 in Figure 31. <Notification flow (registered points at each timing)> Next, regarding notifications based on the registered points at each timing, in response to the detection of suspicious behavior, the notification process will be explained based on the flow shown in Figure 21. In this case, the processor 104 senses the arrival of predetermined timings (for example, every 3 minutes) using the computer's clock function, etc. (s70). If the processor 104 senses the arrival of the predetermined timing as a result of this sensing (s70:Y), it transitions the process to s71.

[0100] In this case, the processor 104 identifies the number of registered items scanned up to the timing in the POS data DB 121 for a particular shopping opportunity of a particular customer (s71). This identification of the number of registered items can be performed by aggregating the number of items purchased indicated by each POS data up to the timing in the POS data related to the shopping opportunity.

[0101] Next, the processor 104 notifies the display 13 of the POS register 10 and the user terminal 30 of the number of registered items identified in s71 above, as well as information prompting the user to check for any missing items (s72), and then terminates this flow. In this case, the processor 104 instructs the display 13 of the POS register 10 and the user terminal 30 to display, for example, the number of items registered up to that point and a message prompting the user to check for any missing items, as shown in screen G16 in Figure 33.

[0102] <Notification (Registration of discount stickers) flow> Next, we will explain the notification based on the scanning action of discount stickers in response to the detection of suspicious behavior, using the flow shown in Figure 22. In this case, the processor 104 refers to the POS data distributed from the POS register 10 each time a customer performs a scanning action, and identifies the scanning action of discount stickers (s80). In the case of the POS data already shown in Figure 4, if there is a value in the "discount sticker" column, it can be identified that a scanning action of discount stickers has occurred.

[0103] Furthermore, the processor 104 outputs a specific sound or message to the speaker 14 or display 13 of the POS register 10 (s81) for the scan operation identified in s80 as the scanning operation for the discount sticker, and terminates this flow. In this case, the processor 104 instructs the POS register 10 or user terminal 30 to display information such as the product to which the discount sticker was attached, "Sirloin Steak 250g," and the discounted amount "1,190 yen," as shown in screen G17 in Figure 34.

[0104] <Notification flow (different output control depending on register, operation, and timing)> Next, we will explain, based on the flow shown in Figure 23, how notifications based on output control that differs depending on the register, action, and timing when detecting suspicious behavior. In this case, the processor 104 refers to the POS data delivered from the POS register 10 each time a customer performs a scan, and identifies a specific action (for example, a cancel action or pressing the subtotal button) or timing (for example, every certain amount of time elapsed since the start of scanning) (s90).

[0105] Furthermore, the processor 104 determines whether the specific result, specific operation, or timing described above has been identified for each different register (s91). If, as a result of this determination, the specific operation or timing has not been identified for each different register, that is, it has been identified for only one register (s91:N), the processor 104 causes the POS register 10 and user terminal 30 to output a different sound for each operation or timing (s92), and then terminates this flow. On the other hand, if, as a result of the above determination, the specific operation or timing has been identified for each different register (s91:Y), the processor 104 causes the POS register 10 and user terminal 30 to output a different sound for each register and for each operation or timing (s93), and then terminates this flow.

[0106] <Notification (review submission) flow> Next, we will explain the notification process for posting reviews in response to the detection of suspicious behavior, based on the flow shown in Figure 24. In this case, the processor 104 obtains at least one of the following pieces of information from the performance information DB 126 (s100): the content or frequency of various suspicious customer behaviors detected for a particular store, or shoplifting incidents or the results of measures taken to prevent shoplifting incidents. The specific content of the information obtained here is as described in the explanation of the performance information DB 126 in Figure 9.

[0107] Next, the processor 104 generates a review (s101) that suggests the store has adequate anti-shoplifting measures, based on the information obtained in s100. This review generation can be performed, for example, by extracting the value of the "review" column from the records in the review table 129 that matches the actual information obtained in s100 (for example, "anti-shoplifting measures"). Alternatively, the generation AI may be given the information obtained in s100 and prompted to generate a review post that expresses that the store has adequate anti-shoplifting measures, and the generation AI may then generate the review.

[0108] Next, processor 104 posts the review obtained in s101 to the destination URL specified for "reviews" in the destination table 130 (s102), and terminates this flow. In this case, processor 104 posts, for example, a review about the target store on the review site published on network 1, as shown in screen G19 in Figure 36, with content such as "State-of-the-art cash registers have been installed, making checkout quick and easy. You don't see the staff very often, but it seems that cameras are installed in various places..." The URL of the review site to which the review is posted is the URL specified for "reviews" to be published in the destination table 130.

[0109] <Flowchart for exit control based on receipts> Next, the process of controlling customer exit from a specific area based on the receipt output by the POS register 10 will be explained using the flow chart in Figure 25. As a prerequisite for this flow chart, it is assumed that the receipt has at least the code G142 or G152, which encodes the receipt's identification information, printed on it by the POS register 10's printer 12, as shown in receipt G14 in Figure 31 and receipt G15 in Figure 32.

[0110] Furthermore, a physical gate exists along the route from the location of the store's POS register 10 to, for example, the store exit, where exit control based on the receipt is implemented. This gate is connected to the POS system 100 via network 1 and is equipped with opening and closing flaps that allow or prohibit customer passage based on instructions from the POS system 100. This gate also has a communication function that reads the barcode on the receipt presented by the customer and notifies the POS system 100 of the reading result via network 1.

[0111] In this case, the processor 104 retrieves information about the receipt issued by the POS register 10 from the POS register 10 and stores it in the receipt management DB 122 (s110). The processor 104 also identifies the receipts from which information was obtained in s110 that have an elapsed time from the time of issue that is within a certain limit (such as "within 20 minutes" as defined in the standard table 128) (s111).

[0112] Furthermore, for the receipt identified in s111, processor 104 assigns the value "Yes" to the "Exit Authority" column of the record in receipt management DB 122, indicating that the customer has the authority to exit from the specific area (s112). Subsequently, processor 104 determines whether the customer presented the receipt, i.e., whether the receipt was scanned at the gate, and whether or not a notification was received from the gate (s113). If, as a result of this determination, the customer did not present the receipt (s113:N), processor 104 terminates this flow.

[0113] On the other hand, if, as a result of the above determination, the customer presents a receipt (s113:Y), the processor 104 checks the status of the granting of exit authority for the receipt in the receipt management DB 122 (s114). This check can be performed by searching for records in the receipt management DB 122 using the receipt identification information as the key, based on the barcode reading result of the receipt as indicated by the notification from the gate.

[0114] Next, the processor 104, based on the result of s114, determines whether the customer who presented the receipt has the right to exit (s115). If the result of this determination is that the customer does not have the right to exit (s115:N), the processor 104, for example, causes the gate's user interface to output a message indicating that exit is not permitted (s116), and terminates this flow. On the other hand, if the result of the above determination is that the customer has the right to exit (s115:Y), the processor 104 notifies the gate to open the opening / closing flap and allow the customer to exit (s117), and terminates this flow.

[0115] <Cancellation process flow> Next, the process of detecting suspicious behavior in response to a customer canceling a transaction at the POS register 10 will be explained based on the flow shown in Figure 26. In this case, the processor 104 refers to the data related to the cancellation process that is delivered from the POS register 10 each time a customer cancels an action, and detects the event of a cancellation operation (s120).

[0116] Furthermore, the processor 104 outputs a message to the POS register 10 or user terminal 30 prompting the user to move the item for which a cancellation operation was detected in s120 to the cancellation box 40 (s121). In this case, the processor 104 will display a screen on the POS register 10's display 13 or the user terminal 30 that notifies the user of information about the item to be canceled and a message prompting the user to place the item in the cancellation box 40, for example, as shown in screen G20 in Figure 37.

[0117] Furthermore, the processor 104 compares the number of items subject to cancellation with the number of items stored in the cancellation box 40 (s122). As previously mentioned, the cancellation box 40 can identify at least the number of items, more preferably the items themselves, of the items subject to cancellation placed in its storage compartment, via the box camera 1063. Therefore, it is assumed that the cancellation box 40 continuously notifies the POS system 100 via the network 1 of information about the items subject to cancellation placed in the cancellation box 40.

[0118] Next, based on the result of s122, processor 104 determines whether the number of items targeted for cancellation at POS register 10 matches the number of items placed in cancellation box 40 (s123). If the number of items matches (s123:N), processor 104 terminates this flow. On the other hand, if the number of items does not match (s123:Y), processor 104 detects suspicious behavior in the customer's shopping activity (s124) and terminates this flow.

[0119] Although the best mode for carrying out the present invention has been described in detail above, the present invention is not limited thereto and can be modified in various ways without departing from its essence.

[0120] (program) In the POS system 100, the operation of each component is stored in the form of a program 111, for example, in the non-volatile area of ​​the auxiliary storage device 101. The processor 104, which implements the necessary functions, reads the program 111 from the non-volatile area and expands it into the volatile area, the main memory 103, and executes the above processing according to the program 111. The processor 104 also allocates the necessary storage area in the main memory 103 according to the program 111.

[0121] Specifically, the program 111 is a program that, in a POS system 100 equipped with a sensor (e.g., camera 106) for observing predetermined events related to customers in a store, applies the observation results obtained from the camera 106 sensor to a learning model 112, which is an algorithm for detecting suspicious behavior, and performs the process of detecting suspicious behavior by a customer, and notifying at least one of the store staff or the customer of the detection result or a predetermined notification content corresponding to that result.

[0122] The auxiliary storage device 101 is an example of a non-temporary tangible medium. Other examples of non-temporary tangible media include magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, and semiconductor memory connected via the communication device 105. Furthermore, if the program 111 is distributed from an external device to the POS system 100 via an appropriate network 1, the POS system 100 that receives the program 111 may load it into the main memory 103 and execute the above processing.

[0123] Furthermore, the program 111 may be for the purpose of implementing some of the functions described above. In addition, the program 111 may be a so-called differential file (differential program) that implements the functions described above in combination with other programs already stored in the non-volatile auxiliary storage device 101.

[0124] While several embodiments of this disclosure have been described above, these embodiments can be implemented in a variety of other forms, and various omissions, substitutions, and modifications are permitted without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0125] 1 Network 10 POS registers 11 Scanners 12 Printers 13 displays 14 speakers 15 Light 20 carts 21 Lower storage 30 User terminals 40 Cancel Box 100 POS systems 101 Auxiliary storage 102 Programs 103 Main storage 104 Processors 105 Communication equipment 106 Camera (Sensor) 1061 Register Camera 1062 Cart Camera 1063 Box Camera 110 OS 111 Programs 112 Learning Models 120 Product management DB 121 POS Data Database 122 Receipt Management Database 123 Observation Results Database 124 Customer management DB 125 Register Management DB 126 Performance Information Database 127 Cancellation Management DB 128 Reference Table 129 Customer Review Table 130 Public access table

Claims

1. A sensor that observes predetermined customer events in a store, A processor that performs the following processes: applying the observation results obtained from the sensor to an algorithm for detecting suspicious behavior to detect the customer's suspicious behavior; and notifying at least one of the store staff or the customer of the detection result or a predetermined notification content corresponding to the result. A POS system having [a certain feature].

2. The aforementioned processor, In response to the aforementioned notification, or in response to an action taken by a staff member who has received such notification, the following process is further executed regarding subsequent procedures relating to the customer: stopping or restricting the process to accept only staff operations. The POS system according to feature 1.

3. The aforementioned processor, In detecting the aforementioned suspicious behavior, the time interval between scanning operations for each product to be registered by the customer is determined based on the observation results of the sensor, and if the time interval does not meet the criteria, it is detected as suspicious behavior by the customer. The POS system according to feature 1.

4. The aforementioned processor, In detecting the aforementioned suspicious behavior, based on the observation results of the sensor, the number of items and the unit price of the items registered by the customer are identified, and if the number of items is above a certain threshold and the unit price is below a certain threshold, it is detected as suspicious behavior by the customer. The POS system according to claim 1 or 2.

5. The aforementioned processor, In detecting the aforementioned suspicious behavior, the system acquires video footage of the customer during a scanning operation related to the product to be registered from the sensor, inputs the customer video footage into the learning model, which is the algorithm, and determines whether the customer's movements or facial expressions constitute suspicious behavior. The POS system according to feature 1.

6. The aforementioned processor, In detecting the suspicious behavior, an image of the bottom of the shopping cart used by the customer is acquired from the camera, which is the sensor, and this image is input into the learning model, which is the algorithm, to determine whether or not there are large items in the bottom of the cart. The POS system according to feature 1.

7. The aforementioned processor, The image of the lower part of the cart is acquired from the camera located under the checkout counter through which the cart passes, or from the camera located under the cart. The POS system according to feature 6.

8. The aforementioned processor, In processing the aforementioned notification, the system further performs a process to highlight the output content related to individual items that also exist as large items among the items to be registered, or to such large items, on at least one of the screens and receipts for at least one of the staff or the customer, based on the information of the items, which is the result of the customer's scanning operation of the items to be registered. The POS system according to feature 6.

9. The aforementioned processor, In processing the aforementioned notification, based on the information of the product, which is the result of the customer's scanning operation of the product to be registered, the system further performs a process to highlight the output content of high-priced products among the products to be registered that have a unit price above a certain threshold, on at least one of the screens and receipts for at least one of the staff or the customer. The POS system according to feature 1.

10. The aforementioned processor, In issuing the aforementioned notification, the customer will be notified of dummy information indicating that an equipment error has occurred, and the staff will be notified of the results of the detection of suspicious behavior by the customer or a predetermined notification content corresponding to those results. The POS system according to feature 1.

11. The aforementioned processor, The notification is executed at any point between the customer scanning the items to be registered and the customer completing the payment for those items, and each notification outputs information about the number of items scanned up to that point and information prompting the customer to check for any missing items. The POS system according to feature 1.

12. The aforementioned processor, When the customer scans a product to be registered, and a discount sticker attached to the product is scanned, the aforementioned notification will be given in a predetermined specific form. The POS system according to feature 1.

13. The aforementioned processor, The system outputs different sounds as notifications for each type or timing of any action, and for each register to which the action is performed, from the customer's scanning of the product to the payment of the product. The POS system according to feature 1.

14. The aforementioned processor, In issuing the aforementioned notification, an image including the customer whose suspicious behavior was detected will be obtained from the sensor, and after masking at least the face portion of the person's image in the image, the image will be published on a designated website that is accessible to at least the customer. The POS system according to feature 1.

15. The aforementioned processor, The images to be published on the aforementioned website are obtained from the sensor, capturing images of at least one of multiple parts of the customer or multiple parts of the cash register while the customer is working, and the obtained images are published on the aforementioned website. The POS system according to feature 14.

16. The aforementioned processor, In issuing the aforementioned notification, information suggesting the enhanced shoplifting prevention measures at the store, based on the content or frequency of the detected suspicious activity, will be published on a designated website. The POS system according to feature 1.

17. The aforementioned processor, In issuing the aforementioned notification, information on shoplifting losses at the store or the measures taken to prevent such shoplifting will be obtained from a designated system, and information indicating the level of shoplifting prevention measures at the store will be published on a designated website, depending on the magnitude of the shoplifting losses or the measures taken as indicated by the obtained information. The POS system according to feature 1.

18. The aforementioned processor, The process involves obtaining identification information and issuance time information of a receipt issued at the register in connection with the payment of goods registered by the customer from a predetermined system; granting exit rights from a predetermined area in the store to receipts whose elapsed time from the issuance time is within a certain limit; having the sensor read the identification information of the receipt presented by the customer attempting to leave the predetermined area, and determining whether the receipt corresponding to the identification information has exit rights; and further executing a process to detect the customer's exit as suspicious behavior based on the result of the determination. The POS system according to feature 1.

19. The aforementioned processor, The system further detects when the customer cancels an item registered at the register and outputs a message prompting the customer to transfer the item to a designated storage device. In detecting the suspicious behavior, the system compares the number of items that were canceled with the number of items stored in the storage device to determine if the cancellation operation is the suspicious behavior. The POS system according to feature 1.

20. The POS system Equipped with sensors to observe predetermined customer events in a store, The process involves applying the observation results obtained from the sensor to an algorithm for detecting suspicious behavior to detect the customer's suspicious behavior, and notifying at least one of the store staff or the customer of the detection result or a predetermined notification content corresponding to the result. A POS system operation method characterized by performing the following.

Images