JP7060353B2 - Sales amount totaling system - Google Patents

Description

The present invention relates to a sales amount totaling system that totals the sales amount of a product sold to a visitor who visits a store having a sales floor space of a predetermined area while using artificial intelligence.

A food and drink management device for managing food and drink contained and transported in a food and drink container placed on a transport path of the food and drink transport device is disclosed (see Patent Document 1). The food and drink container includes a mounting table having a container mounting portion for mounting a container on which food and drink are placed, and a lid for covering the container mounting portion of the mounting table. , It is attached to the container mounting part so that it can be opened and closed. The food and drink management device includes an IC tag attached to the food and drink container, a reader for reading food and drink information from the IC tag of the food and drink unit container transported via the transport path, and a food and drink container. Calculates information on food and drink transported via the transport path based on the open / close detector for detecting the open / close of the lid, the information read by the reader, and the open / close information of the lid detected by the open / close detector. Equipped with a controller for. The food and drink management device does not need to install an IC tag on the plate used to provide food and drink, and ensures information on food and drink that has been eaten and information on food and drink that stays on the transport route for a certain period of time without being eaten. Can be obtained in.

Japanese Unexamined Patent Publication No. 2014-33886

The food and drink management device disclosed in Patent Document 1 divides each eating and drinking space of a restaurant that provides food and drink into the first to nth eating and drinking spaces, and determines the amount of food and drink provided for each eating and drinking space. The total amount of food and drink is not calculated, and the total amount of food and drink for each eating and drinking space of the restaurant cannot be grasped. In addition, investors are solicited for each of the first to nth spaces of restaurants, and investors invest a predetermined amount in each of the first to nth spaces, and food and drink for each of the first to nth spaces. It is not possible to provide investors with a mechanism to kick back (return) a margin (a part of profit) according to the total amount of sales, and it is not possible to build a business model as a new method of collecting funds in stores.

An object of the present invention is to use artificial intelligence to divide the sales floor space of a store into a plurality of first to nth spaces, and to obtain a total amount of the total amount of goods sold in the first to nth spaces. It is an object of the present invention to provide a sales amount totaling system that can be calculated for each of the 1st to nth spaces and can accurately determine the quality of sales of a product in each space of a store. Another object of the present invention is to be able to grasp the total sales amount of the products for each of the first to nth spaces of the store, and to remodel and arrange the sales floor space with reference to the total sales amount for each of the first to nth spaces. The purpose is to provide a sales amount totaling system that can consider modifications for improving sales and profits such as replacement. Another object of the present invention is to solicit investors for each of the first to nth spaces of the store, and the investors invest a predetermined amount in each of the first to nth spaces, and each of the first to nth spaces. It is possible to provide investors with a mechanism to kick back (return) a margin (a part of profit) according to the total sales amount of each product, and to build a business model as a new method of collecting funds in stores. It is to provide a sales amount totaling system that can be used.

The premise of the present invention for solving the above-mentioned problems is a sales amount totaling system that totals the sales amount of products sold to a visitor who visits a store having a sales floor space of a predetermined area while using artificial intelligence.

The feature of the present invention in the above premise is that the sales amount totaling system is formed by a server in which artificial intelligence is set up and a photographing camera installed in the sales floor space, and the sales floor space of the store is a plurality of first to nth. It is divided into spaces, and the server trains artificial intelligence so that it can recognize the required various product images from various sample images with high accuracy, and the recognized various product images are the product images. 1st to nth spaces including product images sold in the 1st to nth spaces using a learning means to train artificial intelligence so that they can be classified into various product types to which the product belongs. Either the space image shooting means for shooting an image in each of the first to nth spaces or the product type in which the product images included in the first to nth space images shot by the space image shooting means are learned by the learning means. The amount of goods sold in the 1st to nth spaces was totaled while specifying the amount of each product judged by the product type judgment means for making artificial intelligence judge whether it belongs to the type of the product and the product type judgment means. It is to have a total amount calculation means for calculating the total amount for each of the first to nth spaces.

As an example of the present invention, the store is a restaurant that provides various kinds of food and drink, the sales floor space is a food and drink space having a predetermined area for eating food and drink, and the food and drink space is a plurality of first to first. n It is divided into eating and drinking spaces, and the learning means is made to learn by artificial intelligence so that various necessary food and drink images can be recognized with high accuracy from various sample images, and various recognized food and drinks are recognized. The images are trained by artificial intelligence so that the images can be classified into various types of foods and drinks to which the food and drink images belong, and the space image photographing means includes the food and drink images provided in the first to nth food and drink spaces. The first to nth eating and drinking space images are taken for each of the first to nth eating and drinking spaces, and the product type determination means is the food and drink images included in the first to nth eating and drinking space images taken by the space image taking means. The first method is to have artificial intelligence determine which of the food and drink types learned by the learning means, and the total amount calculation means specifies the amount of each food and drink determined by the product type judgment means. -The total amount of the total amount of food and drink provided in the nth eating and drinking space is calculated for each of the first to nth eating and drinking spaces.

As another example of the present invention, the space image photographing means includes the food and drink images provided in one eating and drinking act in the first to nth eating and drinking spaces or the eating and drinking act during the business hours of the day. The nth eating and drinking space image is taken for each of the first to nth eating and drinking spaces, and the total amount calculation means is provided in one eating and drinking act in the first to nth eating and drinking spaces or in the eating and drinking act during the business hours of the day. The total amount of food and drink is calculated for each of the first to nth food and drink spaces.

As another example of the present invention, the learning means is trained by artificial intelligence so that various necessary container images can be recognized with high accuracy from various sample images, and the first to nth eating and drinking spaces are used. The period from the time when the first container containing food and drink is provided to the time when the container containing food and drink disappears in the first to nth eating and drinking spaces is artificially determined to be one eating and drinking act. In addition to letting intelligence learn, if a container containing food and drink is provided to the 1st to nth eating and drinking spaces between the opening time and the closing time of the restaurant, it is judged to be eating and drinking during the business hours of the day. Let artificial intelligence learn to do.

As another example of the present invention, in the learning means, various sample images of various correct answers and various sample images of various errors are input to the artificial intelligence, and the sample images of the correct answer and the sample images of the errors of the artificial intelligence are used. Artificial intelligence is trained so that the correct food and drink image can be recognized at a predetermined correct answer rate.

As another example of the present invention, in the learning means, the correct answer rate of artificial intelligence that recognizes the correct food and drink image from those sample images of correct answers and those sample images of errors is in the range of 80 to 100%.

As another example of the present invention, in the learning means, artificial intelligence removes unnecessary images other than food and drink images from various sample images and acquires various necessary food and drink images.

According to the sales amount totaling system according to the present invention, the sales floor space of a store is divided into a plurality of first to nth spaces, and various necessary product images can be recognized with high accuracy from various sample images. A shooting camera installed in the sales floor space by learning artificial intelligence so that it can be learned by artificial intelligence so that various recognized product images can be classified into various product types to which those product images belong. The first to nth space images including the product images sold in the first to nth spaces are photographed for each of the first to nth spaces, and are included in the photographed first to nth space images. It is sold in the 1st to nth spaces while letting artificial intelligence determine which type of product the product image belongs to among the product types learned by the learning means and specifying the amount of each product determined by the artificial intelligence. Since the total amount of the total amount of the products sold is calculated for each of the first to nth spaces, the sales floor space of the store is divided into a plurality of first to nth spaces, and the types of the sold products are classified by artificial intelligence. It can be reliably specified, and the total amount of the total amount of the products sold in the 1st to nth spaces can be calculated for each of the 1st to nth spaces, and the products in each space of the store can be calculated. It is possible to accurately judge the quality of sales. The sales amount totaling system can accurately grasp the total sales amount for each of the first to nth spaces of the store by calculating the accurate total sales amount of the products for each of the first to nth spaces. With reference to the total amount of money for each of the first to nth spaces, it is possible to consider modifications for improving sales and profits such as remodeling and rearrangement of sales floor spaces. The sales amount totaling system, for example, recruits investors for each of the first to nth spaces of a store, and the investors invest a predetermined amount in each of the first to nth spaces, and each of the first to nth spaces. It is possible to provide investors with a mechanism to kick back (return) a margin (a part of profit) according to the total sales amount of each product or service, and build a business model as a new method of collecting funds in stores. can do.

The store is a restaurant that provides various kinds of food and drink, the sales floor space is a food and drink space of a predetermined area for eating food and drink, and the food and drink space is divided into a plurality of first to nth food and drink spaces, and various types of food and drink are provided. Artificial intelligence is trained to recognize various required food and drink images from sample images with high accuracy, and various food and drink images recognized are used as various food and drink types to which the food and drink images belong. The first to nth eating and drinking space images including the eating and drinking images provided in the first to nth eating and drinking spaces are photographed for each of the first to nth eating and drinking spaces. Then, the artificial intelligence is made to judge which of the food and drink types the food and drink images included in the captured 1st to nth eating and drinking space images belong to, and the amount of each food and drink determined by the artificial intelligence. The sales amount totaling system that calculates the total amount of food and drink provided in the 1st to nth eating and drinking spaces for each of the 1st to nth eating and drinking spaces while specifying The types of food and drink provided can be reliably specified by artificial intelligence while being classified into the 1st to nth eating and drinking spaces, and the total amount of food and drink provided in the 1st to nth eating and drinking spaces is totaled. The total amount can be calculated for each of the first to nth eating and drinking spaces, and the quality of sales of food and drink in each eating and drinking space of the restaurant can be accurately determined. The sales amount totaling system accurately grasps the total sales amount for each of the 1st to nth eating and drinking spaces of the restaurant by calculating the accurate total sales amount of the food and drink for each of the 1st to nth eating and drinking spaces. Therefore, it is possible to consider modifications for improving sales and profits such as remodeling and rearrangement of eating and drinking spaces with reference to the total amount of money for each of the first to nth eating and drinking spaces. The sales amount totaling system, for example, recruits investors for each of the 1st to nth eating and drinking spaces of a restaurant, and the investors invest a predetermined amount in each of the 1st to nth eating and drinking spaces, and each 1st to nth eating and drinking space. It is possible to provide investors with a mechanism to kick back (return) a margin (a part of profit) according to the total sales amount of food and drink in the nth eating and drinking space, as a new method of collecting funds at restaurants. You can build a business model.

The 1st to nth eating and drinking space images including the food and drink images provided in one eating and drinking act in the 1st to nth eating and drinking space or the eating and drinking act during the business hours of the day are displayed for each 1st to nth eating and drinking space. The total amount of food and drink provided in one eating and drinking activity in the 1st to nth eating and drinking spaces or in the eating and drinking activities during the business hours of the day is totaled for each 1st to nth eating and drinking space. The calculated sales amount totaling system can reliably identify the type of food and drink provided by artificial intelligence, and in the 1st to nth eating and drinking spaces in one eating and drinking activity or eating and drinking activity during the business hours of the day. The total amount of food and drink provided can be calculated for each of the 1st to nth eating and drinking spaces, and each eating and drinking space in one eating and drinking activity of the restaurant or in the eating and drinking activity during the business hours of the day. It is possible to accurately judge the quality of sales of each food and drink.

The first container that accommodates food and drink is provided in the first to nth eating and drinking spaces by letting artificial intelligence learn so that various necessary various container images can be recognized with high accuracy from various and multiple sample images. From the time when the food is served to the time when the container containing the food and drink disappears in the 1st to nth eating and drinking spaces, the artificial intelligence is trained so that it is judged as one eating and drinking act, and from the opening time of the restaurant. A food and beverage amount totaling system that trains artificial intelligence to determine that a container containing food and drink is provided to the 1st to nth eating and drinking spaces during the business hours of the day as eating and drinking activities. Is once from the time when the first container containing food and drink is provided in the first to nth eating and drinking spaces until the container containing food and drink disappears in the first to nth eating and drinking spaces. By letting the artificial intelligence learn to judge that the eating and drinking behavior is, one eating and drinking behavior in the 1st to nth eating and drinking spaces is clearly classified by the artificial intelligence, and the 1st to nth eating and drinking in one eating and drinking behavior. The total amount of food and drink provided in the space can be calculated for each of the 1st to nth eating and drinking spaces, and the sales of food and drink for each eating and drinking space in one eating and drinking activity of the restaurant can be calculated. It is possible to accurately judge the quality. The sales amount totaling system determines that a container containing food and drink is provided to the 1st to nth eating and drinking spaces between the opening time and the closing time of the restaurant as eating and drinking during the business hours of the day. By letting artificial intelligence learn to do so, the eating and drinking activities of the first to nth eating and drinking spaces during the business hours of the day are clearly classified by the artificial intelligence, and the first to nth eating and drinking activities during the business hours of the day are clearly classified. The total amount of food and drink provided in the food and drink space can be calculated for each of the 1st to nth food and drink spaces, and the food and drink for each food and drink space during the daily business hours of the restaurant can be calculated. It is possible to accurately judge the quality of sales of goods.

Input sample images of various correct answers and sample images of various errors into artificial intelligence, and the artificial intelligence recognizes the correct food and drink images from those sample images of correct answers and those sample images of errors at a predetermined correct answer rate. By setting a predetermined correct answer rate, the sales amount totaling system that trains artificial intelligence so that it can accurately recognize various food and drink images from various food and drink sample images. Can be trained by artificial intelligence, and various recognized food and drink images can be trained by artificial intelligence so that they can be classified into various food and drink types to which the food and drink images belong. In the sales amount totaling system, the types of food and drink are accurately classified by artificial intelligence, and the types of food and drink provided can be reliably specified by artificial intelligence, and the food and drink provided in the first to nth eating and drinking spaces. The total amount of the total amount of food and drink can be calculated for each of the first to nth eating and drinking spaces, and the quality of sales of food and drink in each eating and drinking space of the restaurant can be accurately determined.

The sales amount totaling system that recognizes the correct food and drink image from those sample images of correct answers and those sample images of errors has a correct answer rate of 80 to 100%, and the correct answer rate of artificial intelligence is 80 to 100. By setting it in the range of%, it is possible to train artificial intelligence so that various food and drink images can be accurately recognized from various sample images, and various recognized food and drink images can be recognized. Artificial intelligence can be trained so that the food and drink images can be classified into various types of food and drink to which they belong. In the sales amount totaling system, the types of food and drink are accurately classified by artificial intelligence, and the types of food and drink provided can be reliably specified by artificial intelligence, and the food and drink provided in the first to nth eating and drinking spaces. The total amount of the total amount of food and drink can be calculated for each of the first to nth eating and drinking spaces, and the quality of sales of food and drink in each eating and drinking space of the restaurant can be accurately determined.

The sales amount totaling system that artificial intelligence removes unnecessary images other than food and drink images from various sample images and acquires various necessary food and drink images, recognizes the correct food and drink image at a predetermined correct answer rate. Since the artificial intelligence learned so that it can remove unnecessary images other than food and drink images from the sample images and acquire various necessary food and drink images, various food and drink from various sample images. Images can be reliably acquired, various recognized food and drink images can be classified into various food and drink types to which the food and drink images belong, and the types of food and drink provided can be reliably specified. can do.

The block diagram of the sales amount totaling system shown as an example. The figure which shows an example of a sample image taking. The figure which shows another example of the sample image taking. The figure which shows another example of the sample image taking. The figure which shows an example of taking the food and drink image in the 1st to 4th eating and drinking space. The figure which shows another example of taking the food and drink image in the 1st to 4th eating and drinking space.

The details of the sales amount totaling system according to the present invention will be described below with reference to the attached drawings such as FIG. 1, which is a configuration diagram of the sales amount totaling system 10 shown as an example. Note that FIG. 2 is a diagram showing an example of sample image shooting, and FIG. 3 is a diagram showing another example of sample image shooting. FIG. 4 is a diagram showing another example of taking a sample image.

The sales amount totaling system 10 is a food or drink 13 (at least one of the sold product and the provided service) provided to a visitor who visits the restaurant 12 (store) having the eating and drinking space 11 (sales floor space) having a predetermined area. Aggregate the sales amount. Although the sales amount totaling system 10A will be described by taking the restaurant 12 as an example, the sales amount totaling system 10A is described not only in the restaurant 12 but also in stores of all industries engaged in at least one of selling products and providing services. (Including the sales amount totaling system 10B) can be used. Further, the food and drink 13 is not particularly limited, and the food and drink 13 includes all foods and drinks 13 provided (sold) at various restaurants 12.

The sales amount totaling system 10 includes a server 14 (including cloud computing) on which artificial intelligence is set up, and first to fourth eating and drinking spaces 11a to 11d (first to nth eating and drinking spaces, first to nth sales floor spaces). ), It is formed of the first to fourth photographing cameras 15a to 15d (first to nth photographing cameras) individually installed. The server 14 is installed in a data center contracted by a business operator that manages and provides the sales amount totaling system 10, but may be installed in a restaurant 12 (store) that uses the sales amount totaling system 10.

The eating and drinking space 11 of the restaurant 12 is divided into a plurality of first to fourth eating and drinking spaces 11a to 11d (first to nth eating and drinking spaces). As an example of dividing the eating and drinking space 11 into the first to fourth eating and drinking spaces 11a to 11d, it is divided by 18 units of eating and drinking tables (first to fourth eating and drinking tables 18a to 18d) installed in the restaurant 12, or The eating and drinking space 11 of the restaurant 12 is divided into area units (for example, the area of the eating and drinking space 11 is evenly divided or the area of the eating and drinking space 11 is unevenly divided). In FIG. 1, the eating and drinking space 11 is divided by the eating and drinking table 18.

The sales amount totaling system 10 recruits investors (owners) for each of the first to fourth eating and drinking spaces 11a to 11d, and the investors invest a predetermined amount in each of the eating and drinking spaces 11a to 11d. In the sales amount totaling system 10A, a margin (a part of profit) is kicked back (returned) to the investor according to the total sales amount of the food and drink 13 for each of the first to fourth eating and drinking spaces 11a to 11d. The server 14 is a computer (including a virtual machine) having a central processing unit (CPU or MPU) and a memory (main memory and cache memory) and operated by an independent operating system (OS), and has a large capacity storage. Implements the area. Although not shown, an input device such as a keyboard and a mouse, and an output device such as a display, a touch panel, and a printer are connected to the server 14 via an interface.

The server 14 has various server functions such as a DSN server function, a database server function, a Web server function, a mail server function, and a document server function, and the Internet 16 is connected to the Internet 16 via various various DSN servers and the like. You can access and log in to any other unspecified number of other servers (physical servers and virtual servers) that have no access restrictions. The server 14 stores (stores) an application including artificial intelligence (AI) for implementing each means of the sales amount totaling system 10A.

In the large-capacity storage area of the server 14, the first to fourth eating and drinking spaces 11a to 11d are associated with the first to fourth eating and drinking space specific identifiers that specify the spaces 11a to 11d and the investor identifier that identifies the investor. It is stored (stored) in the state of being stored. In the large-capacity storage area of the server 14, the first photographing camera 15a is stored (stored) in a state associated with the first eating / drinking space specific identifier, and the second photographing camera 15b is associated with the second eating / drinking space specific identifier. It is stored (stored) in a state, and the third photographing camera 15c is stored (stored) in a state associated with the third eating / drinking space specific identifier, and the fourth photographing camera 15d is associated with the fourth eating / drinking space specific identifier. It is stored (stored) in the state of being stored.

The first to fourth shooting cameras 15a to 15d (first to nth shooting cameras) are connected to the router 19 installed in the restaurant 12, and the first to fourth eating and drinking space images (first to fourth eating and drinking space images) are connected via the router 19. It is transmitted to the server 14 using the 1st to nth eating and drinking space images). The server 14 and the router 19 are connected by the Internet 16 (network). The first photographing camera 15a is installed above the first eating and drinking space 11a (first eating and drinking table 18a), and is a first eating and drinking space image (first eating and drinking table image) including food and drink 13 provided to the first eating and drinking space 11a. ). The first photographing camera 15a connects to the Internet 16 by using the router 19, and transmits the photographed first eating and drinking space image to the server 14 via the router 19. The second photographing camera 15b is installed above the second eating and drinking space 11b (second eating and drinking table 18b), and is a second eating and drinking space image (second eating and drinking table image) including the food and drink 13 provided to the second eating and drinking space 11b. ). The second photographing camera 15b connects to the Internet 16 by using the router 19, and transmits the photographed second eating and drinking space image to the server 14 via the router 19.

The third photographing camera 15c is installed above the third eating and drinking space 11c (third eating and drinking table 18c), and is a third eating and drinking space image (third eating and drinking table image) including the food and drink 13 provided to the third eating and drinking space 11c. ). The third photographing camera 15c connects to the Internet 16 by using the router 19, and transmits the photographed third eating and drinking space image to the server 14 via the router 19. The fourth shooting camera 15d is installed above the fourth eating and drinking space 11d (fourth eating and drinking table 18d), and is a fourth eating and drinking space image (fourth eating and drinking table image) including food and drink provided to the fourth eating and drinking space 11d. To shoot. The fourth photographing camera 15d connects to the Internet 16 by using the router 19, and transmits the photographed fourth eating and drinking space image to the server 14 via the router 19.

The business operator who manages and provides the sales amount totaling system 10 starts the server 14, and as shown in FIGS. 2 to 4, the digital shooting function (digital camera function) of the digital camera 20 (2D digital camera and 3D digital camera). ) Is used to shoot various multiple samples (food and drink subjects, container subjects, and various other subjects), and various two-dimensional sample images (sample two-dimensional image data) and three types of the captured samples. 3D sample image (3D image data of the sample) is input to the server 14.

A business operator that manages and provides the sales amount totaling system 10 accesses each portal site (website) using the Internet 16, and various various samples (food and drink subjects and container subjects, etc.) posted on those portal sites. Various plurality of 2D sample images (sample 2D image data) and various plurality of 3D sample images (sample 3D image data) of various other various subjects are input to the server 14. The server 14 stores (stores) a two-dimensional sample image (including an input date and time) and a three-dimensional sample image (including an input date and time) input by a business operator in a large-capacity storage area (sample image storage means).

Users of each restaurant and restaurant use a sample (provided at each restaurant) using a digital camera 20 (2D digital camera and 3D digital camera), a tablet or smartphone with a digital shooting function (digital camera function), etc. Various 2D sample images (sample 2D image data) of various samples taken by shooting food and drink subjects, container subjects used in each restaurant, and various other subjects) and various 3 A three-dimensional sample image (three-dimensional image data of a sample) can be transmitted to the server 14 via the Internet 16. The server 14 receives the 2D sample image and the 3D sample image of the sample transmitted from each restaurant and the user of the restaurant, and also receives the received 2D sample image (including the reception date and time) and the 3D sample image. Stores (stores) (including the reception date and time) in a large-capacity storage area (sample image storage means).

In the 2D sample image and the 3D sample image, in addition to various food and drink images (food and drink itself) and various container images (container itself), food and drink stored in the container and stored in a bag. Foods and drinks, foods and drinks housed in showcases, etc. are included. In FIG. 4, carrots (vegetables), pens (writing tools), and pots (cooking tools) are illustrated as various other subjects, but for various other subjects, a digital camera 20, a tablet, a smartphone, or the like is used. Includes all existing objects (subjects) that can be photographed.

The server 14 starts the application according to the control of the operating system and executes each of the following means. A large amount (tens of thousands) of 2D sample images and 3D sample images are input to the artificial intelligence set up on the server 14 by the business operator who manages and provides the sales amount totaling system 10, and a large amount (tens of thousands). The two-dimensional sample image and the three-dimensional sample image of (sheet)) are transmitted (input) from each restaurant or a user who uses the restaurant.

The server 14 recognizes various necessary two-dimensional food and drink images (two-dimensional product image) and three-dimensional food and drink image (three-dimensional product image) from various two-dimensional sample images and three-dimensional sample images with high accuracy. In addition to training artificial intelligence so that it can be done, it is possible to accurately recognize various required 2D container images and 3D container images from various 2D sample images and 3D sample images. Let artificial intelligence learn as much as possible (learning means).

In the learning means, various sample images (various two-dimensional sample images and three-dimensional sample images) are input to the artificial intelligence, and the artificial intelligence determines the correct food and drink image and the correct container image from the sample images. Artificial intelligence is trained so that it can be recognized by the correct answer rate, and artificial intelligence can classify the correct food and drink images recognized by the artificial intelligence into various types of food and drink to which those food and drink images belong. To learn. The correct answer rate of artificial intelligence that recognizes the correct food and drink image and the correct container image is in the range of 80 to 100%, and the correct food and drink image is classified into various types of food and drink to which the food and drink image belongs. The correct answer rate of artificial intelligence is in the range of 80 to 100%.

In the learning means, various correct 2D sample images and 3D sample images and various multiple error 2D sample images and 3D sample images are input to the artificial intelligence, and the artificial intelligence is the correct 2D. Recognizes food and drink images, 3D food and drink images of solutions, 2D container images of solutions, 3D container images of solutions, 2D and other various subject images of correct answers, and 3D and other various subject images of correct answers at a predetermined accuracy rate. In addition to learning artificial intelligence so that it can be done, the correct 2D food and drink image, the solution 3D food and drink image, the solution 2D container image, the solution 3D container image, the correct answer 2D and various other subjects It is also possible to train artificial intelligence so that an image, three-dimensional correct answers, and various other subject images can be classified into various types to which the images belong at a predetermined correct answer rate. The correct answer rate of artificial intelligence that recognizes the correct answer image from those sample images of correct answers and those sample images of errors, and the correct answer rate of artificial intelligence that divides the correct answer images into various types to which each image belongs is 80 to 100. It can be set in the range of%. In addition, learning in artificial intelligence includes various learning methods such as existing supervised learning and unsupervised learning, and also includes various learning methods to be developed in the future.

Artificial intelligence is a computer that thinks for itself and creates intelligence. As an example, deep learning is performed based on an artificial neural network that simulates the process in which a network of neurons (nerve cells) in the brain learns the correct image gradually and learns how to make its own decisions. Artificial neural networks with complex links aimed at recognizing (understanding) specific foods, containers, and articles can be classified as foods and drinks, containers, and various items when trained by inputting a large number of images. By learning to classify and detect individual foods and drinks, individual containers, and individual articles, you can train foods, containers, and articles that you have seen in the past with training images (sample images of errors). You will be able to recognize even if it is different.

In order to recognize food and drink, containers, and various other articles, individual images of input images (2D sample images and 3D sample images) are analyzed in the input layer of the artificial neural network, and in the next layer (hidden layer). Extracts the unique geometric shape of a particular food or drink image, container image, or other article image. In the middle layer (hidden layer), the characteristics of each food and drink, each container, and other articles are detected, and in the deeper layer (hidden layer), the entire food and drink, the entire container, and all other articles that combine them are recognized. do. Then, in the output layer of the artificial neural network, it is presumed that the food and drink is fried rice, not ramen or dumplings, and that the container is a flat plate, not a bowl or bowl. In addition, he speculates that the vegetables are carrots, not cabbage, that the writing utensils are pens, not rulers, and that the cooking utensils are pots, not frying pans.

When the learning of artificial intelligence progresses in the learning means and the correct answer rate of artificial intelligence becomes in the range of 80 to 100%, the artificial intelligence becomes unnecessary images (table) other than food and drink images and container images from various two-dimensional sample images. Various 2D food and drink images and 2D container images required by removing (cutting out) images of cloth, forks, knives, etc. are acquired, and food and drink images and containers are obtained from various 3D sample images. Unnecessary images other than images (images of table cloths, forks, knives, etc.) are removed (cut out) to acquire various necessary 3D food and drink images and 3D container images.

Furthermore, artificial intelligence is an unnecessary image other than carrot images (other article images), pen images (other article images), and pot images (other article images) (other than article images) from various multiple two-dimensional sample images. Remove (cut out) the (background image) to obtain the required multiple 2D carrot images, 2D pen images, and 2D pot images, as well as carrot images (other articles) from the various 3D sample images. Remove (cut out) unnecessary images (background images) other than (images), pen images (other article images), pot images (other article images), and various necessary 3D carrots. Acquire an image, a three-dimensional pen image, and a three-dimensional pot image.

In the learning means, from the time when the first container containing food and drink is provided in the first to fourth eating and drinking spaces until the container containing food and drink disappears in the first to fourth eating and drinking spaces. Let artificial intelligence learn to judge it as one eating and drinking act. Furthermore, in the learning means, when a container containing food and drink is provided to the first to fourth eating and drinking spaces between the opening hours and the closing hours of the restaurant, it is judged to be eating and drinking during the business hours of the day. Let artificial intelligence learn to do.

The server 14 stores (stores) those food and drink images recognized by the artificial intelligence learning means in a large-capacity storage area (food and drink image storage means), and stores those container images recognized by the artificial intelligence learning means. In addition to storing (storing) in the capacity storage area (container image storage means), other article images recognized by the learning means of artificial intelligence are stored (stored) in the large capacity storage area (article image storage means).

In the large-capacity storage area of the server 14, the food and drink image data (a plurality of food and drink image data of each food and drink) of various food and drink 13 provided in the restaurant 12 and the food and drink name of each food and drink image data are stored. And the food and drink amount (selling price or profit) of the food and drink name are stored (stored) in a state associated with the food and drink specific identifier that identifies the food and drink 13. In the large-capacity storage area of the server 14, container image data of various various containers 17 accommodating various foods and drinks 13 provided in the restaurant 12 are stored (stored).

By setting the correct answer rate of artificial intelligence in the range of 80 to 100%, the sales amount totaling system 10 makes it possible to accurately recognize various food and drink images from various sample images. In addition to being able to learn, artificial intelligence can be trained so that the recognized various food and drink images can be classified into various food and drink types to which the food and drink images belong.

The sales amount totaling system 10 is a plurality of various types required by artificial intelligence learned so that the correct food and drink image can be recognized at a predetermined correct answer rate by removing unnecessary images other than the food and drink image from the sample image. Since the food and drink images are acquired, it is possible to reliably acquire various food and drink images from various sample images, and the recognized various food and drink images can be obtained as various food and drink types to which the food and drink images belong. In addition to being able to classify into, it is possible to reliably identify the type of food and drink provided.

FIG. 5 is a diagram showing an example of shooting food and drink images in the first to fourth eating and drinking spaces 11a to 11d, and FIG. 6 is a diagram showing other shooting of food and drink images in the first to fourth eating and drinking spaces 11a to 11d. It is a figure which shows an example. When the sales amount totaling system 10 (server 14 and the first to fourth shooting cameras 15a to 15d) is started, the server 14 uses the first to fourth shooting cameras 15a to 15d to use the first to fourth eating and drinking spaces 11a. The first to fourth eating and drinking space images including the food and drink 13 provided in ~ 11d are taken for each of the first to fourth eating and drinking spaces 11a to 11d (space image photographing means).

The server 14 transmits the shooting start signal of the first to fourth eating and drinking space images to the first to fourth shooting cameras 15a to 15d (via the router 19) using the Internet 16. When the first to fourth shooting cameras 15a to 15d receive the shooting start signal from the server 14, they start shooting the first to fourth eating and drinking spaces 11a to 11d, and from the start of business hours to the end of business hours at the restaurant 12. During the period, the images of the first to fourth eating and drinking spaces are taken, and the images of the first to fourth eating and drinking spaces are transmitted to the server 14 via the router 19.

In the artificial intelligence, the first container 17 containing the food and drink 13 is provided in the first to fourth eating and drinking spaces 11a to 11d, and the food and drink are eaten from the time when the container image included in the first to fourth eating and drinking space images is recognized. The container 17 containing the food and drink 13 does not exist in the spaces 11a to 11d, and the period until the absence of the container image included in the first to fourth food and drink space images is recognized is recognized as one eating and drinking act.

As shown in FIG. 5, the server 14 is provided with the first (first) container 17 containing the food and drink 13 in the first to fourth eating and drinking spaces 11a to 11d (food and drink table), and the artificial intelligence is the first. When the second container image (container 17) is recognized, it is included in the first to fourth eating and drinking space images transmitted from the first to fourth photographing cameras 15a to 15d triggered by the recognition of the first container image in artificial intelligence. Let artificial intelligence determine the type of food and drink in the food and drink image. In the server 14, the foods and drinks 13 provided in the first to fourth eating and drinking spaces 11a to 11d (food and drink tables) are eaten, and as shown in FIG. 6, the container 17 containing the foods and drinks 13 is the eating and drinking spaces 11a to 11d. When the container image (container 17) disappears from the eating and drinking spaces 11a to 11d, the determination of the food and drink type in the artificial intelligence is stopped by the recognition of the absence of the container image in the artificial intelligence.

For artificial intelligence, containers 17 containing foods and drinks 13 are provided in the first to fourth eating and drinking spaces 11a to 11d from the opening hours of the day at the restaurant, and when the business hours of the day end, the day is reached. Judge that eating and drinking during business hours has ended. The server 14 is provided with the first (first) container 17 containing the food and drink 13 in the first to fourth eating and drinking spaces 11a to 11d (food and drink table) from the business start time of the day, and the artificial intelligence is the first. When the second container image (container 17) is recognized, it is included in the first to fourth eating and drinking space images transmitted from the first to fourth photographing cameras 15a to 15d triggered by the recognition of the first container image in artificial intelligence. The artificial intelligence is made to judge the food and drink type of the food and drink image, and the judgment of the food and drink type in the artificial intelligence is stopped by the time when the business hours of the day are over.

When the first to fourth eating and drinking space images taken by the first to fourth shooting cameras 15a to 15d are transmitted to the server 14, the server 14 uses the learning means to obtain the food and drink images included in the first eating and drinking space image. Have artificial intelligence determine which of the learned food and drink types belongs to each food and drink activity or each food and drink activity during business hours of the day (food and beverage type determination means, product type determination means). The server 14 acquires the type (food and drink name) of the food and drink image (food and drink 13) of the first eating and drinking space image determined by the artificial intelligence from the artificial intelligence, and the type (food and drink 13) of the acquired food and drink image (food and drink 13). While specifying the food and drink amount corresponding to the food and drink name), the total amount of the food and drink 13 provided in one eating and drinking act in the first eating and drinking space 11a or the eating and drinking act during the business hours of the day is totaled. Calculate (total amount calculation means).

The server 14 sets the food and drink name of the food and drink image transmitted from the first shooting camera 15a and the food and drink image determined by artificial intelligence, the food and drink amount of the food and drink name, and the shooting date and time as the first food and drink space identification identifier. , Stored (stored) in a large-capacity storage area in a state associated with the investor identifier (food and drink image storage means), and the total amount of food and drink 13 calculated by the total amount calculation means is used as the first food and drink space specific identifier and investment. Stored (stored) in a large-capacity storage area in a state associated with the house identifier (total amount storage means).

The server 14 multiplies the total amount of food and drink 13 in the first eating and drinking space 11a calculated by calculating the total amount by the margin rate, and calculates the margin (a part of the profit) to be returned to the investor who invested in the first eating and drinking space 11a. (Margin calculation means). The server 14 stores (stores) the margin calculated by the margin calculation means and the calculated date and time in the large-capacity storage area in a state associated with the first eating / drinking space specific identifier and the investor identifier (margin storage means). Food and drink names and food and drink prices and shooting dates and times provided in one eating and drinking activity in the first eating and drinking space 11a or during business hours of the day, total amount and calculated date and time of each food and drink, margin and calculated date and time, The investor name is printed by the printer. The total amount and margin (including the investor name) in the first eating and drinking space 11a can be calculated for each predetermined period such as daily, monthly, 10-day total, and 2-month total.

The server 14 determines which of the food and drink types learned by the learning means the food and drink image included in the second eating and drinking space image belongs to for each eating and drinking activity or for each eating and drinking activity during the business hours of the day. Let artificial intelligence make a judgment (food type judgment means, product type judgment means). The server 14 acquires the type (food and drink name) of the food and drink image (food and drink 13) of the second eating and drinking space image determined by the artificial intelligence from the artificial intelligence, and the type (food and drink 13) of the acquired food and drink image (food and drink 13). While specifying the food and drink amount corresponding to the food and drink name), the total amount of the food and drink 13 provided in one eating and drinking act in the second eating and drinking space 11b or the eating and drinking act during the business hours of the day is totaled. Calculate (total amount calculation means).

The server 14 sets the food and drink name of the food and drink image transmitted from the second shooting camera 15b and the food and drink image determined by artificial intelligence, the food and drink amount of the food and drink name, and the shooting date and time as the second food and drink space identification identifier. , Stored (stored) in a large-capacity storage area in a state associated with the investor identifier (food and drink image storage means), and the total amount of food and drink 13 calculated by the total amount calculation means is stored in the second food and drink space specific identifier and investment. Stored (stored) in a large-capacity storage area in a state associated with the house identifier (total amount storage means).

The server 14 multiplies the total amount of food and drink 13 in the second eating and drinking space 11b calculated by calculating the total amount by the margin rate, and calculates the margin (a part of the profit) to be returned to the investor who invested in the second eating and drinking space 11b. (Margin calculation means). The server 14 stores (stores) the margin calculated by the margin calculation means and the calculated date and time in the large-capacity storage area in a state associated with the second eating / drinking space specific identifier and the investor identifier (margin storage means). Food and drink names and food and drink prices and shooting dates and times provided in one eating and drinking activity in the second eating and drinking space 11b or during business hours of the day, total amount and calculated date and time of each food and drink, margin and calculated date and time, The investor name is printed by the printer. The total amount and margin (including the investor name) in the second eating and drinking space 11b can be calculated for each predetermined period such as daily, monthly, 10-day total, and 2-month total.

The server 14 determines which of the food and drink types learned by the learning means the food and drink image included in the third eating and drinking space image belongs to for each eating and drinking activity or for each eating and drinking activity during the business hours of the day. Let artificial intelligence make a judgment (food type judgment means, product type judgment means). The server 14 acquires the type (food and drink name) of the food and drink image (food and drink 13) of the third eating and drinking space image determined by the artificial intelligence from the artificial intelligence, and the type (food and drink 13) of the acquired food and drink image (food and drink 13). While specifying the food and drink amount corresponding to the food and drink name), the total amount of the food and drink 13 provided in one eating and drinking act in the third eating and drinking space 11c or the eating and drinking act during the business hours of the day is totaled. Calculate (total amount calculation means).

The server 14 sets the food and drink name of the food and drink image transmitted from the third shooting camera 15c and the food and drink image determined by artificial intelligence, the food and drink amount of the food and drink name, and the shooting date and time as the third food and drink space identification identifier. , Stored (stored) in a large-capacity storage area in a state associated with the investor identifier (food and drink image storage means), and the total amount of food and drink 13 calculated by the total amount calculation means is stored in the third food and drink space specific identifier, investment. Stored (stored) in a large-capacity storage area in a state associated with the house identifier (total amount storage means).

The server 14 multiplies the total amount of food and drink 13 in the third eating and drinking space 11c calculated by calculating the total amount by the margin rate, and calculates the margin (a part of the profit) to be returned to the investor who invested in the third eating and drinking space 11c. (Margin calculation means). The server 14 stores (stores) the margin calculated by the margin calculation means and the calculated date and time in the large-capacity storage area in a state associated with the third eating / drinking space specific identifier and the investor identifier (margin storage means). Food and drink names and food and drink prices and shooting dates and times provided in one eating and drinking activity in the third eating and drinking space 11c or during business hours of the day, total amount and calculated date and time of each food and drink, margin and calculated date and time, The investor name is printed by the printer. The total amount and margin (including the investor name) in the third eating and drinking space 11c can be calculated for each predetermined period such as daily, monthly, 10-day total, and 2-month total.

The server 14 determines which of the food and drink types learned by the learning means the food and drink image included in the fourth eating and drinking space image belongs to for each eating and drinking activity or for each eating and drinking activity during the business hours of the day. Let artificial intelligence make a judgment (food type judgment means, product type judgment means). The server 14 acquires the type (food and drink name) of the food and drink image (food and drink 13) of the fourth eating and drinking space image determined by the artificial intelligence from the artificial intelligence, and the type (food and drink 13) of the acquired food and drink image (food and drink 13). While specifying the food and drink amount corresponding to the food and drink name), the total amount of the food and drink 13 provided in one eating and drinking act or the eating and drinking act during the business hours of the day in the fourth eating and drinking space 11d is calculated. Calculate (total amount calculation means).

The server 14 sets the food and drink name of the food and drink image transmitted from the fourth shooting camera 15d and the food and drink image determined by artificial intelligence, the food and drink amount of the food and drink name, and the shooting date and time as the fourth food and drink space identification identifier. , Stored (stored) in a large-capacity storage area in a state associated with the investor identifier (food and drink image storage means), and the total amount of food and drink 13 calculated by the total amount calculation means is stored in the fourth food and drink space specific identifier, investment. Stored (stored) in a large-capacity storage area in a state associated with the house identifier (total amount storage means).

The server 14 multiplies the total amount of food and drink 13 in the fourth eating and drinking space 11d calculated by calculating the total amount by the margin rate, and calculates the margin (a part of the profit) to be returned to the investor who invested in the fourth eating and drinking space 11d. (Margin calculation means). The server 14 stores (stores) the margin calculated by the margin calculation means and the calculated date and time in the large-capacity storage area in a state associated with the fourth eating / drinking space specific identifier and the investor identifier (margin storage means). Food and drink names and food and drink prices and shooting dates and times provided in one eating and drinking activity in the fourth eating and drinking space 11d or during business hours of the day, total amount and calculated date and time of each food and drink, margin and calculated date and time, The investor name is printed by the printer. The total amount and margin (including the investor name) in the fourth eating and drinking space 11d can be calculated for each predetermined period such as daily, monthly, 10-day total, and 2-month total.

The sales amount totaling system 10 is provided while dividing the eating and drinking space 11 (sales floor space) of the restaurant 12 (store) into a plurality of first to fourth eating and drinking spaces 11a to 11d (first to nth eating and drinking spaces). The type of the food and drink 13 can be reliably identified by artificial intelligence, and the food and drink 13 provided in the first to fourth eating and drinking spaces 11a to 11d in one eating and drinking activity or the eating and drinking activity during the business hours of the day. The total amount of money can be calculated for each of the 1st to nth eating and drinking spaces 11a to 11d, and each eating and drinking space 11a in one eating and drinking activity of the restaurant 12 or in the eating and drinking activity during the business hours of the day. It is possible to accurately determine the quality of sales of food and drink every 11d.

The sales amount totaling system 10 calculates the accurate total sales amount of the food and drink 13 for each of the first to fourth eating and drinking spaces 11a to 11d, so that each of the first to fourth eating and drinking spaces 11a to 11d of the restaurant 12 is calculated. It is possible to accurately grasp the total sales amount of the food and drink space 11 and make modifications to improve sales and profits such as remodeling and rearrangement of the food and drink space 11 with reference to the total amount of money for each of the first to fourth eating and drinking spaces 11a to 11d. Can be considered.

The sales amount totaling system 10 recruits investors for each of the 1st to 4th eating and drinking spaces 11a to 11d (each of the 1st to 4th spaces) of the restaurant 12 (store), and each 1st to 4th eating and drinking space. Investors invest a predetermined amount in 11a to 11d, and a margin (a part of profit) is kicked to the investor according to the total sales amount of food and drink 13 for each of the first to fourth eating and drinking spaces 11a to 11d. It is possible to provide a mechanism for backing up (returning), and it is possible to construct a business model as a new method of collecting funds at the restaurant 12.

10 Sales amount totaling system 11 Eating and drinking space (sales floor space)
11a 1st eating and drinking space 11b 2nd eating and drinking space 11c 3rd eating and drinking space 11d 4th eating and drinking space 12 restaurant (store)
13 Food and drink (products sold and services provided)
14 Server 15a 1st shooting camera 15b 2nd shooting camera 15c 3rd shooting camera 15d 4th shooting camera 16 Internet 17 Container 18 Eating and drinking table 18a 1st eating and drinking table 18b 2nd eating and drinking table 18c 3rd eating and drinking table 18d 4th eating and drinking table 19 Router 20 digital camera

Claims (6)

In a sales amount aggregation system that aggregates the sales amount of products sold to visitors who visit a store that has a sales floor space of a predetermined area while using artificial intelligence.
The sales amount totaling system is formed of a server on which the artificial intelligence is set up and a shooting camera installed in the sales floor space.
The store is a restaurant that offers various kinds of food and drink.
The sales floor space is a food and drink space having a predetermined area for eating the food and drink.
The eating and drinking space is divided into a plurality of first to nth eating and drinking spaces.
The server
The artificial intelligence is trained so that the required various food and drink images can be recognized with high accuracy from the various sample images, and the recognized various food and drink images are various and multiple to which the food and drink images belong. Learning means to make the artificial intelligence learn so that it can be classified into food and drink types of
The 1st to nth eating and drinking space images including the food and drink images provided in one eating and drinking act in the 1st to nth eating and drinking space or the eating and drinking act during the business hours of the day by using the photographing camera are taken. Space image shooting means to shoot for each of the 1st to nth eating and drinking spaces ,
A product that causes the artificial intelligence to determine which of the food and drink types learned by the learning means belongs to the food and drink images included in the first to nth eating and drinking space images taken by the space image photographing means. Kind judgment means and
While specifying the amount of each food or drink determined by the product type determination means, the food or drink provided in one eating or drinking act in the first to nth eating or drinking spaces or in the eating or drinking act during the business hours of the day. It has a total amount calculation means for calculating the total amount of the total amount for each of the first to nth eating and drinking spaces.
The learning means
The artificial intelligence is trained so that various necessary container images can be recognized with high accuracy from various sample images.
The period from the time when the first container containing food and drink is provided in the first to nth eating and drinking spaces to the time when the container containing food and drink no longer exists in the first to nth eating and drinking spaces is described in 1. The artificial intelligence is made to learn to judge that it is a time of eating and drinking, and a container containing food and drink is provided to the first to nth eating and drinking spaces between the business start time and the business end time at the restaurant. A sales amount totaling system characterized in that the artificial intelligence learns such a case as eating and drinking during the business hours of the day . In a sales amount aggregation system that aggregates the sales amount of products sold to visitors who visit a store that has a sales floor space of a predetermined area while using artificial intelligence.
The sales amount totaling system is formed of a server on which the artificial intelligence is set up and a shooting camera installed in the sales floor space.
The store is a restaurant that offers various kinds of food and drink.
The sales floor space is a food and drink space having a predetermined area for eating the food and drink.
The eating and drinking space is divided into a plurality of first to nth eating and drinking spaces.
The server
The artificial intelligence is trained so that the required various food and drink images can be recognized with high accuracy from the various sample images, and the recognized various food and drink images are various and multiple to which the food and drink images belong. Learning means to make the artificial intelligence learn so that it can be classified into food and drink types of
The 1st to nth eating and drinking space images including the food and drink images provided in one eating and drinking act in the 1st to nth eating and drinking spaces using the photographing camera are obtained for each of the 1st to nth eating and drinking spaces. Space to shoot Image shooting means and
A product that causes the artificial intelligence to determine which of the food and drink types learned by the learning means belongs to the food and drink images included in the first to nth eating and drinking space images taken by the space image photographing means. Kind judgment means and
While specifying the amount of each food or drink determined by the product type determination means, the total amount of the food or drink provided in one eating or drinking act in the first to nth eating and drinking spaces is the total amount. It has a total amount calculation means for calculating each of the 1st to nth eating and drinking spaces.
The learning means
The artificial intelligence is trained so that various necessary container images can be recognized with high accuracy from various sample images.
The period from the time when the first container containing food and drink is provided in the first to nth eating and drinking spaces until the container containing food and drink disappears in the first to nth eating and drinking spaces is described in 1. A sales amount totaling system characterized in that the artificial intelligence is made to learn to judge that it is an act of eating and drinking once . In a sales amount aggregation system that aggregates the sales amount of products sold to visitors who visit a store that has a sales floor space of a predetermined area while using artificial intelligence.
The sales amount totaling system is formed of a server on which the artificial intelligence is set up and a shooting camera installed in the sales floor space.
The store is a restaurant that offers various kinds of food and drink.
The sales floor space is a food and drink space having a predetermined area for eating the food and drink.
The eating and drinking space is divided into a plurality of first to nth eating and drinking spaces.
The server
The artificial intelligence is trained so that the required various food and drink images can be recognized with high accuracy from the various sample images, and the recognized various food and drink images are various and multiple to which the food and drink images belong. Learning means to make the artificial intelligence learn so that it can be classified into food and drink types of
Using the photographing camera, the 1st to nth eating and drinking space images including the food and drink images provided in the eating and drinking activities during the business hours of the day in the 1st to nth eating and drinking spaces are captured by the 1st to nth eating and drinking spaces. Space image shooting means to shoot for each space,
A product that causes the artificial intelligence to determine which of the food and drink types learned by the learning means belongs to the food and drink images included in the first to nth eating and drinking space images taken by the space image photographing means. Kind judgment means and
While specifying the amount of each food and drink determined by the product type determination means, the total amount of the food and drink provided in the eating and drinking activities during the business hours of the day in the first to nth eating and drinking spaces is totaled. With a total amount calculation means for calculating for each of the first to nth eating and drinking spaces.
The learning means
The artificial intelligence is trained so that various necessary container images can be recognized with high accuracy from various sample images.
When a container containing food and drink is provided to the first to nth eating and drinking spaces between the business start time and the business end time at the restaurant, it is determined that the eating and drinking activity is performed during the business hours of the day. A sales amount totaling system characterized by having the artificial intelligence learn . In the learning means, various sample images of correct answers and sample images of various errors are input to the artificial intelligence, and the food and drink of the correct answer is obtained from those sample images of the correct answer and those sample images of the error. The sales amount totaling system according to any one of claims 1 to 3, wherein the artificial intelligence is made to learn an image so that the image can be recognized with a predetermined correct answer rate. The fourth aspect of the present invention, wherein in the learning means, the correct answer rate of the artificial intelligence that recognizes the correct food and drink image from those sample images of the correct answer and those sample images of the error is in the range of 80 to 100% . Sales amount totaling system. The sales amount according to claim 4 or 5, wherein in the learning means, the artificial intelligence removes unnecessary images other than food and drink images from various sample images and acquires various necessary food and drink images. Aggregation system.

Images