JP7806993B2 - Menu creation support device - Google Patents

Description

The present invention relates to a menu creation support device and a menu creation support method that are useful when planning a menu.

Planning a meal menu is as important now as it has always been, but it is also a time-consuming and laborious task. In the past, people would choose their meals by looking at cooking books and magazines, but recently, by simply searching for the name of a dish, a variety of recipes from various sources are presented online, making it easier than ever to plan a meal menu.

For example, Japanese Patent Application Laid-Open No. 2003-122860 (Patent Document 1) describes a technology that prompts a user who desires menu information to input the conditions for the menu information they desire via a communication line as search conditions (ingredients conditions), searches pre-registered menu information based on the search conditions entered by the user, and provides the user with the search results for menu information.

Japanese Patent Application Laid-Open No. 2003-122860

However, although the invention described in the above Patent Document 1 provides a desired menu that meets the user's requests, such as ingredients and needs, the content is fixed and not extensible, and when the desired dish is not found, it is necessary to start the search again from the beginning, such as by changing the search perspective. Also, although the dish candidates provided reflect the user's search conditions, they merely display dishes that match those conditions.
Therefore, the present invention provides a technology that can provide not only search results such as menus that match the conditions indicated by the user, but also search results that include additional suggestions to the user.

In order to solve the above problems, one embodiment has the following configuration.
In other words, it is a menu creation support device that includes a storage means for recording information such as ingredient data, dish data, and user data, a communication means for exchanging information with the user, a display means for displaying a request for ingredient input to the user and displaying dishes using the ingredients input by the user and compatible ingredients that are ingredients that go well with those ingredients, and a control means for controlling the various means.

In one embodiment, the device has storage means for recording information such as ingredient data, cooking data, and user data, and can provide dishes and recipes that are suitable for the user based on this information. It has communication means for sending and receiving information to and from the user, so information can be exchanged between the device and users in remote locations. It has display means that prompts the user to input ingredients and displays dishes that use the ingredients entered by the user and ingredients that go well with those ingredients, so it can prompt the user to enter the necessary information and display the information required by the user. It has control means for controlling the various means mentioned above, namely the storage means, communication means, and display means, and can use these means to provide the user with dishes that go beyond the ingredients originally entered.

In one embodiment, the display means is a menu creation support device that displays the compatible ingredients and prompts the user to input an ingredient selected from the compatible ingredients.
In one embodiment, the display means displays the compatible ingredients, so that it is possible to present the user with additional compatible ingredients in addition to the ingredients input by the user.Then, the display means displays a request for the user to input an ingredient selected from the compatible ingredients, so that the user can be asked to select an ingredient from among the compatible ingredients, and a response that is more in line with the user's request can be provided.

In one embodiment, the user data may be information selected from personal information such as user identification information such as user ID and user address, information regarding vital data such as gender, age, height, weight, BMI, pulse rate, body temperature, and blood pressure, allergies, chronic illnesses, medical history, health guidance history, treatment history, commonly used medications, dietary restrictions, preferences, and app usage logs.

In one embodiment, the user data is information selected from personal information such as user identification information such as user ID and user address, information related to vital data such as gender, age, height, weight, BMI, pulse rate, body temperature, and blood pressure, allergies, chronic illnesses, medical history, health guidance history, treatment history, commonly used medications, dietary restrictions, preferences, and app usage logs. Therefore, it is possible to provide results that take such user data into consideration as to which ingredients go well together. This makes it possible to present preferred dishes that match the user's individual condition.

One embodiment can be a menu creation support method that executes a first input request step that prompts the user to input ingredients; a compatible ingredient display step that displays the ingredients input by the user and compatible ingredients that go well with the ingredients; a second input request step that prompts the user to input ingredients selected from the compatible ingredients; and a dish display step that displays dishes using the ingredients input in the first input request step and the second input request step.

In one embodiment, the menu creation support method executes a first input request step that prompts the user to input ingredients; a compatible ingredient display step that displays the ingredients input by the user and compatible ingredients that go well with those ingredients; a second input request step that prompts the user to input ingredients selected from the compatible ingredients; and a dish display step that displays dishes using the ingredients input in the first input request step and the second input request step.This makes it possible to present ingredients that go well with the ingredients initially input.The user can then further select their preferred ingredients from the compatible ingredients, making it possible to present dishes that meet the user's requests.

In one embodiment, the compatibility may be determined based on the frequency with which the ingredients are used together in a dish.
In one embodiment, the compatibility is determined based on the frequency with which ingredients are used together in a dish, taking into account the fact that ingredients have been used together in many dishes in the past, and making it possible to provide the user with highly versatile suggestions that incorporate the knowledge and experience of predecessors.

In one embodiment, the compatibility may be determined by taking into consideration the ingredients of the ingredients.
In one embodiment, the compatibility is determined by taking into account the components of the ingredients. Therefore, it can be determined based on, for example, the variation in the nutritional components of the ingredients. Taking such nutritional components into consideration can lead to the provision of nutritionally balanced dishes. Furthermore, the compatibility can be determined by taking into account the nutritional components of the ingredients as well as the frequency with which the ingredients are used together in a dish, or it can be determined based on the nutritional components of the ingredients independently of the frequency with which the ingredients are used together in a dish. In the former case, the compatibility can reflect both the frequency of occurrence and the merits of the nutritional components. On the other hand, in the latter case, it is possible to provide compatibility that focuses on the nutritional components regardless of the frequency of occurrence. In addition to nutritional components such as protein and lipid, ingredients in ingredients include components that may not necessarily be considered nutrients, such as water and dietary fiber, and color, taste, aroma, etc. can also be considered as components.

In one embodiment, the compatibility may be determined taking into consideration the characteristics of the user.
In one embodiment, the compatibility is determined by taking into account the user's characteristics. For example, the compatibility can be determined by taking into account information contained in user data, such as the user's allergies, dietary restrictions, and preferences. By taking such user-specific factors into account, it is possible to provide dishes that are personalized to each user and more tailored to their needs. Furthermore, the compatibility can be determined by taking into account the user's characteristics as well as the frequency with which ingredients are used together in a dish, or by taking into account the user's characteristics independently of the frequency with which ingredients are used together in a dish. In the former case, the compatibility can reflect the merits of both the frequency of occurrence and the user's characteristics. On the other hand, in the latter case, it is possible to provide compatibility that focuses on the user's characteristics regardless of the frequency of occurrence.

According to one embodiment, it is possible to present ingredients that go well with the ingredient selected by the user, and it is possible to present the ingredients with the added benefit of good compatibility.
According to one embodiment, a dish that meets the user's request can be presented from ingredients selected by the user.

1 is a block diagram of a menu creation support system according to an embodiment. 1 is a flowchart showing the flow of a menu creation support method according to one embodiment.

The present invention will be described in further detail based on the following embodiments. Note that duplicated descriptions of identical parts, materials, functions, effects, etc. will be omitted in each embodiment.

First embodiment:
1 is a block diagram showing the configuration of a menu creation support system 10 in which a menu creation support device 11 according to one embodiment is connected to a user terminal 12 used by a user via a network 13. This menu creation support device 11 includes a display means 1, a storage means 2, a communication means 3, and a control means 4.

The display means 1 creates and displays an input request screen that requests the user to input ingredients of their choice, a screen that displays search results, dishes suitable for the user, and the like.
The storage means 2 stores information including various data necessary for presenting menus, such as ingredient data, dish data, and user data.
The communication means 3 exchanges information with the user.
The control means 4 controls the entire device, communicates with the user, and executes programs and the like stored using various information recorded in the storage means 2.
The user terminal 12 used by the user displays the display screen created by the display means 1, stores the user's personal data, and performs the necessary input to be transmitted to the menu creation support device 11 by the user.

The menu creation support device 11 is configured to include a central processing unit (CPU) that functions as a calculation device or control device, a main storage device such as RAM or flash memory, an external storage device such as a hard disk, a communication device such as a LAN card or router, computer programs, etc., and may be a virtual server on a cloud service connected to a network or a computer used by a system manager.

The user terminal 12 used by the user is equipped with a communication device capable of connecting to and communicating with a network such as an in-house LAN, the Internet, or a LAN, and may be a personal computer, a mobile phone, a tablet device, a mobile device capable of recording vital data, or the like, which can retrieve and display data stored in the menu creation support device 11 when a program is started and connected to the network, send data entered by the user to the menu creation support device 11, and receive data from the menu creation support device 11 and output it in the form of email or the like via an output device.

<Execution of menu creation support device 11>:
In one mode of use, a user accesses the web page of the menu creation support device 11 from the user terminal 12, enters his or her own ID, and starts using the device. The subsequent execution flow of the menu creation support device 11 is shown in the flowchart of FIG. 2.

Step 1: Presentation of ingredient input screen When the control means 4 grants the user access, the display means 1 displays an input screen with an input field for ingredients to be used in creating a dish. The user inputs a desired ingredient (e.g., ingredient 1) in the input field of the input screen. The desired ingredient may be an ingredient that the user owns, an ingredient that the user wants to cook, or the like.
This ingredient input field can be configured to allow the user to freely input ingredients. Alternatively, the display means 1 can be configured to present a list of ingredients in the input field, allowing the user to select from the list. Furthermore, if the user has previously input ingredients, the display means 1 can use that history information to display ingredients that the user has previously selected, making it easier for the user to select their preferred ingredients.
The user may input one type of ingredient or two or more types of ingredients.

Ingredients include staple foods such as rice, udon, and soba, as well as main and side dishes such as types of fish, meat, beans, eggs, vegetables, potatoes, mushrooms, and seaweed, as well as fruits, milk, and dairy products. In addition to these, seasonings, additives, processed foods, and the like can also be considered ingredients.
The display unit 1 is preferably configured to display multiple input fields for ingredients. When the user inputs multiple ingredients into one input field, an OR search can be performed, and when the user inputs multiple ingredients into multiple input fields, an AND search can be performed.

In addition to inputting an ingredient, the user may also be able to instruct the cancellation of step 3 (described below), which presents ingredients that go well with the selected ingredient. The display means 1 may display, in addition to the ingredient input field, an input request field required to cancel the screen presenting ingredients that go well with the selected ingredient.

Step 2: Searching for compatible ingredients (good match ingredients) When the user finishes inputting ingredients, the control means 4 searches the database stored in the storage means 2 based on the ingredients (e.g., ingredient 1) input by the user for ingredients that go well with the ingredients input by the user, in other words, ingredients that can be used together with the ingredients selected by the user to create a dish suitable for the user. Alternatively, the control means 4 performs machine learning to search for ingredients that go well with the ingredients input by the user based on various information stored in the storage means 2. For ease of explanation, compatible ingredients are also referred to as "good match ingredients."
The compatibility between ingredients is expressed as a number between 0 and 100% that indicates how well or poorly the ingredients are combined, with 100% being the best and 0% being the worst. The method for determining compatibility will be described later.
When performing an AND search on multiple ingredients entered by a user, ingredients that go well with those multiple ingredients are searched for. In other words, if a user enters n ingredients, +1 ingredient that goes well with the n+1 ingredients is searched for. Also, when performing an OR search on multiple ingredients entered by a user, ingredients that go well with each of those multiple ingredients are searched for.

Step 3: Presentation of compatible ingredients Next, the display means 1 presents to the user an appropriate number of compatible ingredients from the multiple compatible ingredients identified by the control means 4. The number of ingredients presented can be the limit number of ingredient photos that can be displayed on the output screen presented as search results, or the number of ingredients within a predetermined compatibility threshold. Furthermore, it is preferable to present these ingredients in order of compatibility. Furthermore, it is preferable to present the ingredients in order of compatibility, displaying a compatibility number along with the photo.
Ingredients that are considered to go well together include seasonings, additives, processed foods, etc., but it is also possible to exclude these ingredients from the ingredients to be displayed.
The user then selects and inputs an ingredient that the user would like to use in a dish together with the ingredient that was initially input from among the ingredients presented as being compatible with the ingredient (for example, ingredient 2).

Step 4: Presenting dishes using ingredients that go well with the ingredients entered by the user Once the user has finished entering the first ingredient entered (e.g., ingredient 1) and any ingredient (e.g., ingredient 2) from among the ingredients that are subsequently determined to go well together, the control means 4 searches the data stored in the memory means 2 for dishes that use the ingredients selected by the user (here, ingredient 1 and ingredient 2), and the display means 1 presents dishes that meet the conditions.
Alternatively, if step 3 is cancelled, the control means 4 searches for dishes using ingredients that go well with the ingredients initially input by the user, and the display means 1 presents dishes that meet the conditions.

In addition to the dishes displayed in step 4, a portion of the display screen that displays dishes can also display ingredients that go well with the ingredients initially selected by the user but were not selected in step 3 (e.g., ingredients other than ingredient 2). Therefore, even if dishes using ingredients selected by the user are displayed in step 4, if the user does not want that dish and would like to find another dish, they can quickly select other ingredients that go well together.

Step 5: Presentation of Recipe When the user selects a desired dish from the multiple dishes displayed in step 4, the display means 1 displays a recipe showing the ingredients needed for that dish, their quantities, how to prepare the dish, etc. The user can then cook the selected dish according to the recipe.

This menu creation support device 11 can present ingredients that go well with the ingredients initially selected by the user, and can also present dishes that use ingredients that go well with the ingredients initially selected by the user, thereby expanding the range of dishes that can be suggested and providing users with an added bonus when planning their menu.

<How to determine compatibility>:
There are several examples of how to determine compatibility, and these will be explained below, but any of the compatibility determined by the method determined here may be adopted.

Compatibility determination 1:
Based on the idea that if ingredients go well together, they will be used in the same dish, one embodiment of compatibility determination is based on the frequency with which the ingredients are used together in a dish. One specific method of determination is as follows.

First, information on various dishes is collected, and the ingredients used in each dish are extracted. Ingredients can be extracted by natural language processing from the text describing the dish, or by extracting ingredient names from the shape and color of the food image.
Next, the frequency of occurrence of each collected ingredient is counted. If an ingredient is used in one dish, it is counted as "1." If an ingredient is used in 231 different dishes, it is counted as "231."

The frequency with which one ingredient appears in the same dish as another ingredient is also counted. For example, if there are 58 dishes in which ingredients A and B both appear, the co-occurrence frequency of ingredients A and B is counted as "58."

Table 1 is an example showing the relationship between a certain ingredient and another ingredient. In Table 1, the "Ingredient 1 [Count])" column indicates that the ingredient is "tomato" and the number of dishes in which "tomato" is used, while the "Ingredient 2 [Count])" column shows various ingredients and the number of dishes in which they are used. The "Co-occurrence" column indicates the number of dishes in which the ingredient displayed in the "Ingredient 2 [Count])" column is used together with the ingredient "tomato," and the "Pairing Score" column shows the correlation between "tomato" and the ingredient displayed in the Ingredient 2 column.

Here, the correlation between two ingredients indicated by the "pairing score" in Table 1 can be calculated as follows.
First, an embedding model of ingredient names from all the recipe data is created, and an N-dimensional numerical vector representation is generated for each ingredient. A deep learning Siamese network is then used. The Siamese network has two top-level multilayer perceptrons (MLPs), each with the same number of layers, nodes, and weights. The N-dimensional numerical vector representations of ingredients are input to the Siamese network, which outputs an M-dimensional numerical vector. The two M-dimensional vectors are first concatenated and then input to another MLP with a different number of layers, nodes, and weights than the two top-level MLPs, which outputs a K-dimensional vector. Next, the cross product of the two M-dimensional vectors is obtained to generate an M ^2- dimensional vector. The M ^2- dimensional vector is concatenated with the K-dimensional vector to form an M ² + K-dimensional vector. Finally, the M ² + K-dimensional vector is input to another MLP with a one-dimensional scalar output to obtain the desired correlation between ingredients.
Finally, the correlation between the two ingredients was standardized to obtain a pairing score ranging from -100% to 100%, with 100% representing the best pairing and -100% representing the worst pairing.

In this way, compatibility is determined by taking into account the frequency of co-occurrence of ingredients and the specific frequency of occurrence of the ingredients themselves. Compatibility determined by this compatibility determination method is based on the empirical finding that in many dishes that have existed in the past, ingredients go well together when used together.

Compatibility determination 2:
In one embodiment of the compatibility determination method, the compatibility between two ingredients is determined, but the compatibility determination method here determines the compatibility between three or more ingredients. More specifically, when determining the compatibility between three ingredients, the compatibility determination method between two ingredients is extended and applied to three ingredients. The same applies when determining the compatibility between four or more ingredients.

Compatibility determination 3:
Another embodiment of compatibility determination is a method of determining compatibility by taking into account the ingredients of the ingredients between the ingredients. While the compatibility determination methods described above are based on the criterion that the more frequently ingredients are used together in a dish, the better the compatibility. However, this method is different from that idea and is based on the idea that the more nutritionally balanced an ingredient is in a dish, the better the compatibility.
In one aspect of the method for considering nutritional data in determining food compatibility, ingredients based on food composition tables are extracted from the ingredients, and compatibility is determined by focusing on the extent to which the variation in nutritional components, particularly when one ingredient is combined with another, is reduced compared to when either ingredient is consumed alone (i.e., when the ingredients are consumed alone). For example, when comparing rice and pork, the pork compensates for the lack of protein in rice alone, and the rice compensates for the lack of carbohydrates in pork alone, resulting in less variation in nutritional components than when either ingredient is consumed alone. This results in a good compatibility between these ingredients. On the other hand, when comparing pork and beef, the nutritional components contained in these ingredients are similar, and the variation in nutritional components cannot be reduced compared to when pork or beef are consumed alone, resulting in a poor compatibility. Note that while protein and carbohydrates are used as examples of ingredients here, the same applies to other ingredients such as dietary fiber, water, and color.

It is more preferable to add results based on nutritional data to the compatibility based on frequency of occurrence in dishes, as explained in Compatibility Determination 1 and Compatibility Determination 2. This is because the ingredients of the ingredients are taken into consideration in addition to the compatibility based on frequency of occurrence. In addition, if ingredients B and C are used in dish A, a more realistic compatibility can be calculated by extracting the amounts of ingredients B and C used in dish A and using the variation in ingredients based on the nutritional content of those amounts as a factor in determining compatibility.

Compatibility determination #4:
Another embodiment of the compatibility determination method is a method of determining compatibility by taking into account the characteristics of the user. In other words, this method determines compatibility by taking into account user data in addition to data related to the contents of the menu, such as ingredient data or dish data.
As data related to the menu contents, ingredient data includes information such as the names of ingredients, the amounts used, nutritional data of the ingredients, allergy data, frequency of occurrence in the dish, compatibility between ingredients, etc. Dish data includes information such as the name of the dish, the recipe details of the dish, a photo of the dish, and nutritional data such as calories of the dish. User data includes user identification information such as user ID and user address, information related to vital data such as gender, age, height, weight, BMI, pulse rate, body temperature, and blood pressure, personal information such as allergies, chronic illnesses, medical history, health guidance history, treatment history, commonly used medications, dietary restrictions, preferences, and app usage logs.

The control means 4 uses machine learning to analyze the correlations between these various disparate data, and from there, it identifies the elements that are important for maintaining the user's health and for the user's needs, determining the order of importance and compatibility. In addition, items with known strong correlations are dealt with individually. Since there is a strong correlation between a user's allergies and the ingredients that cause them, for example, if the user's personal data indicates that they have a wheat allergy, wheat will be excluded from the suggested ingredients.

By considering compatibility based on the user's characteristics, ingredients can be presented in advance to users with dietary restrictions or allergies based on the user's specific conditions, reducing the risk of violating dietary restrictions or mistakenly selecting an ingredient that causes an allergy. Furthermore, if a user accesses the menu creation support device 11 after engaging in a long period of aerobic exercise, the user's exercise data can be stored in the storage means 2 via the communication means 3 from a mobile device that stores the user's exercise data, making it easier to search for ingredients high in carbohydrates as compatible ingredients based on the user's temporary energy deficiency.

If there is some publicly known compatibility data, such as data on compatibility between ingredients based on their frequency of occurrence, the control means 4 can store this in the storage means 2. Furthermore, if there is no publicly known compatibility data, for example, if there is a new variety of ingredient that has never been produced before, an ingredient listed in a new recipe presented by a new supplier, or a new data item that has been added to the user data, the control means 4 can recalculate new compatibility each time and update the data sequentially. Here, the control means 4 can also perform machine learning to search for and present the latest compatibility data based on the latest data.

Second embodiment:
In step 4 of the first embodiment, dishes using ingredients that go well with the ingredients entered by the user are presented, but the presented dishes include multiple dishes that contain both the ingredients entered by the user and ingredients that go well with the ingredients.
However, when the compatibility between ingredients is determined by taking into account the nutritional components of the ingredients or the user's characteristics, simply presenting dishes containing the data entered by the user may not be sufficient. For example, even if the user selects only one of the presented dishes, that dish alone may not necessarily provide sufficient nutrition. It may be that the user also selects other dishes and combines them into a single menu to obtain sufficient nutrition. In such cases, the display means 1 can present the multiple dishes in pairs.

When determining the compatibility of ingredients, the control means 4 uses compatibility determined by taking into account the nutritional components of each ingredient, and selects a combination of ingredients that provides sufficient nutrition based on the ingredients selected by the user, and presents dishes using such ingredients. For example, if the user selects ingredients A and B, and the combination of ingredients A, B, C, D, and E provides sufficient nutrition, dishes that do not contain the ingredients entered by the user can be presented, such as a dish using ingredients A and B and a dish using ingredients C, D, and E, or a dish using ingredients A and D and a dish using ingredients B, C, and E. Of course, the same ingredient may be included in multiple dishes. In this case, it is preferable to select dishes consisting of combinations of ingredients A, B, C, D, and E taking into account the compatibility based on the frequency of their use in the same dish.

When using compatibility determined taking into account the user's characteristics, the control means 4 also selects a combination of ingredients based on the user's selected ingredients, taking into account the user's health and other values desired by the user, and presents dishes using those ingredients. For example, if the user selects ingredients F and G, and the combination of ingredients F, G, H, I, and J results in a dish that meets the user's desired values, the control means 4 can present dishes that do not contain the ingredients entered by the user, such as a dish using ingredients F and H and a dish using ingredients G, I, and J, or a dish using ingredients F and I and a dish using ingredients G, H, and J. Of course, the same ingredient may be included in multiple dishes. It is preferable to select dishes consisting of combinations of ingredients F, G, H, I, and J taking into account compatibility based on the frequency of their use in the same dish.

When this embodiment is adopted, the display means 1 creates a display screen with an input field for requesting the display of recommended dishes. When the user requests that the system present recommended dishes, a combination of recommended dishes that meets the request is presented.

The above embodiment is an example of the present invention, and may further include known means and methods, and may be modified as appropriate within the scope of the present invention.

REFERENCE SIGNS LIST 1 display means 2 storage means 3 communication means 4 control means 10 menu creation support system 11 menu creation support device 12 user terminal 13 network

Claims (3)

a storage means for recording information on ingredient data, dish data, and user data;
A communication means for transmitting and receiving information to and from a user terminal;
a display means for displaying a request for inputting ingredients on a user terminal and displaying a recipe using the ingredients input by the user and ingredients that go well with the ingredients;
A control means for controlling the various means;
Equipped with
The compatibility is determined based on the frequency with which ingredients are used together in a dish, and is expressed as an N-dimensional numerical vector for each ingredient, which is obtained as an M-dimensional vector using a multi-layer perceptron, a K-dimensional vector obtained by concatenating two of the M-dimensional vectors and using another multi-layer perceptron, an M ^2- dimensional vector obtained by the cross product of the two M-dimensional vectors, and an M ² +K-dimensional vector obtained by concatenating the M ² -dimensional vector and the K-dimensional vector, and the compatibility is standardized by the correlation obtained from the M ² +K-dimensional vector using another multi-layer perceptron. 2. The menu creation support device according to claim 1 , wherein the display means displays the compatible ingredients and prompts the user to input an ingredient selected from the compatible ingredients. The menu creation support device described in claim 1 or claim 2, wherein the user data is information selected from user identification information such as user ID and user address, information regarding vital data such as gender, age, height, weight, BMI, pulse, body temperature, and blood pressure, and personal information such as allergies, chronic illnesses, medical history, health guidance history, medical history, commonly used medications, dietary restrictions, preferences, and app usage logs.