Finished dish nutrient content accurate real-time calculation method based on model

Abstract

The invention provides a model-based finished dish nutrient content accurate real-time calculation method. The food raw materials and cooking data are obtained through front-end interaction, the nutrient intake is accurately calculated in real time, the nutrient intake is compared with a recommended value of a dietary pagoda for evaluation, food category intake is counted in a classified mode, and prompts are generated. The requirement of modern healthy diet for accurate evaluation of nutritional ingredients of dishes is met, and healthy catering development is promoted.

Description

Technical Field

[0001] The present invention relates to the field of catering information technology, and particularly relates to a method for accurately and real-time calculating the nutrient content of finished dishes based on a model. Background Art

[0002] With the continuous growth of the demand for healthy diets, accurately obtaining the nutrient information of finished dishes has become an important technical requirement in the fields of food processing, catering services, and medical health. There are mainly two technical bottlenecks in the existing technical solutions: First, the manual calculation method based on the food composition table has significant efficiency defects, and its static data model is difficult to dynamically reflect the differential effects of different cooking methods on nutrients; Second, the nutrition assessment system based on image recognition faces multiple technical barriers, specifically manifested as insufficient accuracy in identifying mixed ingredients, the lack of a quantification mechanism for nutrient loss during the cooking process, and the lag in the construction of the ingredient database for Chinese-style compound seasonings, resulting in the inability to accurately calculate the nutritional contribution values of special seasonings such as soy sauce and broad bean paste. These technical defects severely restrict the practical value and promotion potential of the nutrition assessment system. Summary of the Invention

[0003] The purpose of this application is to provide a method for accurately and real-time calculating the nutrient content of finished dishes based on a model, including the following steps: obtaining the dish raw material information input by the user, including the types and usage information of ingredients and seasonings. If the cooking method is not input by the user, the predicted cooking method is output through the dish raw material information; converting the ingredient names into the standard names in the standard database, and generating an ingredient weight vector based on the ingredient usage; inputting the preprocessed data into a supervised learning model, and the model calculates the nutrient content through the following formula: Wherein, N k is the content of nutrient k in a certain dish, W i is the weight of the i-th ingredient, C ik is the original content value of nutrient k per unit mass of the i-th ingredient, f j is the optimization coefficient of nutrient k in the i-th ingredient corresponding to cooking method j. A standardized report containing multiple nutrients is generated in real time.

[0004] In one embodiment, it further includes establishing an ingredient alias mapping table to convert non-standard names into the standard names in the standard database; performing mass-volume conversion on seasonings according to the preset density parameters.

[0005] In one embodiment, the cooking method parameters include frying and baking, boiling, deep frying, steaming, stir-frying, stewing, etc., and the corresponding optimization coefficients are pre-determined through experimental data and stored in the database.

[0006] In one embodiment, the constraint is specifically implemented as: the optimization coefficient is limited to the range of 0 to 1.

[0007] In one embodiment, it further includes outputting a real-time standardization report, and the real-time standardization report includes: the core index values of energy, protein, fat, and carbohydrates and their percentages of the daily recommended intake; warning signs for the contents of cholesterol and sodium (if exceeding the preset threshold); and visual annotations on the degree of nutrient loss based on the cooking method.

[0008] Further, the preset threshold can be set according to the following criteria: (1) Based on the national dietary guidelines: According to the age and gender information of the target user group, call the corresponding daily nutrient recommended intake (such as the Dietary Reference Intakes for Chinese Residents, DRIs), and set the warning threshold according to the proportion of a single serving of the dish; (2) Based on the user's health data: When the user associates a personal health record, automatically lower the warning thresholds for sodium and cholesterol according to their disease types (such as hypertension, high cholesterol); (3) Dynamic learning and adjustment: Calibrate the threshold in a feedback manner according to the error rate between the actual detected value and the predicted value of the nutrients of the same type of dishes in the historical data.

[0009] In one embodiment, the data preprocessing further includes: establishing a compensation mechanism for the processing state of ingredients. For the fresh / dried / frozen state of ingredients, different moisture compensation coefficients are used to dynamically correct the ingredient weight vector, and the dried ingredients are restored to the original edible part weight through the rehydration rate parameter.

[0010] The present invention also provides an embodiment, including a device, an input module configured to receive ingredient information and cooking method parameters; a data processing module configured to perform data cleaning, name standardization, and vectorization processing; a model operation module storing a supervised learning model and a pre-set cooking method influence coefficient database; and an output module generating a visual nutrient content report interface.

[0011] In one embodiment, the data processing module is further configured to be able to identify and replace the ingredient names in the user input; and provide a list of standardization suggestions for the unrecognized ingredient names for the user to select.

[0012] In one embodiment, the model operation module includes a standard food composition database and a cooking method influence factor control database. The cooking method influence factor control database includes a cooking method influence factor comparison table, and the cooking method influence factor comparison table contains a variety of Chinese cooking combination data.

[0013] In one embodiment, the output module integrates a nutritional intake assessment function. By calling the daily recommended intake standards in the user health record database, a visual progress bar indicating the degree of nutrient compliance and warning signs for abnormal nutrients are generated in the standardized report.

[0014] Beneficial effects

[0015] The method for accurately and real-time calculating the nutrient content of finished dishes based on a model in the present invention, compared with the traditional manual calculation method based on food composition tables, can quickly process a large amount of data. By considering the effects of different cooking methods on nutrients and precisely processing ingredients and seasonings, it greatly improves the calculation efficiency and accuracy, and controls the error within a very small range. At the same time, it avoids the problems faced by the nutrition assessment system based on image recognition, such as insufficient recognition accuracy of mixed ingredients, lack of a quantification mechanism for nutrient loss during cooking, and lag in the construction of a database of Chinese compound seasoning ingredients, improving the reliability and practicality of nutrition assessment. It can also generate a standardized report containing multiple nutrients in real time and integrate a nutritional intake assessment function, enabling users to clearly understand the nutrient content of the dishes and the comparison with the daily recommended intake, facilitating the reasonable arrangement of diet. For example, in the nutrition report of braised pork ribs, users can intuitively know whether nutrients such as protein, fat, and sodium meet the standards, thus better controlling their dietary health. This method and device can be widely applied in food processing enterprises to accurately control the nutritional components of products, in the catering service industry to provide accurate nutritional information to customers, and in the medical and health field to assist doctors in formulating personalized diet plans for patients, helping patients with diseases such as diabetes and hypertension recover and manage their health, and providing strong support for the development of related fields. Brief description of the drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on the structures shown in these drawings.

[0017] Figure 1 It is a working step diagram of a method for accurately and real-time calculating the nutrient content of finished dishes based on a model provided by the first embodiment of the present invention;

[0018] Figure 2 It is a setting module diagram provided by the second embodiment of the present invention. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0020] It should be noted that if there are directional indications (such as up, down, left, right, front, back...) involved in the embodiments of the present invention, the directional indications are only used to explain the relative position relationship and movement conditions between components in a specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0021] In addition, if there are descriptions such as "first" and "second" involved in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, if "and / or" or "and / or" appears throughout the text, its meaning includes three parallel scenarios. Taking "A and / or B" as an example, it includes scenario A, scenario B, or the scenario where A and B are satisfied simultaneously. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the fact that those of ordinary skill in the art can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0022] Embodiment 1

[0023] The present invention relates to the field of catering information technology, and particularly relates to a method for accurately and real-time calculating the nutrient content of finished dishes based on a model. Through this method, the nutrient content in finished dishes can be calculated quickly and accurately, providing important data support for multiple fields such as food processing, catering services, and medical health.

[0024] Reference Figure 1 - Figure 2 , the purpose of this application is to provide a method for accurately and real-time calculating the nutrient content of finished dishes based on a model, including the following steps: obtaining the dish raw material information input by the user, including the types of ingredients, usage amounts, and condiment information, and outputting predicted cooking method parameters through the dish raw material information; converting the ingredient names into standard names in the standard database and generating an ingredient weight vector based on the ingredient usage amounts; inputting the preprocessed data into a supervised learning model, and the model calculates the nutrient content through the following formula:

[0025] Where Nk is the content of nutrient K in a dish, W i is the weight of the i-th ingredient, C ik is the original content value of nutrient k per unit mass of the i-th ingredient, f j is the optimization coefficient of nutrient k in the i-th ingredient corresponding to cooking method j; the model satisfies the constraint condition: the optimization coefficient is limited to a range of 0 to 1.

[0026] As people pay more and more attention to healthy eating, it is increasingly important to accurately obtain the nutrient information of finished dishes. Under current technical conditions, there are many problems that need to be solved. The existing manual calculation method based on food composition tables is extremely inefficient. The data in the food composition table is often static and cannot be dynamically adjusted according to the differentiated effects of different cooking methods on nutrients. For example, for potatoes, the two different cooking methods of frying and steaming will cause a great change in the nutrient content in potatoes, but the traditional manual calculation method cannot accurately reflect this change. The nutrition assessment system based on image recognition also faces many technical difficulties. When identifying mixed ingredients, the accuracy is difficult to be effectively guaranteed. A dish may contain multiple ingredients. Due to the mixed state of the ingredients and the limitations of the image recognition technology itself, misjudgment often occurs. In addition, during the cooking process, the quantification mechanism of nutrient loss is missing, and the specific impact of cooking on nutrient content cannot be accurately known. At the same time, for Chinese compound seasonings, such as soy sauce and bean paste, the construction of ingredient databases lags behind. The contribution of these special condiments to the nutritional content of dishes is difficult to calculate accurately, which seriously affects the accuracy and practicality of the nutritional assessment system and also limits its promotion in practical applications.

[0027] The present invention provides a model-based method for accurately and in real time calculating the nutrient content of finished dishes to solve the above-mentioned problems existing in the prior art. The method specifically comprises the following steps: obtaining user input information and predicting cooking method parameters: obtaining dish ingredient information input by the user, which includes the type of ingredients, amount and condiment information. Through these dish ingredient information, the system can output predicted cooking method parameters. For example, the user inputs information about potatoes, beef, and condiments such as soy sauce and salt. The system can predict the possible cooking methods as steaming, frying and braising based on the combination of these ingredients and condiments, combined with big data analysis or preset rules.

[0028] When the user does not specify a cooking method, the following prediction process is executed: Feature extraction, extracting ingredient combination features (such as "pork + green pepper + soy sauce"), condiment types (liquid / solid / paste) and dosage ratios from the input data. Model prediction, inputting the feature vector into a pre-trained random forest classification model to output the probability distribution of cooking methods. The model training data comes from the cooking method annotation data of 100,000 Chinese recipes, and the features include ingredient combinations, condiment types, total liquid content, etc. Finally, select the cooking method with the highest probability (such as "steaming / frying / stewing") as the prediction result. Establish an ingredient alias mapping table to convert non-standard names input by the user into standard names in the database. For example, if the user inputs "yam", it can be converted into the standard name "potato" through the alias mapping table. The purpose of this is to ensure that in subsequent data processing and calculations, the corresponding ingredient information can be accurately obtained from the standard database. Ingredient dosage processing, generating an ingredient weight vector based on the ingredient dosage. At the same time, perform mass-volume conversion on the condiments, and perform the conversion according to the preset density parameters. For example, if the user inputs 50 ml of soy sauce and the preset density of soy sauce is known to be 1.15 g / ml, it can be converted into 57.5 g. Establish an ingredient processing state compensation mechanism. For the fresh / dried / frozen state of the ingredients, use different moisture compensation coefficients to dynamically correct the ingredient weight vector. For dried ingredients, restore them to the original edible part weight through the rehydration rate parameter. For example, the rehydration rate of dried shiitake mushrooms is 1:8. If the user uses 10 g of dried shiitake mushrooms, the original edible part weight after calculation is 80 g.

[0029] Furthermore, nutrient content calculation, inputting the preprocessed data into a supervised learning model. This model calculates the nutrient content through the formula where N k is the content of nutrient k in a certain dish, W i is the weight of the i-th ingredient, C ik is the original content value of nutrient k per unit mass of the i-th ingredient, and f j is the optimization coefficient of nutrient k in the i-th ingredient corresponding to cooking method j. For example, to calculate the vitamin C content in a stir-fried shredded potato, given that the weight of the potato is 200 g, the original content coefficient of vitamin C in the potato is 20 mg per 100 g, and the vitamin C optimization coefficient corresponding to the stir-fry cooking method is 0.8, then the vitamin C content in this stir-fried shredded potato can be calculated.

[0030] Generate a standardized report: Generate a standardized report containing multiple nutrients in real time. Output a real-time standardized report, which includes: the core index values of energy, protein, fat, and carbohydrates and their percentages of the daily recommended intake; warning signs for the contents of cholesterol and sodium (if exceeding the preset threshold); visual annotations on the degree of nutrient loss based on cooking methods. The report can display the contents of various nutrients such as protein, fat, carbohydrates, vitamins, and minerals. For example, the report shows that a certain dish contains 15 grams of protein and 8 grams of fat, etc. Moreover, the output module integrates a function for evaluating nutrient intake. By calling the daily recommended intake standards in the user's health record database, a visual progress bar for the compliance degree of nutrients and warning signs for abnormal nutrients are generated in the standardized report. If the daily recommended intake of protein for a certain user is 60 grams, and the calculated protein content in a certain dish is 10 grams, then the contribution ratio of this dish to the daily recommended intake of protein is shown as 16.7% in the form of a progress bar in the report. If the content of a certain nutrient exceeds or is lower than the normal range, a warning sign is given.

[0031] In addition, the present invention also provides a device, specifically including: an input module: configured to receive ingredient information and cooking method parameters. The user can input information such as various ingredients, dosages, seasonings, and cooking methods of the dish through this module. A data processing module: configured to perform data cleaning, name standardization, and vectorization processing. Organize and preprocess the input information, convert non-standard names into standard names, and generate ingredient weight vectors, etc. A model operation module: stores a supervised learning model and a preset database of cooking method influence coefficients. It includes a standard food composition database and a cooking method influence factor control database. The cooking method influence factor control database contains a cooking method influence factor control table, which includes various Chinese cooking combination data. For example, information such as the optimization coefficients of various nutrients under different cooking methods such as frying, boiling, deep-frying, steaming, stir-frying, and braising. An output module: generates a visual interface for the nutrient content report. Presents the calculated nutrient content to the user in an intuitive visual way, and integrates a function for evaluating nutrient intake, showing a visual progress bar for the compliance degree of nutrients and warning signs for abnormal nutrients. The user inputs ingredient information through the input module, including 200 grams of green vegetables, 10 grams of cooking oil, and 2 grams of salt. The system predicts the cooking method as steaming, stir-frying, and braising based on this information, combined with big data analysis or preset rules.

[0032] Steps for data preprocessing, such as standardizing the ingredient names. The name of green vegetables is a standard name and does not need to be converted. Then, process the ingredient dosages to generate a green vegetable weight vector of 200 grams, a cooking oil weight vector of 10 grams, and a salt weight vector of 2 grams. For cooking oil and salt, perform mass-volume conversion according to the preset density parameters.

[0033] Preset food processing status compensation mechanism. When the green vegetables are in the fresh food state, the weight vector of the green vegetables is corrected according to the preset fresh food moisture compensation coefficient (if necessary).

[0034] Input the preprocessed data into the supervised learning model. It is known that the original content coefficient of vitamin C in green vegetables is 30 mg per 100 g, and the optimization coefficient of vitamin C corresponding to green vegetables under cooking methods such as stir-frying, steaming, and braising is 0.7. Then the content of vitamin C in stir-fried green vegetables is 42 mg. For other nutrients such as protein and fat, the calculation is also carried out according to the model formula. Further, the processing module is also configured to: identify and replace the ingredient names in the user input; provide a standardized suggestion list for the user to select for the unrecognized ingredient names.

[0035] The output module generates a visual nutritional content report interface. The report shows that the content of vitamin C in stir-fried green vegetables is 42 mg, the protein content is 2 g, the fat content is 10 g, etc. At the same time, it calls the daily recommended intake standard in the user's health record database to generate a visual progress bar for the compliance degree of nutrients and an early warning sign for abnormal nutrients. If the daily recommended intake of vitamin C for the user is 100 mg, the progress bar shows that the contribution ratio of this dish to the daily recommended intake of vitamin C is 42%. Compared with the traditional manual calculation method based on the food composition table, the model-based calculation method of the present invention can quickly process a large amount of data, greatly improving the calculation efficiency. At the same time, by considering the influence of different cooking methods on nutrients and the precise processing of ingredients and seasonings, the calculation of nutrient content is more accurate. For example, when calculating the nutrient content of stir-fried green vegetables, the traditional manual calculation method may not be able to accurately consider the optimization effect of the stir-frying cooking method on vitamin C, while the present invention can accurately calculate, and the error can be controlled within a very small range.

[0036] The method for accurately and real-time calculating the nutrient content of finished dishes based on a model in the present invention can process a large amount of data quickly compared with the traditional manual calculation method based on food composition tables. By considering the impact of different cooking methods on nutrients and the precise processing of ingredients and seasonings, it greatly improves the calculation efficiency and accuracy, and controls the error within a very small range. At the same time, it avoids the problems faced by the nutrition assessment system based on image recognition, such as insufficient accuracy in identifying mixed ingredients, lack of a quantification mechanism for nutrient loss during the cooking process, and lag in the construction of a database for the ingredients of Chinese compound seasonings, improving the reliability and practicality of nutrition assessment. It can also generate a standardized report containing multiple nutrients in real time and integrate the function of nutrition intake assessment, enabling users to clearly understand the nutrient content of the dishes and the comparison with the daily recommended intake, facilitating the reasonable arrangement of diet. For example, in the nutrition report of braised pork ribs, users can intuitively know whether nutrients such as protein, fat, and sodium meet the standards, so as to better control their dietary health. This method and device can be widely applied in food processing enterprises to accurately control the nutritional components of products, in the catering service industry to provide accurate nutritional information to customers, and in the medical and health field to assist doctors in formulating personalized diet plans for patients, helping patients with diseases such as diabetes and hypertension recover and manage their health, and providing strong support for the development of related fields.

[0037] In summary, the method and device for accurately and real-time calculating the nutrient content of finished dishes based on a model in the present invention have significant innovation and practicality, can effectively solve many problems in the prior art, and provide strong support for the development of related fields.

[0038] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural transformation made under the inventive concept of the present invention by using the content of the specification and drawings of the present invention, or directly / indirectly applied in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. A method for accurately and real-time calculating the nutrient content of finished dishes based on a model, characterized in that, It includes the following steps: Obtain the dish ingredient information input by the user, including the types of ingredients, usage amounts, and condiment information, and output the predicted cooking method parameters based on the dish ingredient information; Convert the ingredient names into the standard names in the standard database, and generate an ingredient weight vector based on the ingredient usage amounts; Input the preprocessed data into a supervised learning model, and the model calculates the nutrient content through the following formula: Among them, N k is the content of nutrient k in a certain dish, W i is the weight of the i-th ingredient, C ik is the original content value of nutrient k per unit mass of the i-th ingredient, f j is the optimization coefficient of nutrient k in the i-th ingredient corresponding to cooking method j. Generate a standardized report containing multiple nutrients in real time.

2. The method for accurately and real-time calculating the nutrient content of finished dishes based on a model according to claim 1, wherein It also includes: establishing an ingredient alias mapping table to convert non-standard names into the standard names in the standard database; performing mass-volume conversion on condiments and converting them according to the preset density parameters.

3. The method for accurately and real-time calculating the nutrient content of finished dishes based on a model according to claim 1, characterized in that, The cooking method parameters include pan-frying and baking types, boiling types, deep-frying types, steaming, frying, and stewing types, and the corresponding optimization coefficients are determined in advance through experimental data and stored in the database.

4. The method for accurately and real-time calculating the nutrient content of a finished dish based on a model according to claim 1, wherein The specific implementation of the constraint conditions is: The range of the optimization coefficient is limited to 0 to 1.

5. The method for accurately and real-time calculating the nutrient content of a finished dish based on a model according to claim 1, characterized in that, Output a real-time standardized report, and the real-time standardized report includes: the core index values of energy, protein, fat, and carbohydrates and their percentages of the daily recommended intake; warning signs for the contents of cholesterol and sodium; visual annotations on the degree of nutrient loss based on the cooking method.

6. The method for accurately and real-time calculating the nutrient content of finished dishes based on a model according to claim 2, wherein The data preprocessing also includes: Establish an ingredient processing state compensation mechanism. For the fresh / dried / frozen state of ingredients, different moisture compensation coefficients are used to dynamically correct the ingredient weight vector, and the dried ingredients are restored to the original edible part weight through the rehydration rate parameter.

7. A device, characterized in that, It includes: An input module configured to receive ingredient information and cooking method parameters; A data processing module configured to perform data cleaning, name standardization, and vectorization processing; A model operation module that stores a supervised learning model and a preset cooking method influence coefficient database; An output module that generates a visual nutrient content report interface.

8. The device according to claim 7, characterized in that, The model operation module includes a standard food composition database and an optimization coefficient comparison database.

9. The device according to claim 7, characterized in that, The data processing module is also configured to: identify and replace the ingredient names in the user input; provide a list of standardization suggestions for the unrecognized ingredient names for the user to select.

10. The device according to claim 8, characterized in that, The output module integrates a nutrient intake evaluation function. By calling the daily recommended intake standard in the user health record database, a visual progress bar for the compliance degree of nutrients and an abnormal nutrient warning sign are generated in the standardized report.

Images