US 12,224,054 B2
Diet recommendation method, device, storage medium and electronic device
Xiying Zhang, Beijing (CN)
Assigned to BOE TECHNOLOGY GROUP CO., LTD., Beijing (CN)
Appl. No. 17/614,454
Filed by BOE TECHNOLOGY GROUP CO., LTD., Beijing (CN)
PCT Filed Dec. 25, 2020, PCT No. PCT/CN2020/139313
§ 371(c)(1), (2) Date Nov. 26, 2021,
PCT Pub. No. WO2022/133985, PCT Pub. Date Jun. 30, 2022.
Prior Publication US 2022/0399098 A1, Dec. 15, 2022
Int. Cl. G16H 20/60 (2018.01); G09B 19/00 (2006.01); G16H 50/20 (2018.01)
CPC G16H 20/60 (2018.01) [G09B 19/0092 (2013.01); G16H 50/20 (2018.01)] 14 Claims
OG exemplary drawing
 
1. A computer-implemented diet recommendation method performed by a server computing device or a terminal computing device, the server computing device or the terminal computing device comprising a central processing unit (CPU), and an input portion connected to an input/output (I/O) interface, the method comprising:
receiving, by the input portion, historical dining data; and
reading, by the CPU, the historical dining data, and determining, based on multiple candidate foods and the historical dining data, a target recommended recipe by using a recommendation model;
determining energy required by a user for each meal according to basic information of the user;
determining a mass of each nutrient required by the user for each meal according to following relationships each between the energy required by the user for each meal and a nutrient required by the user for each meal;

OG Complex Work Unit Math
wherein nutrientCountn represents a mass of a nth nutrient, energy represents an energy value required by the user for each meal in a candidate recipe, ration represents a ratio coefficient of the nth nutrient to the energy value of each meal, and paramn represents a converted energy coefficient of the nth nutrient; and n is a positive integer; and
determining a mass of each food in the target recommended recipe according to following relationships each between a content of a nutrient in a candidate food contained in the target recommended recipe and a mass of the nutrient required for each meal:

OG Complex Work Unit Math
wherein fatm represents a fat content of a mth food, xm represents a mass of the mth food, Count1 represents a fat mass required by the user for each meal in the candidate recipe, proteinm represents a protein content of the mth food, Count2 represents a protein mass required by the user for each meal in the candidate recipe, CHOm represents a carbohydrate content of the mth food, and Count3 represents a carbohydrate mass required by the user for each meal in the candidate recipe; m is a quantity of food types in the target recommended recipe, and m is a positive integer;
presetting, based on a preset meal preparation rule of each meal, a constraint condition of each nutrient required by the user according to the basic information of the user;
determining the mass of each food in the target recommended recipe according to the constraint condition, wherein the constraint condition on intake of carbohydrates includes: c1*x1+c2*x2+ . . . +cm*xm≤Count4, where cm represents a sugar content of the mth food, and Count4 represents a sugar mass required by the user for each meal in the target recommended recipe;
determining a content of each nutrient of each food in the target recommended recipe, and performing One-Hot encoding for the content of each nutrient of each food to obtain a content vector of each nutrient; and
outputting the mass of each food in the target recommended recipe by using the content vector of each nutrient as an input to a neural network model and updating the neural network model according to a loss function, wherein the neural network model is obtained by training with a sample content vector of each nutrient as input and the mass of each food in the target recommended recipe as output, and the loss function of the neural network model is expressed as: Loss=∥σ−σ′∥2+regular term,
wherein σ=(Count1, Count2, Count3, Count4) includes the mass of each nutrient required by the user for each meal, σ′ represents a mass of each nutrient required by the user for each meal predicted by the neural network model, and the regular term is a preset constraint condition of each nutrient required by the user.