Personalized rice noodle forming methods, systems, equipment, and media based on individual characteristics

Abstract

This invention discloses a personalized rice noodle forming method, system, equipment, and medium based on individual characteristics, belonging to the field of artificial intelligence starch-based food processing technology. Individual users input their physical characteristics, including fasting blood glucose, chewing ability, and digestive capacity, through an intelligent collection and display terminal. The rice noodle forming model constructed by the invention outputs personalized rice noodle formula components and nutrient additives, and controls parameters such as extrusion pressure and temperature during the production process, ensuring that the produced rice noodles accurately match the individual user. This truly achieves personalized customization of a single food category based on the food's inherent composition and texture. It solves the problem that current customized rice noodles for specific health characteristics still have a wide range of compatibility, are still produced in a standardized, assembly-line manner, and still have compatibility errors between individual users and the customized rice noodles.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence starch-based food processing technology, specifically relating to a personalized rice noodle forming method, system, equipment, and medium based on individual characteristics. Background Technology

[0002] Traditional rice noodle production uses standardized formulas and fixed processing parameters, which has significant limitations: the formula relies on a fixed proportion of starch raw materials, making it impossible to dynamically adjust the firmness of the rice noodles according to the digestive capacity of the target user; for example, the elderly prefer soft noodles while young adults prefer firm noodles. Although there are rice noodle products on the market specifically for the elderly or those with high blood sugar, these are all artificially adjusted formulas within a broad range of physical characteristics, such as high blood sugar or the elderly. However, each user's characteristics are different, and products made within this broad range are not suitable for every user, nor do they completely match each user's physical data. The core texture and nutritional composition of rice noodles, a single food category, lack precise, independent, and personalized adaptation. Current so-called personalized customization is essentially just standardized production within a broad range or a simple combination of existing foods, without reconstructing the food itself, especially a single food category. It still cannot adapt to individualized health needs, specifically manifested in poor nutritional adaptability, making it difficult to meet the nutritional restrictions of individual individuals such as diabetic patients, the elderly, and other special groups, and the rigid taste, with a single texture unable to accommodate the differences in chewing ability among different age groups within this broad range of physical characteristics. Summary of the Invention

[0003] This invention provides a personalized rice noodle forming method, system, equipment, and medium based on individual characteristics. Based on the age, chewing ability, fasting blood glucose, and other physical characteristics of different individuals, a personalized rice noodle forming method is used to construct a model, generating a customized rice noodle hardness and nutritional formula. Based on this personalized formula, the rice noodle forming system precisely and quantitatively adds the formula ingredients and nutritional additives. Starting from the inherent texture of a single food category, this invention achieves precise and personalized matching for individuals with different health characteristics, satisfying their optimal chewing and nutritional suitability. This invention solves the problem that current customized rice noodles for specific health characteristics still have a wide range of customization options and are still produced in a standardized, assembly-line manner, resulting in compatibility errors between individual individuals and the customized rice noodles.

[0004] To achieve the above-mentioned technical objectives, the present invention is implemented through the following technical solution:

[0005] A personalized rice noodle forming method based on user body characteristics includes the following steps:

[0006] S1: Source data collection, which collects user body characteristic data and rice noodle recipe characteristic data respectively;

[0007] The user's physical characteristics data include: user chewing ability data, digestive ability data, and fasting blood glucose index data;

[0008] The rice noodle recipe and characteristic data include: nutritional parameters, textural parameters and related recipe component ratio parameters, and processing parameters;

[0009] S2: Data preprocessing, removing outliers that do not conform to conventional standards from the collected data, and supplementing missing values ​​by referring to the median value of user data with similar physical characteristics;

[0010] One-hot coding is used to encode the cleaned classification features, and normalized coding is used to encode the cleaned continuous features.

[0011] The collected historical rice noodle recipe data and individual characteristic data were labeled;

[0012] S3: Core feature screening. Pearson correlation coefficient is used for linear correlation analysis. Features with high correlation coefficients are retained, while features with low correlation coefficients are removed. For example, the correlation between resistant starch content and blood sugar is retained, and the correlation between amylose content in rice noodle recipe and GI (glycemic index) value is removed. The correlation between height and amylose content in rice noodle recipe is removed.

[0013] S4: Construct derivative interactions between user characteristics and food characteristics. For diabetes, construct an interaction between fasting blood glucose value and GI value to reflect the superimposed correlation between an individual's fasting blood glucose level and the GI index of food. For chewing ability, construct an interaction between chewing ability level and the texture hardness level of cooked rice noodles to reflect the matching degree between an individual's chewing ability and the texture of cooked rice noodles.

[0014] S5: Basic model construction, based on individual user characteristics and rice noodle recipe ingredient parameters, a basic model is constructed using random forest regression;

[0015] S6: Advanced model, based on multilayer perceptron (MLP) model to capture refined feature interactions, improving the overall output accuracy of the model;

[0016] S7: Model Training. The advanced model is trained by dividing the data into a 7:2:1 ratio: training set, used for model training and learning; validation set, used to validate the output results, such as whether the rice noodle recipe ingredients and proportions, the hardness of the rice noodles obtained from the nutrient composition, and the GI value match the user's chewing ability and diabetes level, and to optimize the model parameters based on the validation results; and test set, used to evaluate the model output results.

[0017] S8: Post-training optimization of the model involves using grid search combined with Bayesian optimization to fine-tune the model parameters; stacking model fusion is used to reduce the generalization error of a single model.

[0018] Preferably, the user chewing ability data is obtained by extracting tooth and periodontal feature data after visual image analysis and processing by the U-net model based on real-time images of individual users' teeth and periodontium, combined with bite force detection data; the digestive ability data is obtained through clinical physical examination; the fasting blood glucose data is obtained by individual users by real-time detection of their own fasting blood glucose value, and the model can determine the individual user's diabetes level based on the fasting blood glucose value.

[0019] Preferably, the nutritional parameters include: resistant starch content, probiotic content, and mushroom extract content; wherein the resistant starch content is related to the GI value, the probiotic content is related to the degree of digestibility, and the mushroom extract content is related to the freshness and flavor of the rice noodles;

[0020] The processing parameters include extrusion pressure and the gelatinization, expansion, and setting temperatures of the mixture.

[0021] Preferably, the mapping relationship between the texture parameters and the proportion parameters of the associated formulation components is as follows:

[0022] The formula for firm rice noodles is: 60%~80% rice flour + 20~30% corn starch + 1~10% tapioca starch;

[0023] The formula for medium-quality rice noodles is: 60%~70% rice flour + 10%~20% corn starch + 10%~20% tapioca starch + 1%~10% potato starch;

[0024] The formula for soft rice noodles is: 50%~70% rice flour + 1%~10% corn starch + 5%~15% tapioca starch + 20%~30% potato starch.

[0025] Preferably, the historical rice noodle recipe data and individual characteristic data are labeled as follows: the rice noodle recipe ingredients and parameters correspond to the user's diabetes level and chewing ability, and the final GI value and hardness match, then it is labeled as qualified; otherwise, it is labeled as unqualified. The criteria for matching qualified are provided by nutrition professionals based on clinical standards.

[0026] Preferably, the basic model is constructed based on the GI value and the ingredients and proportions of the rice noodle recipe, the chewing ability level and the ingredients and proportions of the rice noodle recipe, and the processing parameters.

[0027] Preferably, the input layer of the MLP is set to user individual feature data and rice noodle recipe characteristic data; the hidden layer is set to 2-3 layers, each layer is set to 20-50 neurons; the output layer is set to rice noodle recipe parameters and effect prediction. The recipe parameters include: rice noodle recipe components and proportions, nutrient components and addition amounts, extrusion puffing temperature; the effect prediction includes: the hardness of the cooked rice noodles, the amount of probiotics retained, and the GI value after consumption.

[0028] Preferably, during the model training process, mean squared error loss is used for the regression task to monitor the GI value and texture hardness index of the rice noodle recipe components; cross-entropy loss is used to monitor the accuracy of qualified rice noodle recipe components.

[0029] During model training, early stopping, L2 regularization, and random dropout are combined to prevent overfitting and improve the model's generalization ability. The early stopping method stops training if the training set loss does not decrease after more than five rounds.

[0030] Another object of the present invention is to provide a personalized rice noodle forming system based on user body characteristics, comprising:

[0031] The intelligent data acquisition and display terminal includes a data input terminal for inputting individual users' age, fasting blood glucose level, and digestive function data; an image acquisition terminal for acquiring visual images of teeth and periodontal tissues; and the acquired data for providing to the rice noodle forming model.

[0032] The central processing unit is connected to the intelligent acquisition and display terminal. Based on the acquired individual user body characteristic data, the central processing unit outputs personalized rice noodle recipe ingredients and proportions, nutrient ingredients and addition amounts, and extrusion puffing temperature parameters from the personalized rice noodle forming model of the user body characteristics.

[0033] The multi-channel batching module includes rice flour bins, corn starch bins, cassava starch bins, and potato starch bins, arranged side by side and independently. Each bin is equipped with an electrically controlled metering valve at the bottom, and the electrically controlled metering valve is communicatively connected to the output of the actuator.

[0034] The nutrient flavor module is arranged side by side and independently includes a fungal extract compartment, a probiotic microcapsule compartment, and a resistant starch compartment; each compartment is equipped with an electrically controlled metering valve at the bottom, and the electrically controlled metering valve is communicatively connected to the output of the actuator;

[0035] An extrusion and puffing screw is equipped with a variable frequency motor drive, which is communicatively connected to the output end of the actuator; the actuator controls the rotational speed and extrusion pressure of the extrusion and puffing screw.

[0036] The outlets of each hopper of the multi-channel batching module are connected to the mixing chamber through pipes, and the mixing chamber is connected to the feed inlet of the extrusion and puffing screw through pipes.

[0037] The bottom of the fungus extract chamber and the resistant starch chamber in the nutrient flavor module are connected to the middle injection port of the extrusion screw via a pipe.

[0038] A dynamic temperature-controlled pot, located at the outlet of the extrusion screw, is divided into three independent temperature-controlled zones along the material flow direction: a first gelatinization zone with a temperature of 70-90℃, a middle puffing zone with a temperature of 110-130℃, and a final shaping zone with a temperature of 50-70℃. The bottom of the probiotic microcapsule chamber is connected to the dynamic temperature-controlled pot via a pipe. When the temperature drops to 30-38℃ after shaping, probiotics are added to the rice noodle mixture. A pressure sensor is installed at the connecting flange between the outlet end of the extrusion screw and the inlet end of the dynamic temperature-controlled pot. The pressure sensor is communicatively connected to the input end of the central processing unit. The extrusion pressure of the extrusion screw is adjusted according to the matching of the extrusion pressure monitoring. The extrusion pressure determines the tightness between the starch molecular chains inside the rice noodle body. Based on the rice noodle formula, the extrusion pressure simultaneously affects the texture and hardness of the rice noodles.

[0039] A forming mold is installed at the outlet end of a dynamic temperature-controlled pot via a quick-release flange. The forming mold has an annular array of 0.8-3mm holes. Rice noodles with a diameter of 0.8-3mm are formed through the forming mold.

[0040] The actuator receives PWM signals from the central processing unit and drives the metering valve opening of the multi-channel batching module, the metering pump flow of the nutrient flavor module, and the variable frequency motor speed of the extrusion puffing screw.

[0041] Another object of the present invention is to provide an apparatus comprising a processor and at least one memory for storing a computer-executable program; when at least one of the computer-executable programs is executed by the processor, the processor enables the processor to implement a personalized rice noodle forming method based on user body characteristics.

[0042] Another object of the present invention is to provide a computer-readable storage medium storing a computer-executable program, which, when executed by a processor, implements a personalized rice noodle forming method based on user body characteristics.

[0043] The beneficial effects of this invention are:

[0044] This invention provides a personalized rice noodle forming method based on user biometrics. It constructs an automatic analysis model that correlates the rice noodle recipe's ingredients and proportions with the rice noodle's texture and GI value, and with the correlation between nutritional parameters and digestibility. Based on an individual's age, chewing ability, digestive capacity, and fasting blood glucose level, the method obtains the rice noodle recipe's ingredients and proportions. Traditional methods only focus on creating personalized meals by combining existing food recipes, which doesn't substantially change the food's structure and nutrition. It only achieves nutritional matching of food combinations, without truly reshaping and integrating the food's nutritional composition. The synergistic effects between combined nutrients are limited. Alternatively, it may offer broad applicability customization within a specific characteristic area, such as customized production for diabetic patients, but customization should also consider... The production scope standards are relatively broad and are set by humans, making it impossible to achieve precise individual customization. Even if customized production is within a certain range, it will still differ from the individual's physical characteristics. The idea of ​​this invention focuses on a single food category. Based on the physical characteristic data of an individual, it can adjust the ingredients and processing parameters of rice noodles according to the chewing ability of the individual user to form the appropriate texture of rice noodles, the appropriate GI value of rice noodles after stable consumption according to the diabetes level, the appropriate texture of rice noodles according to the digestive ability, and the addition of probiotics. By changing the structure of rice noodles themselves, taking into account the texture, glycemic index, and digestibility, the invention achieves precise mapping of the food's own components and texture to the user's key dietary characteristics such as taste, chewing, blood sugar, and digestion.

[0045] Breaking through the limitations of a single model, the approach progresses from matching basic user characteristics with rice noodle recipes using a basic model, to deepening and advancing the model using a multilayer perceptron to capture the synergistic effects of refined features such as resistant starch content and extrusion temperature on GI values, which are then mapped back to the rice noodle recipe. Finally, stacking model fusion reduces the generalization error of a single model and improves the model's prediction accuracy for new data. The ultimate goal is to improve the accuracy of rice noodle recipe components and portion parameters, processing parameters, and nutritional addition parameters obtained based on individual user characteristics, thereby enhancing the precise mapping and matching of finished rice noodles to individual users.

[0046] The model directly predicts and outputs rice noodle recipes, nutritional information, and processing parameters, and maps them directly to the automated production module. The central control unit drives each module to directly produce rice noodle products that conform to the model's output results, achieving the goal of seamlessly integrating intelligent systems into assembly line production. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram illustrating the construction process of a personalized rice noodle forming method model based on user body characteristics according to the present invention.

[0049] Figure 2 This is a schematic diagram of a personalized rice noodle forming process based on user body characteristics according to the present invention;

[0050] Figure 3 This is a schematic diagram of a personalized rice noodle forming system based on user body characteristics according to the present invention;

[0051] In the attached diagram, the structural names represented by each number are as follows:

[0052] 101-Intelligent acquisition and display terminal, 102-Central processing unit, 103-Multi-channel ingredient dispensing module, 104-Nutrient flavor module, 105-Extrusion puffing screw, 106-Dynamic temperature control pot, 107-Forming mold, 108-Actuator, 109-Pressure sensor. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Example 1

[0055] like Figure 1 As shown, the relevant model for the personalized rice noodle forming method based on user body characteristics is constructed as follows:

[0056] S1: Source data collection, which collects user body characteristic data and rice noodle recipe characteristic data respectively;

[0057] The user body feature data to be collected includes: user chewing ability data: based on the number of teeth, the condition of molars, and whether there is periodontal atrophy, the chewing ability is judged by visual analysis; in addition, the chewing bite force value of the user is measured by a chewing force meter.

[0058] Digestive capacity data were obtained through clinical testing: measuring the activity and quantity of digestive enzymes such as amylase and lipase in the body; and the respective quantities of probiotics and pathogenic bacteria in the gut.

[0059] Fasting blood glucose index data: After fasting for 8-12 hours, usually in the morning before eating, blood is collected from the fingertip using a blood glucose meter for testing; the fasting blood glucose index is used to map the diabetes level; the criteria for determining the fasting blood glucose value and the diabetes level are based on clinical standards.

[0060] Rice noodle recipe and characteristic data: Based on past production and R&D, the relationship between the firmness of rice noodles and the recipe and ingredient proportions is as follows:

[0061] The formula for firm rice noodles is: 60%~80% rice flour + 20~30% corn starch + 1~10% tapioca starch;

[0062] The formula for medium-quality rice noodles is: 60%~70% rice flour + 10%~20% corn starch + 10%~20% tapioca starch + 1%~10% potato starch;

[0063] The formula for soft rice noodles is: 50%~70% rice flour + 1%~10% corn starch + 5%~15% tapioca starch + 20%~30% potato starch;

[0064] Nutritional parameters: Adding resistant starch to the above-mentioned basic rice noodle formula can control the glycemic index (GI) after eating rice noodles; adding probiotics can adjust the intestinal flora and promote digestion; adding fungal extracts can supplement the amino acids in rice noodles and enhance the freshness and flavor of rice noodles.

[0065] Besides the basic ingredients of rice noodles being an important factor affecting the texture and hardness of rice noodles, controlling the extrusion pressure of the extrusion screw and adjusting the tightness of the starch molecular chains in rice noodles can also synergistically affect the texture and hardness of rice noodles.

[0066] S2: Data preprocessing, removing outliers that do not conform to conventional standards, such as removing abnormally high blood glucose values ​​for those without a history of diabetes or abnormally low blood glucose values ​​for those with a history of diabetes; supplementing some missing values ​​by referring to the median value of user data with similar physical characteristics or by re-recording, to avoid missing key data; if chewing ability data is missing, it can be supplemented by referring to the median or mean value of data with similar age and similar dental and periodontal conditions.

[0067] One-hot coding is used to encode the cleaned classification features, and normalized coding is used to encode the cleaned continuous features.

[0068] The collected historical rice noodle recipe data and individual characteristic data are labeled, specifically the rice noodle recipe ingredients and parameters are corresponding to the diabetes level and chewing ability of the user. The final test results show that if the GI value and hardness match, it is labeled as qualified; otherwise, it is labeled as unqualified.

[0069] S3: Core feature screening. Pearson correlation coefficient was used for linear correlation analysis. Features with high correlation coefficients were retained, while features with low correlation coefficients were removed. For example, the correlation between resistant starch content and blood sugar was retained, and the correlation between amylose content in rice noodle recipe and GI (glycemic index) value was retained. The correlation between age, periodontal condition and chewing degree was retained. Features such as height and amylose content in rice noodle recipe were removed.

[0070] S4: Construct derivative interactions between user characteristics and food characteristics. For diabetes, construct an interaction between fasting blood glucose value and GI value to reflect the superimposed correlation between an individual's fasting blood glucose level and the GI index of food. For chewing ability, construct an interaction between chewing ability level and the texture hardness level of cooked rice noodles to reflect the matching degree between an individual's chewing ability and the texture of cooked rice noodles.

[0071] S5: Basic model construction, based on GI value and rice noodle recipe ingredients and component parameters; chewing ability level and rice noodle recipe ingredients and component parameters; extrusion pressure; the first stage gelatinization, the middle stage puffing, and the final stage setting temperature are constructed using random forest regression to build the basic model;

[0072] S6: Advanced model, based on multilayer perceptron (MLP) model to capture refined feature interactions, improving the overall output accuracy of the model;

[0073] The input layer of the multilayer perceptron is set to user individual feature data and rice noodle recipe characteristics data; the hidden layer is set to 2-3 layers, with 20-50 neurons in each layer; the output layer is set to rice noodle recipe parameters and effect prediction. The recipe parameters include: rice noodle recipe ingredients and proportions, nutrient components and addition amounts, and extrusion puffing temperature; the effect prediction includes: the hardness of the cooked rice noodles, the amount of probiotics retained, and the GI value after consumption.

[0074] S7: Model Training. The advanced model is trained by dividing the data into a 7:2:1 ratio: training set, used for model training and learning; validation set, used to validate the output results, such as whether the rice noodle recipe ingredients and proportions, the hardness of the rice noodles obtained from the nutrient composition, and the GI value match the user's chewing ability and diabetes level, and to optimize the model parameters based on the validation results; and test set, used to evaluate the model output results.

[0075] During model training, mean squared error loss is used for regression tasks to monitor the GI value and texture hardness index of rice noodle recipe components; cross-entropy loss is used to monitor the accuracy of qualified rice noodle recipe components.

[0076] During model training, early stopping, L2 regularization, and random dropout are combined to prevent overfitting and improve the model's generalization ability. The early stopping method stops training if the training set loss does not decrease after more than five rounds.

[0077] S8: Post-training optimization of the model involves using grid search combined with Bayesian optimization to fine-tune the model parameters; stacking model fusion is used to reduce the generalization error of a single model.

[0078] Example 2

[0079] like Figure 3 As shown, this embodiment provides a personalized rice noodle forming system based on user body characteristics, including:

[0080] The intelligent acquisition and display terminal 101 has a data input terminal for inputting the individual user's age, fasting blood glucose level, and digestive function data; an image acquisition terminal for acquiring visual images of teeth and periodontium; and an input terminal for inputting chewing force detection data of the occlusal instrument into the intelligent acquisition and display terminal 101 to provide the model for evaluating chewing ability.

[0081] The central processing unit 102 and the intelligent acquisition and display terminal 101 are communicatively connected to the central processing unit 102. The central processing unit 102 is equipped with at least one set of memory, which stores computer-executable programs. The processor executes the program to complete the rice noodle forming method based on user body characteristics in Example 1. Based on the acquired individual user body characteristic data, the central processing unit 102 outputs personalized rice noodle recipe ingredients and proportions, nutrient ingredients and addition amounts, and extrusion puffing temperature parameters from a personalized rice noodle forming model based on user body characteristics.

[0082] The multi-channel ingredient module 103 is set up side by side and independently includes rice flour bins, corn starch bins, cassava starch bins and potato starch bins. Each bin is equipped with an electrically controlled metering valve at the bottom, which is connected to the output of the actuator 108. According to the composition and proportion output by the central processor, the electrically controlled metering valve at the bottom of each bin adds each component in the formula to the mixing chamber in a quantitative manner.

[0083] The nutrient flavor module 104 is set up side by side and independently, including a mushroom extract compartment, a probiotic microcapsule compartment, and a resistant starch compartment; each compartment is equipped with an electrically controlled metering valve at the bottom, which is communicatively connected to the output of the actuator 108; based on the output of the central processing unit 102, the electrically controlled metering valve at the bottom of each compartment quantitatively adds the required nutrients to the rice noodle mixture.

[0084] The extrusion screw 105 is driven by a variable frequency motor, which is communicatively connected to the output of the actuator 108. The speed and extrusion pressure of the extrusion screw are controlled by the actuator 108. In this embodiment, the speed is set to 180 rpm, and the extrusion pressure needs to be set by the system according to the chewing data and digestive function data of the user's body characteristics.

[0085] The outlets of each hopper of the multi-channel batching module 103 are connected to the mixing chamber through pipes, and the mixing chamber is connected to the inlet of the extrusion and expansion screw 105 through pipes; the raw materials of the formula are mixed evenly and then extruded and expanded by the extrusion and expansion screw 105 to promote changes in the physical form of the physical mixture and fully integrate and agglomerate;

[0086] The bottom of the resistant starch bin and the fungal extract bin of the nutrient flavor module 104 are connected to the middle section injection port of the extrusion screw 105 through a pipe; fungal extract, resistant starch and other nutrients are added in the middle section of the raw material extrusion and puffing so that they are fully integrated with the raw material in the later stage of extrusion and puffing.

[0087] The dynamic temperature control pot 106 is located at the outlet of the extrusion screw 105. The rice noodle mixture, after being fully extruded, flows into the dynamic temperature control pot 106. The dynamic temperature control pot 106 is divided into three independent temperature control zones: the first gelatinization zone, with a temperature set at 80℃, to fully gelatinize the starch and prevent the rice noodles from being undercooked or easily broken after forming; the middle puffing zone, with a temperature set at 120℃, to ensure the rice noodles are smooth and elastic; and the final shaping zone, with a temperature set at 60℃, to fully agglomerate and shape the noodles, facilitating subsequent extrusion into strips. The probiotic microcapsule chamber is connected to the dynamic temperature control pot through a pipe. When the rice noodle mixture has been shaped and the temperature has dropped to 30~38℃, probiotic microcapsules are added to the rice noodle mixture.

[0088] A pressure sensor 109 is installed at the connecting flange between the outlet end of the extrusion screw 105 and the inlet end of the dynamic temperature control pot 106. The pressure sensor 109 is communicatively connected to the input end of the central processing unit 102. The extrusion pressure of the extrusion screw 105 is adjusted by monitoring the extrusion pressure. The magnitude of the extrusion pressure of the extrusion screw 105 affects the compactness of the starch molecular chains after the rice noodles are formed, that is, the softness and hardness of the rice noodles. Therefore, in addition to controlling the softness and hardness of the rice noodles by adjusting the ingredients of the formula, the uniformity of the softness and hardness of the rice noodles is achieved by detecting the extrusion pressure of the extrusion screw 105 in real time.

[0089] The forming mold 107 is set at the outlet end of the dynamic temperature control pot 106 via a quick-release flange. The forming mold 107 has a 0.8~3mm annular array of channels. The material that has been gelatinized, puffed and shaped is extruded through the forming mold 107 to form thin rice noodles with a diameter of 0.8~3mm.

[0090] The actuator 108 receives the PWM signal from the central processing unit 102 and drives the metering valve opening of the multi-channel batching module 103, the metering pump flow of the nutrient flavor module 104, and the variable frequency motor speed of the extrusion puffing screw 105.

[0091] Example 3

[0092] User 1, age: 30;

[0093] Fasting blood glucose test value: The blood glucose test is performed by collecting blood from the fingertip using a blood glucose meter before breakfast in the morning;

[0094] Fasting blood glucose level: 4.2 mmol / L, fasting blood glucose is normal;

[0095] Description of teeth and periodontal characteristics: The gingival margin is coral pink in color, and the gingival margin is slightly rounded; the teeth are closely arranged, with no missing teeth or looseness; the chewing force of the molars measured by the occlusal apparatus is 170N, indicating good chewing ability;

[0096] Digestive capacity data: Serum amylase level was low at 32 U / L, and the number of beneficial bacteria in the intestine was low; Helicobacter pylori was detected.

[0097] In summary, User 1 has normal chewing ability, no diabetes, but weak digestive function;

[0098] The above-mentioned physical characteristics of User 1 are input into the intelligent acquisition and display terminal in Example 2, wherein the images of the user's teeth and periodontal condition are acquired on-site or entered using recent photos; the provided physical characteristic data is processed by the personalized rice noodle forming model based on the user's physical characteristics provided in Example 1, and the rice noodle recipe ingredients and proportions, and the amount of added nutrients are output; the above-mentioned data output by the model are fed into the central processing unit in Example 2, and the central processing unit starts the personalized customized rice noodle production by outputting control commands, starting from the addition of raw materials;

[0099] The rice noodles produced in this embodiment were tested as follows:

[0100] Hardness testing: Wet rice noodles after cooking were taken, with a forming diameter of 3mm. They were blanched in clean boiling water for 1 minute to simulate the normal rice noodle cooking process. The blanched rice noodles were then tested for hardness using a texture analyzer with a diameter of 36mm cylindrical probe, a deformation setting of 50%, and a testing speed of 1mm / second. The average weight of multiple tests was 260g, indicating that the rice noodles were of medium texture. Although User 1 has good chewing ability, considering their poor digestive function, the final rice noodles produced by the system were of medium texture, preserving the taste while making them easy to digest.

[0101] Ingredient analysis: The produced rice noodles were preserved to ensure they remained moist and then sent to a third-party testing center for analysis of the ingredients. The results are as follows (rounded to the nearest integer):

[0102] The test results showed that the rice flour content was 65%, the corn starch content was 15%, the tapioca starch content was 15%, and the potato starch content was 5%.

[0103] No resistant starch was detected in the nutrients, and the survival rate of the added probiotics reached 85%.

[0104] Example 4

[0105] User 2, age: 45;

[0106] Fasting blood glucose test value: The blood glucose was measured by finger prick blood using a blood glucose meter before the morning meal without insulin injection. The value was 7.1 mmol / L, which is high. The clinical diagnosis is type 2 diabetes.

[0107] Description of dental and periodontal characteristics: The gums are pink, full and without recession, the teeth are closely arranged, there are no missing teeth, and there is no looseness; the chewing force of the molars measured by the occlusal instrument is 158N, indicating good chewing ability;

[0108] Digestive capacity data: Serum amylase level was 86 U / L, which is normal; probiotic count was within the normal range; Helicobacter pylori was negative.

[0109] In summary, User 2 has normal chewing ability, good digestion, high blood sugar, and a history of diabetes.

[0110] The above-mentioned physical characteristics of User 2 are input into the intelligent acquisition and display terminal in Example 2, wherein the images of the user's teeth and periodontal condition are acquired on-site or entered using recent photos; the provided physical characteristic data is processed by the personalized rice noodle forming model based on the user's physical characteristics provided in Example 1, and the rice noodle recipe ingredients and proportions, and the amount of added nutrients are output; the above-mentioned data output by the model are fed into the central processing unit in Example 2, and the central processing unit starts the personalized customized rice noodle production by outputting control commands, starting from the addition of raw materials;

[0111] The rice noodles produced in this embodiment were tested as follows:

[0112] Hardness test: Take wet rice noodles after cooking, with a diameter of 3mm, and blanch them in clean boiling water for 1 minute to simulate the normal rice noodle cooking process. After blanching, use a texture analyzer to test the hardness of the rice noodles. The texture analyzer uses a cylindrical probe with a diameter of 36mm, the deformation is set to 50%, and the test speed is 1mm / second. The average value of multiple tests is 320g, indicating that the rice noodles are hard and suitable for adults with normal chewing function.

[0113] Ingredient analysis: The produced rice noodles were preserved to ensure they remained moist and then sent to a third-party testing center for analysis of the ingredients. The results are as follows (rounded to the nearest integer):

[0114] The test results showed that the rice flour content was 70%, the corn starch content was 25%, and the tapioca starch content was 5%.

[0115] The resistant starch content was found to be 35% in the nutrients, meaning that the total mass of rice flour, corn starch, and tapioca starch in the rice noodle ingredients was 1:0.35 with the mass of resistant starch.

[0116] On the morning of the first day, before eating, User 2 consumed 200g of rice noodles produced by this invention. The noodles were blanched in water for 1 minute before consumption. Two hours after consumption, the GI value was measured to be 65.

[0117] The next morning, before eating, 200g of commercially available rice noodles were consumed, and the noodles were blanched in water for 1 minute. Two hours later, the GI value remained at 78. Therefore, after consuming the rice noodles produced by this invention, the GI value of user 2 decreased from 78 to 65.

[0118] Example 5

[0119] User 3, age: 70;

[0120] Fasting blood glucose test value: The blood glucose was tested using a finger prick blood sample taken with a blood glucose meter before breakfast in the morning. The test value was 5.3 mmol / L, which is normal for fasting blood glucose and indicates no diabetes.

[0121] Description of dental and periodontal characteristics: Significant gum recession with partial root exposure, thinning of enamel, tooth loss, and looseness of remaining teeth; bite force measured by an occlusal apparatus is 80N, which is significantly low.

[0122] Digestive capacity data: Although serum amylase of 30 U / L is within the normal range, it is close to the critical value. The activity of probiotics is not strong and the number of beneficial bacteria is low. Overall digestive function is weak.

[0123] In summary, User 3 has weak chewing and digestive functions, but normal blood sugar levels.

[0124] The above-mentioned physical characteristics of user 3 are input into the intelligent acquisition and display terminal in Example 2, wherein the images of the user's teeth and periodontal condition are acquired on-site or entered using photos; the provided physical characteristic data is processed by the personalized rice noodle forming model based on the user's physical characteristics provided in Example 1, and the rice noodle recipe ingredients and proportions, and the amount of added nutrients are output; the above-mentioned data output by the model are fed into the central processing unit in Example 2, and the central processing unit starts the personalized customized rice noodle production by outputting control commands, starting from the addition of raw materials;

[0125] The rice noodles produced in this embodiment were tested as follows:

[0126] Hardness test: Take cooked rice noodles with a diameter of 3mm and blanch them in clean boiling water for 1 minute to simulate the normal cooking process of rice noodles. After blanching, use a texture analyzer to test the hardness of the rice noodles. The texture analyzer uses a cylindrical probe with a diameter of 36mm, the deformation is set to 50%, and the test speed is 1mm / second. The average value of multiple tests is 190g, indicating that the rice noodles are soft and suitable for elderly people with weak chewing and digestive functions.

[0127] Ingredient analysis: The produced rice noodles were preserved to ensure they remained moist and then sent to a third-party testing center for analysis of the ingredients. The results are as follows (rounded to the nearest integer):

[0128] The test results showed that the rice flour content was 60%, the corn starch content was 5%, the tapioca starch content was 10%, and the potato starch content was 25%.

[0129] No resistant starch was detected in the nutrients, and the survival rate of the added probiotics reached 80%.

[0130] Example 6

[0131] Compared with Example 2, this embodiment uses a two-stage temperature control pot with the same temperature as in Example 2, but the forming stage is eliminated; the rest of the processing flow and parameters are the same as in Example 2; and the ingredients are fed according to the hard rice noodle formula.

[0132] The hardness of the produced rice noodles was tested by a texture analyzer. The overall hardness unevenness error reached 12%, and the hardness of some areas reached 350g, which exceeded the upper limit threshold of chewing power of young adults with good chewing function. In addition, the rice noodles were brittle and easy to break.

[0133] Example 7

[0134] Compared with Example 2, this embodiment eliminates the pressure sensor between the extrusion screw and the dynamic temperature control pot, while the remaining processing flow and parameters remain the same as in Example 2; the ingredients are fed according to the hard rice noodle formula for production.

[0135] The hardness of the produced rice noodles was tested using a texture analyzer. The hardness of the entire batch of rice noodles fluctuated greatly, with the minimum hardness measured in some areas reaching 200g and the maximum hardness measured in some areas reaching 400g. The hardness of the rice noodles produced was uneven.

Claims

1. A personalized rice noodle forming method based on user body characteristics, characterized in that, Includes the following steps: S1: Source data collection, which collects user body characteristic data and rice noodle recipe characteristic data respectively; The user's physical characteristics data include: user chewing ability data, digestive ability data, and fasting blood glucose index data; The rice noodle recipe and characteristic data include: nutritional parameters, textural parameters and related recipe component ratio parameters, and processing parameters; S2: Data preprocessing, removing outliers that do not conform to conventional standards from the collected data, and supplementing missing values ​​by referring to the median value of user data with similar physical characteristics; One-hot coding is used to encode the cleaned classification features, and normalized coding is used to encode the cleaned continuous features. The collected historical rice noodle recipe data and individual characteristic data were labeled; S3: Core feature selection: Pearson correlation coefficient is used for linear correlation analysis to retain features with high correlation coefficients and remove features with low correlation coefficients. S4: Construct derivative interactions between user characteristics and food characteristics, including interactions between fasting blood glucose value and GI value for diabetes; and interactions between chewing ability level and texture hardness level of cooked rice noodles for chewing ability. S5: Basic model construction, based on individual user characteristics and rice noodle recipe ingredient parameters, a basic model is constructed using random forest regression; S6: Advanced model, based on MLP model to capture refined feature interactions, improving the overall output accuracy of the model; S7: Model training, training the constructed advanced model, dividing the data into: training set, used for model training and learning; validation set, used for validating the output results; and test set, used for evaluating the model output results. S8: Post-training optimization of the model involves using grid search combined with Bayesian optimization to fine-tune the model parameters; stacking model fusion is used to reduce the generalization error of a single model.

2. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, The user chewing ability data is obtained by extracting tooth and periodontal feature data from real-time images of individual users' teeth and periodontium using the U-net model after visual image analysis and processing, and combining it with bite force detection data. The digestive capacity data is obtained through clinical physical examinations; the model determines the individual user's diabetes level based on the fasting blood glucose value. The nutritional parameters include: resistant starch content, probiotic content, and fungal extract content; The processing parameters include extrusion pressure and the gelatinization, expansion, and setting temperatures of the mixture.

3. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, The mapping relationship between the texture parameters and the proportion parameters of the associated formulation components is as follows: The formula for firm rice noodles is: 60%~80% rice flour + 20~30% corn starch + 1~10% tapioca starch; The formula for medium-quality rice noodles is: 60%~70% rice flour + 10%~20% corn starch + 10%~20% tapioca starch + 1%~10% potato starch; The formula for soft rice noodles is: 50%~70% rice flour + 1%~10% corn starch + 5%~15% tapioca starch + 20%~30% potato starch.

4. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, The historical rice noodle recipe data and individual characteristic data are labeled as follows: the rice noodle recipe ingredients and parameters correspond to the user's diabetes level and chewing ability. If the final test GI value and hardness match, it is labeled as qualified; otherwise, it is labeled as unqualified.

5. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, The basic model is constructed based on the GI value and the ingredients and proportions of the rice noodle recipe, the chewing ability level and the ingredients and proportions of the rice noodle recipe, and the processing parameters.

6. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, The input layer of the MLP is set to user individual feature data and rice noodle recipe characteristic data; the hidden layer is set to 2 to 3 layers, and each layer is set to 20 to 50 neurons; The output layer is set to rice noodle recipe parameters and effect prediction. The recipe parameters include: rice noodle recipe ingredients and proportions, nutrient components and addition amounts, and extrusion puffing temperature. The effect prediction includes: the hardness of the cooked rice noodles, the amount of probiotics retained, and the GI value after consumption.

7. The personalized rice noodle forming method based on user body characteristics according to claim 1, characterized in that, During the model training process, mean squared error loss is used for the regression task to monitor the GI value and texture hardness index of the rice noodle recipe ingredients; cross-entropy loss is used to monitor the accuracy of qualified rice noodle recipe ingredients. During model training, early stopping, L2 regularization, and random dropout are combined to prevent overfitting and improve the model's generalization ability. The early stopping method stops training if the training set loss does not decrease after more than five rounds.

8. A personalized rice noodle forming system based on user body characteristics, characterized in that, include: The intelligent data acquisition and display terminal and the data input terminal are used to input individual users' age, fasting blood glucose level, and digestive function data. The image acquisition terminal acquires visual images of teeth and periodontium, and the acquired data is provided to the rice noodle forming model constructed according to any one of claims 1 to 7; The central processing unit is connected to the intelligent acquisition and display terminal. Based on the acquired individual user body characteristic data, the central processing unit outputs personalized rice noodle recipe ingredients and proportions, nutrient ingredients and addition amounts, and extrusion puffing temperature parameters from the personalized rice noodle forming model of the user body characteristics. The multi-channel batching module includes rice flour bins, corn starch bins, cassava starch bins, and potato starch bins, arranged side by side and independently. Each bin is equipped with an electrically controlled metering valve at the bottom, and the electrically controlled metering valve is communicatively connected to the output of the actuator. The nutrient flavor module is arranged side by side and independently includes a fungal extract compartment, a probiotic microcapsule compartment, and a resistant starch compartment; each compartment is equipped with an electrically controlled metering valve at the bottom, and the electrically controlled metering valve is communicatively connected to the output of the actuator; The extrusion and expansion screw is equipped with a variable frequency motor drive, and the variable frequency motor is communicatively connected to the output end of the actuator. The speed and extrusion pressure of the extrusion and expansion screw are controlled by the actuator; The outlets of each hopper of the multi-channel batching module are connected to the mixing chamber through pipes, and the mixing chamber is connected to the feed inlet of the extrusion and puffing screw through pipes. The mushroom extract chamber and the resistant starch chamber in the nutrient flavor module are connected to the middle injection port of the extrusion screw via pipes. A dynamic temperature control pot, located at the outlet of the extrusion screw, is divided into three independent temperature control zones along the material flow direction: a first gelatinization zone with a temperature of 70-90℃, a middle puffing zone with a temperature of 110-130℃, and a final shaping zone with a temperature of 50-70℃. A probiotic microcapsule chamber is connected to the dynamic temperature control pot via pipes. When shaping is complete and the temperature drops to 30-38℃, probiotic microcapsules are added to the rice noodle mixture. A pressure sensor is installed at the connecting flange between the outlet end of the extrusion screw and the inlet end of the dynamic temperature control pot, and the pressure sensor is communicatively connected to the input end of the central processing unit. The forming mold is set at the outlet end of the dynamic temperature control pot via a quick-release flange, and the forming mold has 0.8~3mm annular array channels. The actuator receives PWM signals from the central processing unit and drives the metering valve opening of the multi-channel batching module, the metering pump flow of the nutrient flavor module, and the variable frequency motor speed of the extrusion puffing screw.

9. A device, characterized in that, It includes a processor and at least one memory for storing a computer-executable program; when at least one of the computer-executable programs is executed by the processor, the processor enables the processor to implement the personalized rice noodle forming method based on user body characteristics as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer-executable program, which, when executed by a processor, implements the personalized rice noodle forming method based on user body characteristics as described in any one of claims 1 to 7.

Images