Method for measuring resting metabolic rate based on digital anthropometric phenotype
Through a machine learning model based on digital anthropometric phenotype, the existing resting metabolic rate measurement methods are solved, and the existing resting metabolic rate measurement is achieved, which is universal and high precision.
Patent Information
- Application Number
- CN202510007032.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-03
- Publication Date
- 2025-06-03
AI Technical Summary
The existing resting metabolic rate measurement methods have the problem of high equipment accuracy but complex operation and time-consuming, and the existing formula has low measurement accuracy, which is affected by a variety of factors and is difficult to accurately estimate.
Using machine learning models based on digital anthropometry phenotypes, we use digital anthropometry data to predict resting metabolic rates without professional equipment and complex operations.
It achieves fast and accurate measurement of resting metabolic rate, which takes extremely short time (about 10 seconds to 30 seconds), and is universal and highly accurate, surpassing the prediction effect of classic formulas.
Smart Images
Figure CN120089261A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of physical anthropology methods, and particularly relates to a method for measuring resting metabolic rate based on digital anthropometric phenotypes. Background Art
[0002] In today's society, with the transformation of lifestyle and the change of eating habits, physical health has become one of the focuses of people's attention. The resting metabolic rate (RMR) refers to the minimum energy consumption required for the body to maintain basic life activities, which intuitively reflects the metabolic level of the body in a quiet state. Understanding one's own resting metabolic rate can help people better understand the body's energy needs, thus providing a scientific basis for health management. A higher resting metabolic rate usually means a stronger basal metabolic capacity, which helps to consume calories more efficiently in daily life. This is not only beneficial for weight management but also can improve physical strength and health status.
[0003] Traditional methods for measuring resting metabolic rate include the following two [1]:
[0004] (1) Indirect calorimetry: By measuring the oxygen consumption and carbon dioxide emission of a person at rest, and using the respiratory quotient (RQ) to calculate the calorie consumption, where the oxygen consumption (VO 2 ) and carbon dioxide consumption (VCO 2 ) are important parameters for calculating the resting metabolic rate. The specific formula is:
[0005] Resting metabolic rate (RMR) = (VO 2 ×3.941) + (VCO 2 ×1.106).
[0006] Currently, professional equipment commonly used for this method includes a respiratory gas analyzer, or a mask and nose and mouth cannula are used to measure the concentrations of oxygen and carbon dioxide, and the measurement time is relatively long, usually 20 - 30 minutes.
[0007] (2) Prediction formula method: Use some common formulas to predict the resting metabolic rate by measuring factors such as height, weight, age, gender, and resting heart rate method. Common formulas include the Harris - Benedict formula [2][3] and the Mifflin - St Jeor formula [4], etc. The accuracy of the results of these formulas is relatively low, and they are only suitable for simple estimation and cannot directly reflect the actual metabolic situation of individuals, especially for special populations (such as athletes [5], the elderly [6], etc.).
[0008] The deficiencies of the prior art are reflected in two aspects: wearable devices and measurement accuracy.
[0009] (1) Indirect calorimetry devices have high precision but complex measurement operations. They require wearing a respiratory gas analyzer, and the measurement time is relatively long, usually about 30 minutes, which is inconvenient to operate.
[0010] (2) When using existing formulas to measure resting metabolic rate, although it is convenient and does not require special equipment, the accuracy is relatively low. It is affected by many factors (such as mood, environment, health status, etc.) and is difficult to accurately estimate. Summary of the Invention
[0011] In view of the above deficiencies of the prior art, the present invention aims to propose a simple, convenient and low-cost method for measuring resting metabolic rate.
[0012] The method for measuring resting metabolic rate proposed by the present invention uses digital human anthropometric phenotypes to establish a machine learning model for predicting resting metabolic rate. Without professional equipment or complex procedures, accurate resting metabolic rate can be obtained. The main steps of this method include data processing, model generation and calculation of resting metabolic rate; the specific steps are as follows:
[0013] The overall process of the method for measuring resting metabolic rate based on digital human anthropometric phenotypes provided by the present invention is as Figure 1 shown, and the specific steps are as follows:
[0014] The first step: Collection of digital human anthropometric phenotype data and resting metabolic rate data;
[0015] The data used includes resting metabolic rate and digital human anthropometric phenotype data. Among them, resting metabolic rate data is collected by a human energy metabolism instrument; digital human anthropometric phenotype data is collected by a 3D scanner or three-dimensional reconstruction after 2D image collection, including but not limited to gender, height, weight, age, chest circumference, abdominal circumference, neck circumference, left and right arm circumferences, lower limb circumferences, head height, back width, waist-to-hip ratio, etc.
[0016] Correspond each digital human anthropometric phenotype data with the simultaneously measured resting metabolic rate data, using the digital human anthropometric phenotype data as input and the resting metabolic rate index as the label.
[0017] The second step: Data preprocessing;
[0018] Preprocess the data collected in the first step, mainly including: handling missing values, removing abnormal data, data standardization, normalization, etc.
[0019] (1) Handling missing values: Delete samples with missing data or fill in the missing data.
[0020] (2) Abnormal data cleaning: Detect abnormal data values and delete or replace them to smooth the data.
[0021] (3) Data standardization: To eliminate the influence of dimensional differences on model parameters, data needs to be standardized in advance.
[0022] (4) Data transformation: Convert the data into a form suitable for data modeling through methods such as smoothing aggregation, data generalization, and normalization. This includes normalizing the data, feature selection, dimensionality reduction, etc.
[0023] Step 3: Construct different machine learning models;
[0024] Use the preprocessed data to construct different machine learning models. Respectively, use the resting metabolic rate data as the label data and the digital anthropometric phenotype data as the input data to construct different machine learning models. The machine learning models here include but are not limited to multiple linear regression models, ridge regression models, Lasso regression models, decision tree regression models, and non-linear random forest regression models.
[0025] Step 4: Calculate the resting metabolic rate;
[0026] (1) Construct a mathematical model to predict the resting metabolic rate through the input digital anthropometric phenotype data (such as gender, height, weight, age, chest circumference, waist circumference, hip circumference, etc.) and the established statistical model.
[0027] (2) Perform data mapping. The core of the resting metabolic rate calculation formula depends on the training model, and the formula is:
[0028] Resting Metabolic Rate (RMR) =
[0029] f (circumferences of various body parts, widths of various body parts, lengths of various body parts, weight, gender, age)
[0030] The function f is a machine learning model trained based on a large number of real data samples (such as: multiple linear regression models, Lasso regression models, ridge regression models, decision tree regression models, random forest regression models, etc.).
[0031] Step 5: Optimize the machine learning model;
[0032] Optimize the model by changing the parameters of different machine learning models or feature selection methods. The optimization criteria include the number of independent variables, the goodness of fit of the model (such as R 2 , R-Squared), root mean squared error (RMSE), etc.
[0033] In the present invention, the evaluation metrics for training the machine learning model are goodness of fit (R 2 , R-Squared) and root mean squared error (RMSE, Root Mean Squared Error).
[0034] In the present invention, using digital anthropometric phenotypes as independent variables, a resting metabolic rate prediction model is constructed. The digital anthropometric phenotype data of the user to be measured is input into the prediction model for calculation to obtain the predicted value of the resting metabolic rate, and the basic metabolic capacity of the user is inferred.
[0035] Compared with the prior art, the features and beneficial effects of the present invention are as follows:
[0036] (1) The present invention has high efficiency. Only by providing digital anthropometric data can the measurement of the resting metabolic rate be completed. The whole process takes an extremely short time (about 10 seconds to 30 seconds), realizing rapid calculation and feedback.
[0037] (2) The present invention has universality: It does not require complex operation processes and supports multiple scenarios, such as home health management and daily monitoring by fitness coaches.
[0038] (3) The present invention has high precision. By means of machine learning technology, the resting metabolic rate is predicted from digital anthropometric phenotypes. The results are better than those of classical formulas such as the Harris-Benedict formula, making this method more convenient and more accurately guiding in clinical practice. Description of the Drawings
[0039] Figure 1 is a schematic flow chart of a method for measuring resting metabolic rate based on digital anthropometric phenotypes. Detailed Embodiments
[0040] The present invention will be introduced in detail below with reference to the drawings and specific embodiments.
[0041] The method for measuring resting metabolic rate based on digital anthropometric phenotypes provided by the present invention has the overall process as Figure 1 shown, and specifically includes the following steps:
[0042] First step, digital anthropometric phenotype measurement and resting metabolic rate measurement.
[0043] The data used in the present invention includes resting metabolic rate and digital anthropometric phenotype data. Among them, the resting metabolic rate data is collected by a human energy metabolism instrument; the digital anthropometric phenotype data can be collected by a 3D scanner or three-dimensional reconstruction after 2D image collection, including but not limited to human anthropometric phenotypes such as gender, height, weight, age, chest circumference, abdominal circumference, neck circumference, left and right arm circumferences, lower limb circumferences, head height, back width, and waist-to-hip ratio.
[0044] Correspond each digital anthropometric phenotype data with the simultaneously measured resting metabolic rate data, using the digital anthropometric phenotype data as input and the resting metabolic rate index as the label.
[0045] Step 2: Data preprocessing.
[0046] Preprocess the collected data, mainly including: handling missing values, removing abnormal data, data standardization, normalization, etc.
[0047] (1) Handling missing values: Delete samples with missing data or fill in the missing data.
[0048] (2) Removing abnormal data: Detect data outliers and delete or replace them to smooth the data.
[0049] (3) Data standardization processing: To eliminate the influence of dimensionality on model parameters, data standardization needs to be performed in advance.
[0050] Step 3: Construct different machine learning models.
[0051] Use the preprocessed data to construct different machine learning models, using the resting metabolic rate data as the label data and the digital anthropometric phenotype data as the input data to construct different regression models. The regression models here include but are not limited to multiple linear regression models, ridge regression models, Lasso regression models, decision tree regression models, and non-linear random forest regression models.
[0052] Step 4: Calculation of resting metabolic rate.
[0053] (1) Data input: Digital anthropometric data (such as gender, height, weight, age, chest circumference, waist circumference, hip circumference, etc.)
[0054] (2) Prediction model: Use the trained machine learning model (such as a regression model or a deep learning neural network) to calculate the resting metabolic rate.
[0055] Example formula:
[0056] Resting metabolic rate (RMR) = a · waist circumference + b · hip circumference + c · height + d · age + e;
[0057] a, b, c, d, e: Regression coefficients in the model, obtained from the training data.
[0058] (3) Result output: Output the resting metabolic rate and provide suggestions according to the standard health guidelines, for example: "The normal resting metabolic rate is 1,600 - 1,800 kcal / day (for men) and 1,200 - 1,400 kcal / day (for women)."
[0059] Step 5: Optimize the machine learning model.
[0060] Optimize the model by changing the parameters of different machine learning models or the feature selection method. The optimization criteria include the number of independent variables, the goodness of fit R 2 of the model, the root mean square error RMSE, etc.
[0061] In a specific embodiment, the machine learning model is trained using the data of the healthy population cohort. Taking the digital anthropometric phenotype data as the input, 157 kinds of three-dimensional human body measurement data are obtained. Respectively using the resting metabolic rate as the label, some test results of the established model are shown as follows:
[0062]
[0063] *HB, Harris - Benedict Formula; MS, Mifflin - St Jeor Formula; Henry, Henry Formula; Correlation, Pearson correlation coefficient between the predicted value and the true value.
[0064] It can be seen from Table 1 that R 2 measures the goodness of fit of the model / formula to the data. The closer R 2 is to 1, the better the prediction effect of the model / formula; RMSE measures the average deviation between the predicted value and the true value of the model / formula. The lower the value of RMSE, the more accurate the prediction of the model / formula. The determination coefficient R 2 values of the classical formulas for predicting the resting metabolic rate are mostly between 0.1 and
[0065] 0.3, the root mean square error RMSE is above 0.8, and the correlation coefficient with the true value is less than 0.6, indicating that the prediction effect of the classical formulas is poor and the error is large. However, the R 2 values of the multiple linear regression model, Lasso regression model, Ridge regression model, and support vector regression model constructed by the present invention for predicting the resting metabolic rate are all greater than 0.5 (for example, the R 2 of the random forest regression model is 0.68, the R 2 of the Lasso regression model is 0.64, the R 2 of the Ridge regression model is 0.64, and the R 2is 0.53), that is, the goodness of fit is relatively high, indicating that the model constructed by the present invention has a good prediction effect. Compared with the classical formula, the prediction effect has been greatly improved; the RMSE generated by all regression models is less than 0.7 (for example, the RMSE of the random forest regression model is 0.68, the RMSE of the Lasso regression model is 0.60, the RMSE of the Ridge regression model is 0.60, and the RMSE of the support vector regression model is 0.68), that is, the error is small, indicating that the model constructed by the present invention has high prediction accuracy. Compared with the classical formula, the prediction error has been greatly reduced; the Pearson correlation coefficients between the predicted values and the true values of all regression models are greater than 0.7 (for example, the correlation coefficient of the random forest regression model is 0.80, the correlation coefficient of the Lasso regression model is 0.80, the correlation coefficient of the Ridge regression model is 0.80, and the correlation coefficient of the support vector regression model is 0.74), that is, the fitting consistency is good, indicating that the prediction trend of the model constructed by the present invention is consistent with the actual data. Compared with the classical formula, the prediction consistency has been greatly improved.
[0066] References
[0067] [1] Frankenfield D, Roth-Yousey L, Compher C, et al. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review[J]. Journal of the American Dietetic association, 2005, 105(5): 775-789.
[0068] [2] FG H J A B. A biometric study of human basal metabolism Proc Natl Acad Sci USA4370-373. Harris, JA, and Benedict, FG(1918). A biometric study of human basal metabolism[J]. Proc Natl Acad Sci USA, 1918, 4: 370-373.
[0069] [3]Roza A M, Shizgal H M. The Harris Benedict equation reevaluated: resting energy requirements and the body cell mass[J]. The American journal of clinical nutrition, 1984, 40(1): 168-182.
[0070] [4]Mifflin M D, St Jeor S T, Hill L A, et al. A new predictive equation for resting energy expenditure in healthy individuals[J]. The American journal of clinical nutrition, 1990, 51(2): 241-247.
[0071] [5]O’Neill J E R, Corish C A, Horner K. Accuracy of resting metabolic rate prediction equations in athletes: A systematic review with meta-analysis[J]. Sports Medicine, 2023, 53(12): 2373-2398.
[0072] [6]Griffith R, Shean R, Petersen C L, et al. Validation of resting energy expenditure equations in older adults with obesity[J]. Journal of nutrition in gerontology and geriatrics, 2022, 41(2): 126-139。
Claims
1. A method for measuring resting metabolic rate based on digital anthropometric phenotype, characterized in that: Using digital anthropometric phenotypes, a machine learning model for predicting resting metabolic rate was established to obtain an accurate resting metabolic rate; the specific steps are as follows: (i) Digital anthropometric phenotypic data and resting metabolic rate data collection; The data used include resting metabolic rate and digital anthropometric phenotype data; the resting metabolic rate data are collected by human energy metabolism instrument; the digital anthropometric phenotype data are collected by 3D scanner or 3D reconstruction after 2D image collection, including gender, height, weight, age, chest circumference, abdominal circumference, neck circumference, left and right arm circumference, lower limb circumference, head height, back width, waist-to-hip ratio; Each digital anthropometric phenotype data is matched with the resting metabolic rate data measured simultaneously, with the digital anthropometric phenotype data as input and the resting metabolic rate index as a label; Step 2: Data preprocessing; The data collected in the first step are preprocessed, including: processing missing values, removing abnormal data, standardizing data, and normalizing data; among which: (1) Handling missing values: deleting samples with missing data or filling in missing data; (2) Clear abnormal data: detect data outliers and delete or replace them to smooth the data; (3) Data standardization: To eliminate the impact of dimensional differences on model parameters, the data is standardized; (4) Data transformation: transforming data into a form suitable for data modeling through smooth aggregation, data generalization, and normalization; including data normalization, feature selection, and dimensionality reduction; Step 3: Build different machine learning models; Use the preprocessed data to build different machine learning models; specifically, use resting metabolic rate data as label data and digital anthropometric phenotype data as input data to build different machine learning models; Step 4: Calculate resting metabolic rate; (1) Construct a mathematical model to predict resting metabolic rate using input digital anthropometric phenotypic data and established statistical models; (2) Perform data mapping. The core of the resting metabolic rate calculation formula depends on the training model. The formula is: Resting metabolic rate (RMR) = f(body circumference, body width, body length, weight, gender, age); Function f is a machine learning model trained based on a large number of real data samples; Step 5: Optimize the machine learning model; By changing the parameters of different machine learning models or feature selection methods, the model is optimized. The optimization criteria are the number of independent variables, the goodness of fit of the model, and the root mean square error (RMSE).
2. The method for measuring resting metabolic rate based on digital anthropometric phenotype according to claim 1, characterized in that: The machine learning model is a multiple linear regression model, a ridge regression model, a Lasso regression model, a decision tree regression model or a nonlinear random forest regression model.