A body weight management method for analyzing coupling metabolism by observing scale body multi-frequency impedance spectrum

Abstract

The present application is suitable for the field of health information technology and biological signal processing, and discloses a body weight management method for observing scale and knowing body multi-frequency impedance spectrum analysis coupled metabolism, comprising the following steps: S1, multi-frequency bioelectrical impedance signal acquisition and pretreatment; S2, impedance spectrum feature extraction and body composition inversion; S3, time series data file construction; S4, primary and secondary index coupled metabolism analysis; S5, dynamic identification of metabolic state; S6, comprehensive evaluation and personalized intervention; the present application, multi-frequency impedance spectrum depth analysis, helps to improve the accuracy of body composition inversion; primary and secondary index dynamic coupling analysis helps to realize early warning of metabolic state; hierarchical classification state recognition system helps to refine the granularity of metabolic analysis; multi-dimensional information fusion evaluation helps to build a full-cycle precise intervention closed loop.

Description

Technical Field

[0001] This invention belongs to the field of health information technology and biological signal processing, and particularly relates to a weight management method that uses multi-frequency impedance spectrum analysis coupled with metabolism based on body weight observation. Background Technology

[0002] With the increasing health awareness of residents, body fat scales have become a common device for daily health monitoring in households. Bioelectrical impedance analysis is the core measurement principle of current body fat scales. Existing technologies have conducted considerable research on the accuracy of body fat measurement and data application. Patent document CN121845552A discloses a method, device, and medium for real-time monitoring of multi-parameter human body fat data, which uses multi-frequency impedance amplitude-phase sequence extraction to extract spectral characterization to calculate body fat percentage. Patent document CN120974438A discloses a method, medium, and device for monitoring the status of target users, which uses multi-device collaborative data collection and performs difference calculation and weight fusion. Patent document CN121938563A discloses a dynamic metabolic rate prediction method and system based on multimodal biological data, which solves the nutritional requirement ratio through multi-source data fusion. While these technologies have made some progress in measurement accuracy, multi-source data fusion, and metabolic prediction, the following technical limitations still exist:

[0003] Insufficient spectral analysis depth and inadequate utilization of raw impedance information: Existing body fat scale technologies mostly directly output the final body composition index values ​​without performing in-depth spectral analysis of the raw multi-frequency bioelectrical impedance signals; most schemes only use impedance amplitudes at different frequencies for simple weighted calculations, without fully exploring frequency domain features such as phase angle spectrum and spectral energy distribution, and without achieving accurate fitting of physiological parameters such as extracellular fluid resistance and intracellular fluid resistance through equivalent circuit models; this limitation leads to insufficient underlying signal support for body composition inversion, making it difficult to capture the impedance characteristic changes corresponding to subtle metabolic changes such as fat breakdown and water transfer, thus limiting measurement sensitivity and the accuracy of subsequent metabolic analysis;

[0004] The analysis of indicators is static and fragmented, lacking a dynamic coupling and correlation mechanism: existing technologies generally treat indicators such as weight, body fat percentage, and water percentage as independent variables for static evaluation, focusing only on the absolute value and daily variation of a single indicator, without establishing a dynamic coupling and correlation model between indicators; existing solutions do not consider the time lag characteristics of indicator changes, and it is also difficult to quantify the linkage strength and causal direction between indicators, making it difficult to identify changes in metabolic state from indicator combination patterns; this limitation means that body fat data can only reflect the current body composition, making it difficult to reveal the internal linkage law of metabolic processes, resulting in a significant lag in early warning of metabolic abnormalities, and it is also difficult to distinguish weight fluctuations of different causes;

[0005] The identification of metabolic stages is limited by a single dimension and lacks the ability to analyze states in a hierarchical and categorized manner: Existing weight management technologies have difficulty effectively distinguishing between weight loss caused by fat consumption and water loss, and even more so the ability to identify the efficient fat decomposition and metabolism stage; most programs only use weight loss and body fat percentage reduction as the criteria for determining the effectiveness of fat loss, without systematically attributing the causes of changes in fat burning rate, and it is difficult to distinguish between different types of changes such as body self-regulation, insufficient energy and deviation from the program, and intervention and adjustment lack precise basis;

[0006] While intervention programs are generally generic, they lack a multi-dimensional, precise matching mechanism: existing weight management programs are mostly general dietary templates, making it difficult to accurately adapt to the dynamic changes in an individual's metabolic state. Although some programs incorporate subjective feedback, they lack a multi-dimensional evaluation system that integrates objective data, intervention dosage, and subjective feelings, and lack a unified quantitative evaluation index. This limitation leads to insufficient matching between intervention intensity and metabolic state, easily resulting in low fat loss efficiency due to insufficient intervention or excessive metabolic stress due to over-intervention, and also makes it difficult to achieve dynamic closed-loop regulation throughout the entire weight management cycle.

[0007] Therefore, a weight management method that combines observation of body weight with multi-frequency impedance spectrum analysis and metabolic coupling is needed to solve the above problems. Summary of the Invention

[0008] The purpose of this invention is to provide a weight management method based on observation of body weight and multi-frequency impedance spectrum analysis coupled with metabolism, in order to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A weight management method based on observation of body weight and multi-frequency impedance spectral analysis coupled with metabolism includes the following steps:

[0011] S1. Multi-frequency bioelectrical impedance signal acquisition and preprocessing: Acquire bioelectrical impedance amplitude and phase angle data at at least two frequency points, perform noise reduction, baseline correction and outlier removal preprocessing on the raw impedance signal, and generate a standardized impedance signal matrix.

[0012] S2. Impedance Spectrum Feature Extraction and Volume Component Inversion: Perform frequency domain decomposition on the standardized impedance signal to extract multi-dimensional spectral feature parameters; fit physiological equivalent parameters based on the bioimpedance equivalent circuit model to invert and generate multiple sets of volume component structural indicators.

[0013] S3. Construction of Time Series Data Archives: Establish an independent body composition time series database for each subject, and perform trend extraction, period identification and fluctuation decomposition processing on the time series data;

[0014] S4. Coupling Metabolic Analysis of Primary and Secondary Indicators: The structural indicators of body composition are divided into three major indicators and three minor indicators. A dynamic coupling correlation model of primary and secondary indicators is established. The coupling correlation parameters, time lag characteristic parameters and synergistic coupling characterization parameters between indicators are calculated, and multi-level early warning intervals are divided.

[0015] S5. Dynamic identification of metabolic state: Based on the results of coupling analysis and time-series change characteristics, identify the high fat-burning metabolic state and its state transition events, classify the attribution type of state change, and construct a multi-stage metabolic cycle model.

[0016] S6. Comprehensive assessment and personalized intervention: Integrating objective body composition data, intervention dosage suitability and subjective body sensation information, the morning three knowledge comprehensive assessment index (CAI) is generated, and the graded assessment results and corresponding energy-renewing nutritional intervention plan are output in combination with the current metabolic stage.

[0017] A complete technical chain from signal acquisition to intervention output has been constructed, forming a three-layer progressive technical architecture of spectrum analysis, coupling analysis, and state intervention. This enables the upgrade of body fat scale data from static numerical measurement to dynamic metabolic assessment, providing full-process technical support for weight management.

[0018] In a further technical solution, in step S1, the noise reduction adopts Kalman filtering, the baseline correction adopts the sliding window mean method, and the outlier removal adopts the threshold determination method based on 3σ statistical distribution; the preprocessing parameters include the filter covariance parameter, the sliding window width, and the outlier determination threshold.

[0019] This scheme defines three core methods and parameter systems for signal preprocessing, which can effectively reduce the impact of skin contact noise and baseline drift, improve impedance signal quality, and provide a stable and reliable data foundation for subsequent spectrum analysis.

[0020] In a further technical solution, in step S2, the frequency domain decomposition includes two methods: fast Fourier transform and wavelet transform. The extracted spectral features include impedance amplitude spectrum, phase angle spectrum, and spectral energy distribution. The bioimpedance equivalent circuit model adopts the Cole-Cole model, and solves for four types of equivalent parameters—extracellular fluid resistance, intracellular fluid resistance, characteristic time constant, and diffusion coefficient—through nonlinear fitting.

[0021] By defining two implementation paths for frequency domain decomposition and the parameter composition of the equivalent circuit model, and combining dual frequency domain analysis with model fitting, accurate solutions for physiological equivalent parameters can be achieved, thereby improving the accuracy of underlying signal support for volume composition inversion.

[0022] In a further technical solution, in step S2, the body composition structural indicators are divided into two categories: the three major indicators and the three minor indicators. The three major indicators include body mass index, body fat percentage, and visceral fat level. The three minor indicators include water content, bone mineral content, and protein content. The two types of indicators together constitute the complete dimension of body composition assessment.

[0023] The hierarchical classification of body composition indicators was clarified, and a hierarchical indicator system consisting of primary evaluation indicators and secondary verification indicators was established, providing a clear basis for the division of indicator dimensions for subsequent coupling analysis and state identification.

[0024] In a further technical solution, in step S4, the coupling correlation parameters are calculated using cross-correlation analysis and divided into two levels, significant coupling and strong coupling, based on their absolute values; the time delay characteristic parameters are solved by iterating through a preset time interval, and the time delay corresponding to the highest coupling strength is taken as the optimal time delay; the cooperative coupling characterization parameters include two categories, synchronous coupling index and cooperative change index, both of which are divided into multi-level early warning intervals based on their absolute values.

[0025] It defines the core computational logic of coupling analysis and two types of collaborative parameters. It quantifies the linkage relationship of indicators through cross-correlation analysis and time-delay traversal. The multi-dimensional coupling parameters can be adapted to different analysis scenarios, providing a quantitative benchmark for early warning of metabolic abnormalities.

[0026] In a further technical solution, in step S5, the high fat-burning state is identified using a three-element feature combination mode: the first element is a false upward trend in body fat percentage, the second element is a downward trend in water percentage, and the third element is an upward trend in protein content; when the three-element feature direction requirements are met simultaneously for multiple consecutive measurement cycles and the weight continues to decrease, it is determined that the state has entered the high fat-burning state.

[0027] A three-dimensional feature recognition model for the high-fat-burning state was proposed, breaking through the traditional cognitive limitation that only a decrease in body fat percentage counts as fat loss. It can identify the metabolic stage of efficient fat decomposition from the body composition linkage model and clarify the golden window period for fat loss.

[0028] In a further technical solution, in step S5, the attribution type of the state change is divided into three categories, which are determined in order of priority: the first category is the scheme deviation type, which is characterized by a sudden interruption of the metabolic state and the reversal of the indicators; the second category is the self-repair type, which is characterized by a slowdown in the metabolic rate and the aggravation of existing symptoms; the third category is the qi and blood deficiency type, which is characterized by a gradual slowdown in the metabolic rate and no new discomfort symptoms.

[0029] Three types of metabolic state changes were classified into attribution types, and differentiated judgment characteristics were established. These characteristics can distinguish between metabolic fluctuations caused by the body's own regulation and those caused by external factors, providing targeted directions for adjusting intervention programs.

[0030] In a further technical solution, in step S5, the multi-stage metabolic cycle includes four stages in sequence: the high fat-burning period, the fat-burning shutdown period, the rest period, and the normal period. Each stage corresponds to different indicator change characteristics and coupling states. The stage transition is triggered by preset judgment conditions. Each stage transition is set with a corresponding observation period and confirmation period. The cycle repeats to form a metabolic state evolution mode.

[0031] A four-stage metabolic cycle pattern was constructed, the transition trigger logic and transition rules of each stage were clarified, and the weight management process was broken down into identifiable and controllable continuous stages, forming an evolutionary framework for the gradual optimization of metabolic state.

[0032] In a further technical solution, in step S6, the Morning Three Knowledges Comprehensive Assessment Index (CAI) is calculated using a weighted linear combination method. The weight dimensions include data-driven dimension, dosage-adaptive dimension, and body sensation dimension. Different metabolic stages correspond to different graded threshold ranges. The revitalizing nutritional intervention program is divided into three modes according to metabolic stages: full meal replacement, stepped refeeding, and self-maintenance. The intervention intensity is adjusted by floating based on the assessment grade.

[0033] It defines the calculation method of the comprehensive assessment index and the three-stage intervention model, deeply linking the intervention plan with the metabolic stage and assessment grading, so as to achieve the matching of nutritional supply and metabolic status, which helps to improve the effectiveness and safety of the intervention.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] This invention, through multi-frequency impedance spectrum in-depth analysis, helps improve the accuracy of body composition inversion: by collecting impedance amplitude and phase angle data at multiple frequencies, combining fast Fourier transform and wavelet transform to extract multi-dimensional spectral features, and then fitting the Cole-Cole equivalent circuit model to obtain physiological parameters such as extracellular fluid resistance and intracellular fluid resistance. This processing path utilizes the differences in the ability of currents of different frequencies to penetrate cells. Low-frequency signals reflect the state of extracellular fluid, while high-frequency signals reflect the characteristics of the whole cell. By mining deep physiological information from the signal source, it helps improve the accuracy and sensitivity of body composition index inversion and captures the changes in impedance characteristics corresponding to subtle metabolic changes such as fat decomposition and water transfer.

[0036] This invention utilizes dynamic coupling analysis of primary and secondary indicators to facilitate early warning of metabolic state. It establishes a primary evaluation indicator plus secondary verification indicator coupling analysis framework, calculates the coupling coefficient between indicators using a cross-correlation algorithm, locates the optimal response time by traversing time lag intervals, and outputs two types of collaborative coupling characterization parameters. This analysis system can quantify the linkage strength and causal direction between indicators, identify metabolic state anomalies from the perspective of indicator change patterns, and, through a graded early warning mechanism, detect metabolic fluctuations in advance by observing changes in the linkage relationship between indicators before any single indicator shows obvious abnormalities, thus reserving a time window for intervention and adjustment.

[0037] This invention, a hierarchical classification state recognition system, helps to refine the granularity of metabolic analysis: it proposes a three-dimensional feature recognition model of increased body fat percentage, decreased water percentage, and increased protein, which can identify the high-fat-burning state of efficient fat decomposition and can distinguish between weight changes caused by fat consumption and simple water loss; at the same time, it establishes a state change attribution classification system to locate the core reasons for changes in fat-burning rate, reduce the subjectivity and ambiguity of experience judgment, and provide targeted directions for intervention and adjustment.

[0038] This invention utilizes multi-dimensional information fusion assessment to facilitate the construction of a closed-loop system for precise intervention throughout the entire lifecycle. It establishes a multi-dimensional information fusion mechanism, generating a comprehensive assessment index through weighted fusion of data-driven, dosage-appropriate, and subjective experience dimensions. This incorporates objective data, intervention suitability, and subjective feelings to achieve quantitative evaluation of metabolic status. Furthermore, it establishes a phased dynamic nutritional intervention program, employing different intervention modes at different metabolic stages and adjusting the intervention intensity based on the assessment grading. This forms a complete closed loop of status identification, comprehensive assessment, precise intervention, and status feedback, which helps improve the individual suitability and implementation safety of the intervention program.

[0039] To more clearly illustrate the structural features and effects of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0042] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0043] Example 1

[0044] like Figure 1 As shown, this embodiment of the invention provides a weight management method based on multi-frequency impedance spectrum analysis coupled with metabolism, including multi-frequency bioelectrical impedance signal acquisition and preprocessing steps, impedance spectrum feature extraction and body composition inversion steps, time series data archive construction steps, primary and secondary index coupled metabolic analysis steps, metabolic state dynamic identification steps, and comprehensive assessment and personalized intervention steps.

[0045] In this embodiment, the complete execution flow of the method is as follows:

[0046] S1. Multi-frequency bioelectrical impedance signal acquisition and preprocessing: Using the multi-frequency bioelectrical impedance analysis module built into the body fat scale, AC signals at three characteristic frequencies—5kHz, 50kHz, and 250kHz—are applied to the human body, and the impedance amplitude and phase angle data at each frequency are acquired. The raw impedance signals are then subjected to noise reduction, baseline correction, and outlier removal processes to generate a standardized multi-frequency impedance signal matrix. The low-frequency signal reflects the extracellular fluid state, the mid-frequency signal comprehensively reflects whole-cell characteristics, and the high-frequency signal penetrates the cell membrane to reflect the intracellular fluid state. The combined analysis of these three frequencies provides a multi-dimensional signal basis for body composition calculation.

[0047] S2. Impedance Spectrum Feature Extraction and Volume Component Inversion: The standardized impedance signal is subjected to frequency domain decomposition to extract multi-dimensional spectral feature parameters such as impedance amplitude spectrum, phase angle spectrum and spectral energy distribution; nonlinear fitting is performed based on the bioimpedance equivalent circuit model to solve the physiological equivalent parameters, and then multiple sets of volume component structural indices are generated by calculating through the preset inversion model.

[0048] S3. Construction of Time Series Data Archives: Establish an independent body composition time series database for each subject, continuously storing the body composition structural index data measured daily; perform trend extraction, period identification and fluctuation decomposition processing on the time series data to separate long-term trend terms, periodic fluctuation terms and random disturbance terms.

[0049] S4. Coupling Metabolic Analysis of Primary and Secondary Indicators: The structural indicators of body composition are divided into two categories: major primary indicators and minor tertiary indicators. A dynamic coupling correlation model between primary and secondary indicators is established. The coupling correlation parameters, time lag characteristic parameters and synergistic coupling characterization parameters between each pair of indicators are calculated, and multi-level early warning intervals are divided according to the absolute value of the parameters.

[0050] S5. Dynamic identification of metabolic state: Based on the results of coupling analysis and the change characteristics of time series data, identify the major fat-burning metabolic state and its state transition events; classify the attribution types of state changes and construct a multi-stage metabolic cycle model.

[0051] S6. Comprehensive Assessment and Personalized Intervention: Integrating objective body composition data, intervention dosage suitability, and subjective body sensation information, the Morning Three Knowledges Comprehensive Assessment Index (CAI) is generated through weighted calculation; combined with the current metabolic stage, the graded assessment results and corresponding revitalizing nutritional intervention plans are output, and the intervention intensity is dynamically adjusted according to the assessment results.

[0052] Example 2

[0053] The difference between this embodiment and Embodiment 1 is that the core parameter configuration and execution logic of signal preprocessing in step S1 are further refined, and the quantitative judgment criteria for various preprocessing methods are clarified.

[0054] In this embodiment, noise reduction employs Kalman filtering, which uses two iterative steps—state update and observation update—to suppress high-frequency noise caused by fluctuations in skin contact resistance and electrode polarization. Baseline correction uses a sliding window mean method, updating the baseline point by point to eliminate slowly varying baseline drift. Outlier removal uses a threshold determination method based on 3σ statistical distribution to identify and remove abnormal sampling points.

[0055] The core parameters of the Kalman filter are fixed as follows: the state noise covariance Q is set to 0.01 to control the noise level of state prediction; the observation noise covariance R is set to 0.1 to control the noise level of the observation data; and the initial error covariance P0 is set to 1, representing the error covariance at the initial moment of filtering. The sliding window width W for baseline correction is set to 7 sampling periods to calculate the baseline reference value. The threshold coefficient k for outlier removal is set to 3.0, i.e., the 3σ criterion is used; the number of consecutive outlier detection points N is set to 2, meaning that points exceeding the threshold consecutively are considered outliers.

[0056] During execution, the mean μ and standard deviation σ of the data within the sliding window are calculated. When the values ​​of two consecutive sampling points exceed the range of μ ± 3σ, they are identified as outliers and removed. Linear interpolation of adjacent effective values ​​is then used to fill in the missing positions. Through the above standardized preprocessing procedure, the quality of the impedance signal can be improved, providing stable and consistent input data for subsequent spectrum analysis.

[0057] Example 3

[0058] The difference between this embodiment and embodiment 1 is that the spectral feature extraction and equivalent circuit model fitting logic in step S2 are further refined, as well as the time-series data decomposition method and parameter system in step S3.

[0059] In this embodiment, the frequency domain decomposition uses a combination of fast Fourier transform and wavelet transform to complement each other. The extracted spectral features include impedance amplitude spectrum, phase angle spectrum, and spectral energy distribution. The bioimpedance equivalent circuit model adopts the Cole-Cole model, and the four types of equivalent parameters are solved by fitting the nonlinear least squares method.

[0060] The parameters for the Fast Fourier Transform (FFT) were set as follows: the number of sampling points N_FFT was 2048, the Hanning window function was used to suppress spectral leakage, and the window overlap rate was 50% to extract steady-state spectral features. The wavelet transform used the db4 wavelet basis function, and the number of decomposition layers J was set to 5 to extract local features in the time-frequency domain.

[0061] The initial fitting parameters for the Cole-Cole model were set as follows: the initial value of extracellular fluid resistance R_e0 was 200Ω, the initial value of intracellular fluid resistance R_i0 was 500Ω, and the initial value of the characteristic time constant τ0 was 1.0×10⁻ 5s, the initial value of the dispersion coefficient α0 is 0.8.

[0062] The expression for the Cole-Cole impedance model is:

[0063]

[0064] In the formula: Angular frequency The corresponding complex impedance; The extracellular fluid resistance is obtained by fitting a low-frequency impedance signal and reflects the conductivity of the extracellular fluid. The intracellular fluid resistance is obtained by fitting a high-frequency impedance signal and reflects the conductivity of the intracellular fluid. The characteristic time constant characterizes the relaxation properties of tissue impedance. The diffusion coefficient reflects the frequency dispersion characteristics of tissue impedance, and its value ranges from 0 to 1.

[0065] The fitting process aims to minimize the sum of squared residuals between the impedance spectrum calculated by the model and the measured impedance spectrum. The optimal equivalent parameters are obtained by iterative solution, and then the volume composition structural index is generated by calculating through the preset volume composition inversion model.

[0066] Time series data processing employed two methods: Holt-Winters exponential smoothing and STL decomposition, with periodicity verification performed using ACF / PACF. The parameters for Holt-Winters exponential smoothing were set as follows: horizontal smoothing coefficient α_HW = 0.3, reflecting the responsiveness to current observations; trend smoothing coefficient β_HW = 0.1, reflecting the responsiveness to trend changes; seasonal smoothing coefficient γ_HW = 0.1, reflecting the responsiveness to seasonal variations; and the seasonal period L = 7 days, corresponding to the circadian rhythm of human metabolism.

[0067] The parameters for the STL decomposition method are set as follows: the number of robust iterations N_robust is 2, used for robust handling of outliers; the seasonal window width W_seasonal is 7, covering a complete cycle; and the trend window width W_trend is 15, used to smooth long-term trends.

[0068] The parameters for the ACF / PACF periodicity check are set as follows: the maximum lag order max_lag is 14 days, the significance threshold is ±1.96 / √n, corresponding to a 95% confidence interval, and the presence of a periodic component is determined when three consecutive lag orders exceed the threshold.

[0069] By using the frequency domain analysis and time series decomposition methods described above, deep physiological information can be extracted from the original impedance signal, thereby improving the accuracy of body composition inversion and the sensitivity of metabolic analysis.

[0070] Example 4

[0071] The difference between this embodiment and embodiment 1 is that it further refines the classification system of primary and secondary indicators, the calculation rules of coupling analysis, the two types of collaborative coupling parameters, and the multi-level anomaly judgment logic in step S4.

[0072] In this embodiment, the structural indicators of body composition are divided into two categories: the three major indicators and the three minor indicators. The three major indicators include body mass index, body fat percentage, and visceral fat level, which serve as the main dimensions for assessing physical condition. The three minor indicators include water content, bone mineral content, and protein content, which serve as the verification dimensions for metabolic activity.

[0073] The coupling correlation parameter is calculated using cross-correlation analysis, and the formula is as follows:

[0074]

[0075] In the formula: For time delay The coupling coefficient between the i-th primary indicator and the j-th secondary indicator; The covariance function characterizes the degree of covariance between the two variables; Main indicator Standard deviation; secondary indicator The standard deviation.

[0076] Coupling strength is divided into two levels based on absolute value: when the absolute value of the coupling coefficient is greater than 0.5, it is considered significant coupling, indicating that there is a moderate or greater mutual influence between the indicators; when the absolute value of the coupling coefficient is greater than 0.8, it is considered strong coupling, indicating that there is a high correlation between the indicators.

[0077] The time delay characteristic parameters are solved by iterating through the interval [-7, +7] days with a step size of 1 day. The time delay corresponding to the maximum absolute value of the coupling coefficient is taken as the optimal time delay. This reflects the time difference between the responses of the two indicators.

[0078] Cooperative coupling characterization parameters include two categories: synchronous coupling index and cooperative change index. The first category is the synchronous coupling index (SCI), calculated as follows:

[0079]

[0080] In the formula: The coupling coefficient at the optimal time delay; The absolute value of the optimal time delay is used; this index combines coupling strength and time delay decay characteristics to reflect the degree of synchronous coupling between indicators. Its warning rule is as follows: a coupling warning is triggered when the absolute value of the synchronous coupling index is greater than or equal to 0.6, and a high-risk warning is triggered when it is greater than or equal to 0.8.

[0081] The second category is the collaborative change index, calculated using the following formula:

[0082]

[0083] In the formula: It represents the maximum absolute value of the coupling coefficient within the time delay range, reflecting the strength of the coupling correlation; The sign of the coupling coefficient at the optimal time delay is: a positive value represents a change in the same direction, and a negative value represents a change in the opposite direction.

[0084] The Co-change Index adopts a four-level early warning classification: the absolute value in the range of [0, 0.4) is the normal level, indicating that the coupling relationship between indicators is within the normal fluctuation range; the absolute value in the range of [0.4, 0.6) is the attention level, indicating that the coupling relationship between indicators has some abnormality and needs attention; the absolute value in the range of [0.6, 0.8) is the warning level, indicating that the coupling relationship between indicators is significantly abnormal and the metabolic state may change significantly; the absolute value in the range of [0.8, 1.0] is the high-risk level, indicating that the coupling relationship between indicators is severely abnormal and the metabolic state fluctuates violently.

[0085] This plan focuses on monitoring the coordinated change characteristics of three key indicator pairs, with the corresponding reference intervals as follows:

[0086] Body fat percentage-water percentage co-change index (SCI) 24 ): Under normal high fat-burning conditions, it should be a negative value, with an absolute value in the range of [0.5, 0.9], reflecting the inverse linkage between the increase in body fat percentage and the decrease in water percentage;

[0087] Body fat percentage-protein co-change index (SCI) 25 ): Under normal high fat-burning conditions, it should be a negative value, with an absolute value in the range of [0.4, 0.7], reflecting the inverse linkage between the increase in body fat percentage and the increase in protein.

[0088] Moisture-Bone Mineral Co-change Index (SCI) 46 ): Under normal conditions, the absolute value is in the range of [0.3, 0.6], reflecting the body's water and electrolyte balance.

[0089] Based on the results of coupling analysis, a three-tiered metabolic abnormality judgment rule was established: the judgment condition for severe metabolic abnormality is that all three major indicators are abnormal, and at least two of the three minor indicators are synergistically abnormal, and the absolute value of the synergistic change index of at least one pair of key indicators is greater than 0.6; the judgment condition for occult metabolic risk is that only the visceral fat level is high, the other three major indicators are all within the normal range, all three minor indicators are normal, and the absolute value of the synergistic change index of key indicators is less than 0.5; the judgment condition for metabolic change warning is that at least two of the three minor indicators are abnormal and the absolute value of the synergistic change index of at least one pair of key indicators is greater than 0.6, or the change of a single key synergistic change index within 3 days exceeds 0.3.

[0090] Example 5

[0091] The difference between this embodiment and embodiment 1 is that it further refines the three-element feature recognition mode, core discrimination conditions and physiological mechanism of the high fat burning state in step S5.

[0092] In this embodiment, the high-fat-burning state is identified using a three-factor combination model. The first characteristic corresponds to body fat percentage, which shows a false increase. The physiological mechanism is that fat cells shrink in size, resulting in water loss from the cells, and the proportion of fat measured by impedance method increases temporarily, with a typical increase of 0.10% to 0.20% per day. The second characteristic corresponds to water percentage, which shows a decrease. The physiological mechanism is that the intracellular water content decreases during fat breakdown, accompanied by increased excretion of metabolic products, with a typical decrease of 0.10% to 0.20% per day. The third characteristic corresponds to protein content, which shows an increase. The physiological mechanism is that after the reduction of adipose tissue, the relative proportion of lean body mass in total weight increases, with a typical increase of 0.05% to 0.15% per day.

[0093] The core criteria for identifying a state of intense fat burning include four aspects: First, the ternary characteristic amplitude condition: within a continuous observation window of 3 days or more, the cumulative increase in body fat percentage is greater than or equal to 0.3%, the cumulative decrease in water percentage is greater than or equal to 0.3%, and the cumulative increase in protein is greater than or equal to 0.1%. Second, the weight change condition: within the observation window, weight continuously decreases, with an average daily decrease of 0.4-0.8 kg. Third, the bone mineral content coordination condition: bone mineral content shows a steady decrease or remains stable, with a daily fluctuation of less than 0.03 kg. Fourth, the coupling verification condition: the body fat percentage-water percentage coordinated change index is negative and its absolute value is in the range of 0.5-0.9, and the body fat percentage-protein coordinated change index is negative and its absolute value is in the range of 0.4-0.7. When all four conditions are met simultaneously, it is determined that the state of intense fat burning has been entered.

[0094] The physiological essence of the high fat-burning state is as follows: Under the dual drive of negative energy balance and nutritional intervention, the body initiates a metabolic mode with fat oxidation as the main energy supply pathway; glycerol and fatty acids produced by fat decomposition enter the tricarboxylic acid cycle for oxidation and energy supply. During this process, extracellular fluid water decreases along with the excretion of metabolic products, resulting in a false increase in body fat percentage due to water loss; at the same time, under the condition of sufficient protein supply, the body maintains a positive nitrogen balance, muscle protein synthesis metabolism is active, and protein content indicators increase.

[0095] Visceral fat levels are assessed using a three-tiered grading system, with different levels corresponding to different health risks. The standard range corresponds to a level of less than 6 for women and less than 9 for men, indicating good metabolic status; continued balanced nutrition and appropriate exercise are recommended. The above-standard range corresponds to levels 6-10 for women and 9-14 for men, indicating that health has been somewhat affected and attention should be paid to weight and dietary control. The high-risk range corresponds to levels 11 or higher for women and 15 or higher for men, indicating a higher metabolic risk; strict control of weight and dietary structure is required.

[0096] The aforementioned hierarchical identification system can accurately pinpoint the state of intense fat burning, providing a basis for refined intervention.

[0097] Example 6

[0098] The difference between this embodiment and Embodiment 1 is that: the three attribution criteria for the changes in the high fat-burning state in step S5 are further refined, and quantitative rules for the two transitional states of rest period and metabolic shutdown period are added.

[0099] In this embodiment, when the rate of high fat burning slows down or is interrupted, it is divided into three types according to the cause, and judged in order of priority: first, it is judged whether it meets the characteristics of the deviation from the plan, then the self-repair type, and finally the qi and blood deficiency type. At any given time, it can only belong to one type.

[0100] The first category is deviation from the plan, characterized by sudden interruption, requiring only 1-2 days; the core indicators are a sharp increase in water content, with a daily increase of more than 0.3%, a continuous rebound in bone mineral content, with a daily rebound of more than 0.02 kg, a rapid decrease in body fat percentage, accompanied by weight rebound; the physical symptoms may include gastrointestinal discomfort; the core cause is the failure to strictly implement the intervention plan, and external dietary factors disrupting the metabolic state.

[0101] The second type is the self-repair type, characterized by a gradual slowdown over several days; the core indicators are a continuous decrease in protein, with a daily decrease of 0.1%-0.5%, abnormal fluctuations in bone mineral content, with daily fluctuations greater than 0.03 kg, and an absolute value of the synergistic change index greater than 0.6; the physical symptoms are an exacerbation of existing chronic symptoms; the core cause is that the body activates its self-protection mechanism, shifting metabolic resources to tissue repair.

[0102] The third type is the Qi and Blood Deficiency type, characterized by a gradual slowdown over several days; the core indicators are: daily change in body fat percentage is less than 0.1%, water percentage rises slightly with a daily increase of 0.1%-0.5%, bone mineral content is highly stable with a daily fluctuation of less than 0.02 kg, and protein is stable or slightly increased; the physical symptoms are: no new discomfort symptoms; the core cause is a relative lack of energy supply to the body, leading to an adaptive downward adjustment of the metabolic rate.

[0103] At the same time, two quantitative judgment criteria for transitional states are added. The first type is the rest period, and the core judgment criteria are that the body fat percentage changes from rising to falling, the water percentage changes from falling to stabilizing, and the protein percentage changes from rising to stabilizing, and the daily weight change is less than 0.3 kg / day; the auxiliary judgment criteria are that the variance of the synergistic change index is less than 0.05 for 3 consecutive days, the body fat percentage decreases for 2 consecutive days, and the water percentage is stable or slightly rising for 2 consecutive days.

[0104] The second category is the fat-burning shutdown period. The core criteria for judgment are a continuous decrease in body fat percentage with a cumulative decrease of more than 0.3%, a continuous increase in water percentage with a cumulative increase of more than 0.3%, and a continuous increase in bone mineral content for 3 consecutive days with a cumulative increase of more than 0.03 kg. The auxiliary criteria are a daily weight change of less than 0.2 kg / day and a variance of the synergistic change index of less than 0.05.

[0105] The above-mentioned quantitative judgment system can objectively distinguish the different causes of fat burning fluctuations, providing a basis for targeted adjustments to the intervention plan.

[0106] Example 7

[0107] The difference between this embodiment and Embodiment 1 is that it further refines the stage characteristics, transition triggering conditions, and buffer period rules of the four-stage metabolic cycle in step S5.

[0108] In this embodiment, the metabolic cycle includes four phases in sequence: the major fat-burning phase, the fat-burning shutdown phase, the rest phase, and the normal phase. The core indicators of the major fat-burning phase are a false increase in body fat percentage, a decrease in water percentage, an increase in protein, and a sustained and stable decrease in weight; the coupling characteristics are a strong negative coupling between body fat percentage and water percentage, and between body fat percentage and protein, with the synergistic change index in the range of attention to warning; the metabolic meaning is that fat oxidation for energy is dominant, which is the golden window for fat loss.

[0109] The core characteristics of the fat-burning shutdown period are that body fat percentage changes from rising to falling, water percentage changes from falling to rising, and the rate of weight loss slows down significantly; the coupling state is characterized by a gradual decrease in coupling strength and a return of the synergistic change index to the normal range; the metabolic meaning is that the metabolic mode changes from fat oxidation as the main source to mixed energy supply, entering an adjustment and transition period.

[0110] The core indicators of the rest period are: body fat percentage continues to decrease and tends to stabilize, and water percentage and protein gradually return to baseline levels; coupling status is characterized by the coupling relationship tending to be stable, and the synergistic change index stabilizing within the normal range; metabolic meaning is that the metabolic system gradually restores balance, and various indicators converge to baseline levels.

[0111] The core characteristics of the normal period are that all six indicators are stable within the individual's normal baseline range with small fluctuations; the coupling state characteristics are that the coupling relationship between the indicators is stable and the synergistic change index is consistently within the normal range; the metabolic meaning is that metabolism is in a dynamic equilibrium state and the physical condition has completed a round of improvement.

[0112] The triggering conditions and buffer period settings for the four-stage transition are as follows:

[0113] Transition from the major fat-burning period to the fat-burning shutdown period: The transition is triggered when any of the following conditions are met and last for 2 consecutive days: body fat percentage changes from increasing to decreasing, water percentage changes from decreasing to increasing, daily weight loss is less than 0.1 kg, and key synergistic change index is less than 0.4; set a 1-2 day observation period, and complete the transition after confirming that the status is stable;

[0114] The transition from the fat-burning off-season to the rest-season occurs when body fat percentage continues to decrease, water percentage continues to rise, and the rate of change of various indicators gradually slows down, and the coupling relationship tends to stabilize.

[0115] Transition from the rest period to the normal period: The transition is triggered when the six major indicators are stable within ±2% of the individual baseline for three consecutive days, and the comprehensive assessment index is in the superior range of the corresponding stage for three consecutive days; a three-day confirmation period is set to complete the transition after confirming metabolic stability.

[0116] Transition from normal period to major fat-burning period: After a professional assessment confirms that the body is in a suitable condition, a new round of intervention cycle is initiated.

[0117] The four-stage cycle repeats itself. Each complete cycle represents a step-by-step improvement in physical condition, forming a spiral optimization and evolutionary pattern of metabolic state.

[0118] Example 8

[0119] The difference between this embodiment and embodiment 1 is that the comprehensive assessment system of the three knowledges upon waking in the morning, the calculation rules of CAI parameters, and the energy-renewing nutritional intervention program in step S6 are further refined.

[0120] In this embodiment, the Morning Three-Knowledge Comprehensive Assessment System includes three layers: the first layer is knowing the data, which means acquiring the body composition index data, frequency domain characteristic parameters and coupling analysis results of the day; the second layer is knowing the dosage, which means determining the amount of nutritional products consumed and the dynamic adjustment rules of the day based on the metabolic stage and assessment level; the third layer is knowing the body feeling, which means collecting the user's subjective body feeling score as a supplementary dimension of the assessment system.

[0121] The Morning Three Knowledges Comprehensive Assessment Index (CAI) is calculated using a weighted linear combination method, and the formula is as follows:

[0122]

[0123] In the formula: The weights of the Data-Driven Index (DDI) reflect the overall state of objective physiological indicators. The weights for the dose-appropriateness index (DQI) reflect the degree of fit of the nutritional intervention program. The weights of the Body Sensation Index (BSI) reflect subjective body sensation; and satisfy the following conditions: .

[0124] The calculation rules for each sub-index are as follows:

[0125] The Data-Driven Index (DDI) is calculated using the following formula:

[0126]

[0127] In the formula: The normalization function maps the index deviation to the [0,1] interval; , , These are the reference values ​​for body fat percentage, water percentage, and bone mineral content for the corresponding metabolic stages; , , , And satisfy .

[0128] The dose-fit index (DQI) is calculated using the following formula:

[0129]

[0130] In the formula: This refers to the actual amount of nutritional products ingested. This is the optimal intake for this metabolic phase; When the usage deviation reaches 100%, the DQI decreases to 50.

[0131] The Body Feeling Index (BSI) is calculated using the following formula:

[0132]

[0133] In the formula: , , , , The five categories of physical sensations are rated on a scale of 5: mental state, hunger, bowel movements, sleep quality, and pain / discomfort. The BSS is the overall physical sensation score.

[0134] Different metabolic stages correspond to different CAI (Cost Acid Intake) grading intervals. The grading criteria for the peak fat-burning phase are: [65, 75] is excellent, [55, 65) is good, [45, 55) is of concern, and less than 45 is a warning. The grading criteria for the adjustment phase are: [70, 85] is excellent, [60, 70) is good, [50, 60) is of concern, and less than 50 is a warning. The grading criteria for the normal phase are: [80, 95] is excellent, [70, 80) is good, [60, 70) is of concern, and less than 60 is a warning.

[0135] The Energizing Nutritional Intervention Program is divided into three modes based on metabolic stages:

[0136] The first type is the full meal replacement mode, which is suitable for the peak fat-burning period. The core rule is to mainly consume the formula nutritional products throughout the day, accompanied only by warm water, and avoid other foods and drinks. Supplements can be taken temporarily when hungry. The basic dosage is 6 packets per day, which is adjusted according to the CAI level. For the good level, add 2 packets; for the attention level, add 4 packets; and for the warning level, add 6 packets.

[0137] The second method is the step-by-step refeeding model, which is suitable for the adjustment period. It involves gradually restoring a regular diet in three stages: initially, the diet consists mainly of light liquid foods; in the middle stage, high-quality protein and vegetables are gradually added; and in the later stage, normal soft foods are restored. Nutritional products are maintained before breakfast and dinner, with a basic dosage of 4 packets per day, 2 packets before breakfast and dinner, and high-oil, high-salt, and high-sugar foods are avoided throughout the process.

[0138] The third type is the self-maintenance mode, which is suitable for normal periods. The core rule is to establish a long-term balanced diet and you can choose whether to maintain your intake of nutritional products. The basic dosage is 0-4 packets per day, and you can choose as needed.

[0139] The method is implemented through four layers of logic: the bottom layer is multi-frequency impedance data acquisition, the second layer is spectrum analysis and volume composition inversion, the third layer is coupling analysis and state identification, and the fourth layer is comprehensive evaluation and intervention output, providing data support and decision-making basis for users and managers.

[0140] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A weight management method based on observation of body weight and multi-frequency impedance spectrum analysis coupled with metabolism, characterized in that, Includes the following steps: S1. Multi-frequency bioelectrical impedance signal acquisition and preprocessing: Acquire bioelectrical impedance amplitude and phase angle data at at least two frequency points, perform noise reduction, baseline correction and outlier removal preprocessing on the raw impedance signal, and generate a standardized impedance signal matrix. S2. Impedance Spectrum Feature Extraction and Volume Component Inversion: Perform frequency domain decomposition on the standardized impedance signal to extract multi-dimensional spectral feature parameters; fit physiological equivalent parameters based on the bioimpedance equivalent circuit model to invert and generate multiple sets of volume component structural indicators. S3. Construction of Time Series Data Archives: Establish an independent body composition time series database for each subject, and perform trend extraction, period identification and fluctuation decomposition processing on the time series data; S4. Coupling Metabolic Analysis of Primary and Secondary Indicators: The structural indicators of body composition are divided into primary evaluation indicators and secondary verification indicators. A dynamic coupling correlation model of primary and secondary indicators is established. The coupling correlation parameters, time lag characteristic parameters and synergistic coupling characterization parameters between indicators are calculated, and multi-level early warning intervals are divided. S5. Dynamic identification of metabolic state: Based on the results of coupling analysis and time-series change characteristics, identify the high fat-burning metabolic state and its state transition events, classify the attribution type of state change, and construct a multi-stage metabolic cycle model. S6. Comprehensive Assessment and Personalized Intervention: Integrate objective body composition data, intervention dosage suitability, and subjective body sensation information to generate a comprehensive assessment index, and combine the current metabolic stage to output graded assessment results and corresponding nutritional intervention plans.

2. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S1, the noise reduction adopts Kalman filtering, the baseline correction adopts the sliding window mean method, and the outlier removal adopts the threshold determination method based on 3σ statistical distribution. Preprocessing parameters include the filter covariance parameter, the sliding window width, and the outlier detection threshold.

3. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S2, the frequency domain decomposition includes two methods: fast Fourier transform and wavelet transform. The extracted spectral features include impedance amplitude spectrum, phase angle spectrum and spectral energy distribution. The bioimpedance equivalent circuit model adopts the Cole-Cole model, and solves for four types of equivalent parameters, namely extracellular fluid resistance, intracellular fluid resistance, characteristic time constant and diffusion coefficient, through nonlinear fitting.

4. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S2, the body composition structural indicators are divided into two categories: the three major indicators and the three minor indicators. The three major indicators include body mass index, body fat percentage, and visceral fat level. The three minor indicators include water content, bone mineral content, and protein content. Together, these two types of indicators constitute the complete dimensions of body composition assessment.

5. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S4, the coupling correlation parameters are calculated using cross-correlation analysis and divided into two levels, significant coupling and strong coupling, based on their absolute values. The time delay characteristic parameters are solved by iterating through the preset time interval, and the time delay corresponding to the highest coupling strength is taken as the optimal time delay. The cooperative coupling characterization parameters include two categories: synchronous coupling index and cooperative change index, both of which are divided into multi-level early warning intervals based on their absolute values.

6. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S5, the high fat-burning state is identified using a three-factor combination mode: the first factor is an increasing trend in body fat percentage, the second factor is a decreasing trend in water content, and the third factor is an increasing trend in protein content. When the three-factor direction requirements are met simultaneously for multiple consecutive measurement cycles and the weight continues to decrease, it is determined that the state has entered the high fat-burning state.

7. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S5, the attribution types of the state changes are divided into three categories, which are determined in order of priority: the first category is the plan deviation type, which is characterized by a sudden interruption of the metabolic state and the reversal of the indicators; the second category is the self-repair type, which is characterized by a slowdown in the metabolic rate and the aggravation of existing symptoms; the third category is the qi and blood deficiency type, which is characterized by a gradual slowdown in the metabolic rate and no new discomfort symptoms.

8. The weight management method based on multi-frequency impedance spectral analysis coupled with metabolism according to claim 1, characterized in that, In step S5, the multi-stage metabolic cycle includes four stages in sequence: the high fat-burning period, the fat-burning shutdown period, the rest period, and the normal period. Each stage corresponds to different indicator change characteristics and coupling states. The stage transition is triggered by preset judgment conditions. Each stage transition is set with a corresponding observation period and confirmation period. The cycle repeats to form a metabolic state evolution mode.

9. The weight management method based on multi-frequency impedance spectrum analysis coupled with metabolism according to claim 1, characterized in that, In step S6, the comprehensive assessment index is the Morning Three Knowledges Comprehensive Assessment Index (CAI), which is calculated using a weighted linear combination method. The weight dimensions include data-driven dimension, dosage-appropriate dimension, and body sensation dimension. Different metabolic stages correspond to different graded threshold ranges. The nutritional intervention program is the Energy Renewal Method Standardized Intervention System, which is divided into three modes according to metabolic stages: full meal replacement, step-by-step refeeding, and self-maintenance. The intervention intensity is adjusted by floating based on the assessment grade.

Images