Wearable intelligent weight reduction system and method based on syndrome differentiation and treatment of traditional Chinese medicine

Abstract

The invention provides a wearable intelligent weight reduction system and method based on traditional Chinese medicine syndrome differentiation and treatment, and the system comprises a multi-modal data collection hardware layer which is used for collecting the physiological data, tongue picture data and symptom data of a user; the traditional Chinese medicine AI dialectical reasoning engine is used for receiving, fusing and processing the multi-modal data and outputting a dynamic traditional Chinese medicine syndrome weight vector through a deep learning dialectical model; and the personalized precise intervention feedback layer is used for generating and pushing a personalized intervention scheme according to the syndrome weight vector. According to the method and the system, the problems that for the design of current health monitoring, on traditional Chinese medicine diagnosis digitization such as tongue diagnosis APP, isolated and static analysis is often performed, the analysis cannot be combined with continuous physiological data flow, and dynamic syndrome differentiation and closed-loop intervention cannot be formed are solved.

Description

Technical Field

[0001] This invention relates to the field of health monitoring technology, and more specifically, to a wearable intelligent weight loss system and method based on the principles of traditional Chinese medicine syndrome differentiation and treatment. Background Technology

[0002] Existing smart weight loss devices and methods, such as smart bracelets, body fat scales, and related apps, are primarily based on Western nutritional and exercise physiology theories. Their core logic is a "calorie deficit," which provides recommendations by monitoring calorie expenditure (e.g., steps, exercise duration) and intake (user-inputted diet). This approach has significant limitations: 1. Limited perspective: It ignores the vast differences in individual physiological and metabolic states. The same calorie intake will have completely different metabolic effects on people with different body types; 2. Lack of a holistic view of traditional Chinese medicine: It is unable to identify the user's TCM constitution type (such as phlegm-dampness constitution, stomach heat constitution, spleen deficiency constitution, etc.), and therefore cannot provide a root-cause-treating intervention plan based on TCM theory, such as clearing heat, resolving phlegm, strengthening the spleen, and soothing the liver. 3. Delayed and passive feedback: The recommendations are mostly post-event summaries, lacking real-time and forward-looking "prevention-oriented" interventions.

[0003] In existing technologies, although there have been some attempts to digitize TCM diagnosis in the process of health monitoring (such as tongue diagnosis apps), these are often isolated and static analyses that fail to be combined with continuous physiological data streams, thus failing to form dynamic syndrome differentiation and closed-loop intervention. Therefore, the market urgently needs a systematic solution that can integrate the advantages of TCM and Western medicine to achieve real-time, dynamic, and personalized health management. Therefore, we have made improvements to this by proposing a wearable intelligent weight loss system and method based on TCM syndrome differentiation and treatment. Summary of the Invention

[0004] The purpose of this invention is to address the problem that current health monitoring designs, such as those in TCM diagnosis digitization apps like tongue diagnosis apps, often involve isolated and static analysis that fails to integrate with continuous physiological data streams, thus hindering dynamic diagnosis and closed-loop intervention.

[0005] To achieve the above-mentioned objectives, the present invention provides the following technical solution: A wearable intelligent weight loss system and method based on the principles of traditional Chinese medicine syndrome differentiation and treatment, in order to improve the above-mentioned problems.

[0006] The application is as follows: A wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment includes: The multimodal data acquisition hardware layer is used to collect users' physiological data, tongue image data, and symptom data. The multimodal data acquisition hardware layer includes a non-invasive or minimally invasive continuous glucose monitoring module for monitoring users' blood glucose concentration change curves, a tongue image acquisition module for acquiring standardized tongue surface images, a basic physiological monitoring module for monitoring physiological data, a user input interface, and a data synchronization mechanism. The TCM AI diagnostic reasoning engine is used to receive and fuse multimodal data. It outputs a dynamic TCM syndrome weight vector through a deep learning diagnostic model. The TCM AI diagnostic reasoning engine includes a data fusion center for time alignment, cleaning and normalization of multi-source heterogeneous data. The personalized precision intervention feedback layer is used to generate and push personalized intervention plans based on the syndrome weight vector.

[0007] As a preferred technical solution of this application, the tongue image acquisition module is integrated into a wearable device or mobile terminal. The tongue image acquisition module includes a camera, a ring light, and a color calibration component. The color temperature range of the ring light is 5300K~5700K, and the color rendering index is not less than 95%. The illuminance of the ring light is 900~1100 Lux at the shooting distance.

[0008] As a preferred technical solution of this application, the physiological data monitored by the basic physiological monitoring module includes at least one of heart rate monitoring, heart rate variability, sleep data and stress data, and the user input interface is used to receive TCM constitution assessment questionnaire data and symptom information actively input by the user.

[0009] As a preferred technical solution of this application, the data synchronization mechanism is used to align data from different devices on the timeline, and includes: An initial unified time synchronization and coarse synchronization layer is used for clock synchronization during device pairing; A fine synchronization and alignment layer for continuous monitoring data is used to perform time interpolation and alignment of continuous monitoring data at different frequencies; A precise calibration layer for event-triggered data is used to associate discrete events triggered by users with continuous data streams.

[0010] As a preferred technical solution in this application, the deep learning dialectical model includes: The tongue image recognition sub-model extracts tongue color, coating color, and coating texture features based on a convolutional neural network. A time series data analysis sub-model is used to analyze time series data on blood glucose and heart rate variability based on a hybrid neural network. The time series data analysis sub-model includes a 3D convolutional neural network, a long short-term memory network, and a Transformer encoder. The dynamic syndrome fusion and determination module fuses multimodal features based on an attention mechanism and outputs a dynamic syndrome weight vector.

[0011] As a preferred technical solution of this application, the dynamic syndrome weight vector quantifies the contribution of the user's current multiple TCM syndromes in the form of a probability distribution. The TCM syndromes include at least two of the following: stomach heat, spleen deficiency, phlegm dampness, liver stagnation, and kidney qi deficiency.

[0012] As a preferred technical solution in this application, the personalized precision intervention feedback layer includes: The behavior suggestion generation module is used to generate diet, exercise and rest suggestions from a preset knowledge base based on the current syndrome weight vector; The physical intervention linkage module is used to communicate with external smart hardware, generate and send control commands to perform acupoint stimulation therapy.

[0013] As a preferred technical solution of this application, the physical intervention linkage module communicates with the intelligent acupoint stimulator via Bluetooth Low Energy. The control instructions are encapsulated in JSON format, including acupoint code, stimulation mode, intensity, duration and waveform parameters.

[0014] A smart weight loss method based on the principles of traditional Chinese medicine syndrome differentiation and treatment, using a wearable smart weight loss system, is characterized by the following steps: S1: Continuously collect user data through the multimodal data acquisition hardware layer; S2: Upload the data to the TCM AI diagnostic reasoning engine for analysis and processing; S3: Obtain the dynamic TCM syndrome weight vector output by the engine; S4: Generate and push personalized behavioral and physical intervention plans based on the syndrome weight vector; S5: Based on subsequent user data, dynamically update the syndrome assessment and intervention plan to form a closed-loop optimization.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: In the scheme of this application: By deeply integrating CGM dynamic blood glucose data with data from TCM tongue diagnosis and HRV, this system provides objective and quantitative assessment indicators for the TCM function of the "spleen and stomach," realizing the digitalization and dynamism of TCM syndrome differentiation. It also offers precise suggestions based on individual real-time physiological states and in accordance with TCM theory, allowing for personalized adjustments to the constitution from the root cause. Furthermore, by monitoring minute changes in physiological indicators in real time, it can intervene before significant weight gain or the onset of discomfort symptoms, reflecting the advanced TCM concept of "superior doctors treat disease before it manifests." It achieves an automated closed loop from "monitoring" to "analysis" to "intervention," and especially through its linkage with intelligent physical intervention devices, it elevates health management from the "suggestion" level to the "action" level. Attached Figure Description

[0016] Figure 1A general architecture block diagram of a wearable intelligent weight loss system based on the syndrome differentiation and treatment of traditional Chinese medicine is provided for this application; Figure 2 A flowchart of an intelligent weight loss method for a wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment, provided for this application; Figure 3 A schematic diagram of the tongue image acquisition module of a wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment provided in this application, displayed in an APP. Figure 4 A schematic diagram of the workflow of the AI ​​syndrome differentiation reasoning engine for a wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment provided in this application. Figure 5 This is an example diagram showing the dynamic syndrome weight vector and intervention suggestions in the APP interface of a wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment, which is provided in this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention.

[0018] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely to illustrate some embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. It should be noted that, unless otherwise specified, the embodiments, features, and technical solutions in the embodiments of the present invention can be combined with each other.

[0019] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0020] like Figure 1-5 As shown, this embodiment proposes a wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment, including: The multimodal data acquisition hardware layer is used to collect users' physiological data, tongue image data, and symptom data. The multimodal data acquisition hardware layer includes a non-invasive or minimally invasive continuous glucose monitoring module (CGM module) for monitoring the user's blood glucose concentration change curve, a tongue image acquisition module for acquiring standardized tongue surface images, a basic physiological monitoring module for monitoring physiological data, a user input interface, and a data synchronization mechanism. Specifically, in the multimodal data acquisition hardware layer, the CGM module can use a commercial patch sensor (such as the Sinocare continuous glucose meter) and communicate with the smartwatch via Bluetooth; the miniaturized tongue image acquisition module is a detachable watch strap accessory. Before brushing their teeth every morning, the user inserts their tongue into the center of the module's ring structure, and the module automatically triggers supplemental lighting and takes a picture, and performs image quality verification and color calibration through algorithms. The physiological data monitored by the basic physiological monitoring module includes heart rate, heart rate variability (HRV), acceleration information, sleep onset time, number of awakenings at night, total sleep duration, deep sleep ratio, stress, blood oxygen saturation, etc. This module acts as the master clock node, and usually collects data continuously at a high frequency (such as 1Hz) and adds local hardware timestamps. The user input interface is used to receive TCM constitution assessment questionnaire data and symptom information actively input by the user. The basic physiological monitoring module includes smart bracelets or watches, etc. Non-invasive or minimally invasive continuous glucose monitoring (CGM) module: used to monitor the glucose concentration change curve of interstitial fluid. This module has a built-in real-time clock to measure and store a blood glucose value at fixed intervals (such as every 5 minutes). The tongue image acquisition module is integrated into wearable devices or mobile terminals. The tongue image acquisition module includes a camera, a ring light, and a color calibration component. The color temperature range of the ring light is 5300K~5700K, and the color rendering index is not less than 95%. The illuminance of the ring light is 900~1100 Lux at the shooting distance.

[0021] Specific optical parameters: Lighting system: Integrated ring LED fill light, providing a color temperature of 5500K±200K (D55 standard white light) and a color rendering index (CRI) ≥95%, ensuring accurate reproduction of the color of the tongue and tongue coating, and avoiding interference from ambient light. Illuminance remains stable at 1000±100 Lux at a shooting distance (3-5cm), ensuring uniform image brightness and no shadows.

[0022] Imaging System: Uses the phone's main camera (or a customized high-definition camera) with a resolution of at least 12 megapixels, supporting autofocus and macro mode. During shooting, white balance and exposure parameters must be automatically locked, and beautification filters must be turned off to ensure image objectivity and repeatability.

[0023] Color calibration: Equipped with a physical 24-color standard color chart (such as X-Rite ColorChecker Classic) as a color calibration component. Before or periodically taking tongue images, users need to place the color chart in the shooting field of view, and the APP guides them through the color calibration process. The algorithm converts the image colors to a standard color space (such as sRGB), thereby eliminating color deviations caused by different device cameras.

[0024] The user input interface is integrated into the mobile APP to receive user-initiated data from constitution identification questionnaires (such as questionnaires on the nine constitutions in traditional Chinese medicine) and daily symptom information (such as appetite, sleep, and bowel movements). This type of data is discrete event data, and the APP adds an accurate timestamp when it is submitted.

[0025] The data synchronization mechanism is used to align data from different devices on the timeline. This system adopts a "layered hybrid synchronization mechanism" to solve the data time alignment problem of multi-source heterogeneous devices (continuous monitoring devices and discrete acquisition devices). This data synchronization mechanism consists of the following three layers: An initial unified time synchronization and coarse synchronization layer is used for clock synchronization during device pairing; Specifically, when a user first wears and pairs the device with the mobile app, all wearable devices (smart bracelets, CGM sensors) synchronize with the mobile phone via Bluetooth using the Network Time Protocol (NTP) to align their internal clocks with the phone's Coordinated Universal Time (UTC) and keep the error within the second range. This process ensures that all devices have a unified time base.

[0026] A fine synchronization and alignment layer for continuous monitoring data is used to perform time interpolation and alignment of continuous monitoring data at different frequencies; Specifically, in the master-slave device collaboration, the smartphone is designated as the master device (Hub), and the smart bracelet acts as the master clock node due to its highest sampling frequency. After each measurement, the CGM module, in addition to recording its own timestamp, also sends a signal with a measurement sequence number to the smart bracelet via Bluetooth Low Energy (BLE). Upon receiving the signal, the bracelet immediately records the received local high-precision timestamp. After subsequent data upload, the server-side algorithm can use these two timestamps for interpolation to align the CGM data points with the bracelet's high-frequency physiological data (such as HRV) with millisecond-level precision, effectively eliminating the physiological time difference caused by fluid delay in CGM.

[0027] A precise calibration layer for event-triggered data is used to associate discrete events triggered by users with continuous data streams; Specifically, for discrete events triggered by the user, such as tongue image capture and questionnaire / symptom input, software-triggered synchronization is used. When a user launches the APP to capture a tongue image or submit a questionnaire, the APP will immediately send a "synchronization event signal" via Bluetooth to the smart bracelet and CGM reader (if within range). After receiving the signal, these devices will record the latest data collection point sequence number at that moment (such as the XXXXth heart rate data point of the bracelet, the XXXth blood glucose value of the CGM). Through this signal, discrete events are accurately correlated with continuous physiological data streams, ensuring that during AI analysis, tongue image features and user symptoms can accurately correspond to the physiological state (such as blood glucose level, stress value) within a period of time before and after the capture.

[0028] Through the above synchronization mechanism, the system ensures that data from different sources, frequencies, and delays can be unified onto the same high-precision timeline, thereby enabling the AI ​​engine to perform truly meaningful cross-modal correlation and dynamic dialectical analysis.

[0029] The TCM AI diagnostic reasoning engine is used to receive and fuse multimodal data. It outputs a dynamic TCM syndrome weight vector through a deep learning diagnostic model. The TCM AI diagnostic reasoning engine includes a data fusion center for time alignment, cleaning and normalization of multi-source heterogeneous data. The data fusion center performs time-stamp-based alignment, cleaning, and normalization of time-series physiological data (blood glucose, HRV), TCM constitution assessment questionnaire data, image data (tongue image), and user input data to form a unified multimodal feature vector, which prepares for subsequent model analysis. The TCM AI diagnostic reasoning engine is deployed on a local terminal or cloud server to receive and process synchronized multimodal data collected by the hardware layer.

[0030] Specifically, the TCM AI syndrome differentiation and reasoning engine needs to be pre-trained. By cooperating with TCM colleges and universities, a large amount of tongue image data (corresponding to syndrome types) and synchronous physiological data (such as blood glucose and HRV) labeled by senior TCM doctors are collected to train the tongue image recognition sub-model and the time series data analysis sub-model respectively. Then, the final syndrome fusion and judgment module is trained using a multimodal dataset. Deep learning dialectical models include: The tongue image recognition sub-model extracts tongue color, coating color, and coating texture features based on a convolutional neural network. Specifically, the model is built upon a convolutional neural network (CNN) to extract deep features from standardized tongue images. It employs a deep residual network (ResNet) as the backbone feature extractor (e.g., preferably ResNet-50 or a customized network of similar depth) to address gradient degradation during deep network training and ensure accurate feature extraction. At the network's end, a global average pooling layer and multiple parallel fully connected layers, along with a softmax classifier, are connected to independently output quantified probability distributions for tongue color (e.g., pale red, red, purplish-red), tongue coating color (e.g., white, yellow, grayish-black), and tongue coating texture (e.g., thin, thick, greasy, dry). To further improve the recognition accuracy of key regions, the model integrates an attention mechanism, allowing the network to focus more on specific areas corresponding to different internal organs, such as the tongue's edge, tip, middle, and root. A time series data analysis sub-model is used to analyze time series data on blood glucose and heart rate variability based on a hybrid neural network. The time series data analysis sub-model includes a 3D convolutional neural network, a long short-term memory network, and a Transformer encoder. Specifically: First, a one-dimensional convolutional neural network (1D-CNN) is used as the front end to automatically extract short-term, high-dimensional features (such as sharp rises in blood glucose and short-term fluctuation patterns of HRV) from local time series and to achieve dimensionality reduction. Subsequently, the feature sequences extracted by 1D-CNN are input into a Long Short-Term Memory (LSTM) network. Due to its unique gating mechanism (input gate, forget gate, output gate), LSTM units are very good at capturing long-term dependencies such as the diurnal rhythm of blood glucose, metabolic patterns over long postprandial periods, and long-term dynamic changes in the autonomic nervous system represented by HRV. Finally, a Transformer encoder architecture is used to enhance the high-level features of the LSTM output. Its self-attention mechanism dynamically calculates the importance weights between features at different time points, thereby more accurately identifying the key time periods and patterns that have the greatest influence on the overall metabolic state. The output of this sub-model is a set of quantitative feature vectors reflecting the TCM concepts of "Qi, Blood, Yin, Yang" and the functional state of the internal organs.

[0031] The dynamic syndrome fusion and determination module fuses multimodal features based on an attention mechanism and outputs a dynamic syndrome weight vector.

[0032] The dynamic syndrome weight vector quantifies the contribution of a user's current multiple TCM syndromes in the form of a probability distribution. TCM syndromes include at least two of the following: stomach heat, spleen deficiency, phlegm dampness, liver stagnation, and kidney qi deficiency.

[0033] Specifically, the dynamic syndrome fusion and judgment module receives high-dimensional feature outputs (tongue appearance features, temporal physiological features, and embedded questionnaire and symptom features) from each sub-model. This module uses a multimodal feature fusion network based on an attention mechanism (not just a simple fully connected network or expert system) for final decision-making. This network first calculates an attention weight for each modality's features, dynamically evaluating the importance of data such as tongue appearance, blood glucose, HRV, and questionnaires to syndrome differentiation at the current moment. Then, all weighted feature vectors are concatenated and nonlinearly fused through several layers of fully connected neural networks, ultimately outputting a dynamic syndrome weight vector. This vector quantifies the contribution of the user's various current TCM syndromes (such as stomach heat excess, spleen deficiency with dampness, liver qi stagnation, kidney qi deficiency, etc.) in the form of a probability distribution (e.g., stomach heat: 0.60, spleen deficiency: 0.30, liver qi stagnation: 0.10), realizing the digitization, quantification, and dynamization of syndrome differentiation.

[0034] The personalized and precise intervention feedback layer is usually presented in the form of a mobile APP. It receives the syndrome weight vector output by the inference engine and generates and pushes personalized intervention suggestions. Specifically, the knowledge base of the personalized precision intervention feedback layer was jointly constructed by experts in nutrition, sports rehabilitation, and traditional Chinese medicine. It has established a mapping relationship database of "syndrome-intervention measures" to ensure the scientific validity and effectiveness of the recommendations. When users use this system, their data is encrypted and uploaded to the cloud for processing. The results are then returned to the user's app. The entire system provides users with a seamless, intelligent, and highly TCM-featured healthy weight loss experience.

[0035] The personalized precision intervention feedback layer includes: The behavior suggestion generation module is used to generate diet, exercise and rest suggestions from a preset knowledge base based on the current syndrome weight vector; Specifically, the behavior suggestion generation module generates personalized dietary, exercise, and lifestyle suggestions based on the current symptoms and matching them from a preset knowledge base; for example, "Currently, there is significant stomach heat, so it is recommended to avoid spicy foods and recommend mung bean soup." The physical intervention linkage module is used to communicate with external smart hardware, generate and send control commands to perform acupoint stimulation therapy.

[0036] Specifically, the physical intervention linkage module communicates with external smart hardware (such as a smart acupoint stimulator) to generate physical intervention instructions based on the current symptoms. For example, "It is recommended to stimulate the Zusanli acupoint for 15 minutes," and the accompanying equipment can be started with one click. This module is responsible for converting abstract symptom conclusions into specific parameters that the equipment can execute, and managing the entire communication and control process. The physical intervention linkage module communicates with the smart acupoint stimulator via Bluetooth Low Energy. The control commands are encapsulated in JSON format, including acupoint codes, stimulation modes, intensity, duration, and waveform parameters.

[0037] The specific technical implementation of the personalized precision intervention feedback layer includes: Communication protocols and connection methods: This system prioritizes Bluetooth Low Energy (BLE) as the core communication protocol for interaction with external smart hardware. BLE is characterized by low power consumption, widespread integration into mobile devices (phones), stable connections, and suitability for intermittent data transmission, making it a perfect fit for the usage scenarios of wearable health devices.

[0038] As an alternative or supplementary solution, the system can also support protocols such as Wi-Fi or Near Field Communication (NFC) (for quick pairing) to accommodate the connectivity capabilities of different hardware devices.

[0039] Specifically, the communication data format follows a lightweight data exchange standard based on JSON (JavaScript Object Notation), which is characterized by its good readability, ease of parsing, and extensibility.

[0040] Data interaction strictly follows the client-server (CS) model, where the mobile app acts as the GATT client and the smart hardware device acts as the GATT server.

[0041] Smart hardware devices need to predefine their GATT profile, which includes specific services, characteristics, and descriptors to implement specific functions.

[0042] Core services include: Device Control Service: Contains characteristics for writing and receives instructions from the APP; Status Feedback Service: Includes features for notification or reading, sending information such as the device's current status (e.g., running, low battery, completed) and error codes to the app.

[0043] Command and data packet structure: When a user confirms the execution of an intervention suggestion (such as "stimulate Zusanli acupoint for 15 minutes"), the APP will generate a structured command data packet, which is encapsulated in JSON format and sent to the corresponding feature value of the device control service through a write operation.

[0044] { "command_id": "CMD_20230901_102301", / / Unique command ID "command_type": "start_acupoint_stimulation", / / Command type: Start acupoint stimulation "parameters": { "acupoint_code": "ST36", / / Acupoint code (standardized, such as using WHO standards) "acupoint_name": "Zusanli", / / Acupoint name "stimulation_mode": "electro_acupuncture", / / Stimulation mode: electrical pulse "intensity_level": 3, / / Intensity level (1-5) "duration": 900, / / Duration (in seconds) "waveform_pattern": "dispersive_dense" / / Waveform pattern: Dispersive-dense wave }, "metadata": { "syndrome_focus": "Stomach_Heat", / / Target syndrome: Stomach heat "timestamp": "2023-09-01T10:23:01Z" / / Command generates timestamp } } Security and pairing mechanisms: Upon first use, the device needs to be paired and bound via Bluetooth through the app to ensure the legitimacy and uniqueness of the connection.

[0045] The transmission of all critical control commands can be protected using the built-in encryption function of the BLE protocol to prevent unauthorized access and manipulation, thus ensuring user security.

[0046] User Interface (UI): Used to visualize diagnostic results, historical trends, and various intervention guidelines.

[0047] A smart weight loss method based on the principles of traditional Chinese medicine syndrome differentiation and treatment, using a wearable smart weight loss system, is characterized by the following steps: S1: Continuously collect user data through the multimodal data acquisition hardware layer, including user physiological data, tongue image data, TCM constitution assessment questionnaire data, and receive symptom data input by the user. S2: Upload the data to the TCM AI diagnostic reasoning engine for analysis and processing. The reasoning engine fuses multi-source data and uses a deep learning diagnostic model for feature extraction and analysis. S3: Obtain the dynamic TCM syndrome weight vector output by the engine. The inference engine outputs a dynamic and quantified TCM syndrome weight vector. S4: Based on the syndrome weight vector, generate and push personalized behavioral and physical intervention plans. The personalized precision intervention feedback layer generates and pushes personalized plans that include diet, exercise and physical intervention measures based on the syndrome weight vector. S5: Based on subsequent user data, dynamically update the syndrome assessment and intervention plan to form a closed-loop optimization.

[0048] In this application, during use: In the multimodal data acquisition hardware layer, the CGM module can use a commercially available patch sensor (such as the Sinocare continuous glucose meter) and communicate with the smartwatch via Bluetooth. The miniaturized tongue image acquisition module is a detachable watchband accessory. Before brushing their teeth each morning, the user inserts their tongue into the center of the module's ring structure. The module automatically triggers supplemental lighting and takes a picture, and performs image quality verification and color calibration through algorithms. The TCM AI syndrome differentiation and reasoning engine needs to be pre-trained. Through cooperation with TCM colleges, a large amount of tongue image data (corresponding to syndrome types) and synchronous physiological data (such as blood glucose and HRV) labeled by senior TCM doctors are collected to train the tongue image recognition sub-model and the time-series data analysis sub-model, respectively. Subsequently, the multimodal dataset is used to train the final syndrome fusion judgment module. The knowledge base of the personalized precision intervention feedback layer is jointly constructed by experts in nutrition, sports rehabilitation, and TCM, establishing a mapping relationship library of "syndrome-intervention measures" to ensure the scientificity and effectiveness of the recommendations. When users use this system, their data is encrypted and uploaded to the cloud for processing, and the results are returned to the user's APP. The entire system provides users with a seamless, intelligent, and highly TCM-featured healthy weight loss experience.

[0049] The above embodiments are only used to illustrate the present invention and are not intended to limit the technical solutions described herein. Although the present invention has been described in detail with reference to the above embodiments, the present invention is not limited to the specific embodiments described above. Therefore, any modifications or equivalent substitutions to the present invention, as well as all technical solutions and improvements that do not depart from the spirit and scope of the invention, are covered within the scope of the claims of the present invention.

Claims

1. A wearable intelligent weight loss system and method based on traditional Chinese medicine syndrome differentiation and treatment, characterized in that, include: The multimodal data acquisition hardware layer is used to collect users' physiological data, tongue image data and symptom data. The multimodal data acquisition hardware layer includes a non-invasive or minimally invasive continuous glucose monitoring module for monitoring the user's blood glucose concentration change curve, a tongue image acquisition module for acquiring standardized tongue surface images, a basic physiological monitoring module for monitoring physiological data, a user input interface and a data synchronization mechanism. The TCM AI diagnostic reasoning engine is used to receive and fuse the multimodal data, and output a dynamic TCM syndrome weight vector through a deep learning diagnostic model. The TCM AI diagnostic reasoning engine includes a data fusion center for time alignment, cleaning and normalization of multi-source heterogeneous data. A personalized precision intervention feedback layer is used to generate and push personalized intervention plans based on the syndrome weight vector.

2. The wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 1, characterized in that, The tongue image acquisition module is integrated into a wearable device or mobile terminal. The tongue image acquisition module includes a camera, a ring light, and a color calibration component. The color temperature range of the ring light is 5300K~5700K, and the color rendering index is not less than 95%. The illuminance of the ring light is 900~1100 Lux at the shooting distance.

3. The wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 2, characterized in that, The physiological data monitored by the basic physiological monitoring module includes at least one of heart rate, heart rate variability, sleep data, and stress data. The user input interface is used to receive TCM constitution assessment questionnaire data and symptom information actively input by the user.

4. The wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 1, characterized in that, The data synchronization mechanism is used to align data from different devices on the timeline, and includes: An initial unified time synchronization and coarse synchronization layer is used for clock synchronization during device pairing; A fine synchronization and alignment layer for continuous monitoring data is used to perform time interpolation and alignment of continuous monitoring data at different frequencies; A precise calibration layer for event-triggered data is used to associate discrete events triggered by users with continuous data streams.

5. A wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 1, characterized in that, The deep learning dialectical model includes: The tongue image recognition sub-model extracts tongue color, coating color, and coating texture features based on a convolutional neural network. A time-series data analysis sub-model is used to analyze time-series data on blood glucose and heart rate variability based on a hybrid neural network. The time-series data analysis sub-model includes a 3D convolutional neural network, a long short-term memory network, and a Transformer encoder. The dynamic syndrome fusion and determination module fuses multimodal features based on an attention mechanism and outputs a dynamic syndrome weight vector.

6. The wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 5, characterized in that, The dynamic syndrome weight vector quantifies the contribution of the user's current multiple TCM syndromes in the form of a probability distribution. The TCM syndromes include at least two of the following: stomach heat, spleen deficiency, phlegm dampness, liver stagnation, and kidney qi deficiency.

7. The wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 1, characterized in that, The personalized precision intervention feedback layer includes: The behavior suggestion generation module is used to generate diet, exercise and rest suggestions from a preset knowledge base based on the current syndrome weight vector; The physical intervention linkage module is used to communicate with external smart hardware, generate and send control commands to perform acupoint stimulation therapy.

8. A wearable intelligent weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in claim 7, characterized in that, The physical intervention linkage module communicates with the intelligent acupoint stimulator via Bluetooth Low Energy. The control commands are encapsulated in JSON format, including acupoint codes, stimulation modes, intensity, duration, and waveform parameters.

9. A smart weight loss method based on a wearable smart weight loss system based on traditional Chinese medicine syndrome differentiation and treatment as described in any one of claims 1-9, characterized in that, Includes the following steps: S1: Continuously collect user data through the multimodal data acquisition hardware layer; S2: Upload the data to the TCM AI diagnostic reasoning engine for analysis and processing; S3: Obtain the dynamic TCM syndrome weight vector output by the engine; S4: Generate and push personalized behavioral and physical intervention plans based on the syndrome weight vector; S5: Based on subsequent user data, dynamically update the syndrome assessment and intervention plan to form a closed-loop optimization.

Images