Digital medical health parameter monitoring device

Abstract

The invention discloses a digital medical health parameter monitoring device in the technical field of health parameter monitoring, and the device comprises a first collection unit which is used for collecting the physiological data of the temple area of a user; the second acquisition unit is used for acquiring differential pressure of the nose wings of the user; the third acquisition unit is used for acquiring eye movement data of eyeballs of the user, and the eye movement data comprises reflected light information and rotation information; the fourth acquisition unit is used for acquiring pulse condition data of the user; the system further comprises a master control center and an output module. The main control center is used for carrying out time-space alignment on the multi-source data and carrying out multi-modal feature fusion; matching a preset pathological model library based on the fusion result, and outputting a user health assessment result; and the output module is used for transmitting the evaluation result to terminal equipment for display or cloud storage. According to the scheme, the comprehensiveness of health monitoring is improved through multi-node collaborative collection of the eyes, the nose, the wrist and the like.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of health parameter monitoring, and specifically relates to a digital medical health parameter monitoring device. BACKGROUND

[0002] In the traditional medical field, the monitoring of human health parameters mainly relies on professional equipment in hospitals. These devices are usually large in size and complex in operation, and require patients to be examined in a specific hospital environment. Moreover, they can only obtain data at specific time points, and cannot realize real-time and continuous monitoring of health parameters. For example, early electrocardiogram monitoring requires patients to lie on a hospital bed and connect complex electrodes and wires, which not only brings inconvenience to patients, but also cannot capture the electrocardiogram changes of patients in daily activities.

[0003] With the improvement of people's health awareness and the high incidence of chronic diseases, the demand for real-time monitoring of personal health status is increasing. Many chronic diseases such as hypertension, diabetes, and heart disease require patients to pay long-term attention to their health parameters and adjust treatment plans according to the data. The traditional monitoring method is difficult to meet this demand.

[0004] The sensor is the core component of the health parameter monitoring device. In recent years, sensor technology has made significant progress, with smaller size, higher sensitivity, and better accuracy. For example, the continuous upgrading of optical heart rate sensors, blood pressure sensors, and blood glucose sensors enables the monitoring device to more accurately obtain various physiological parameters of the human body. At the same time, the emergence of new sensors such as flexible sensors and biosensors provides more possibilities for the diversification and comfort of monitoring devices, enabling the monitoring of more health parameters such as skin conductance and body temperature.

[0005] With the development of technology, wearable health monitoring devices have gradually emerged, such as smart bracelets and smart watches, which can conveniently monitor basic health parameters such as heart rate, exercise steps, and sleep, providing users with preliminary health information. For example, the patent with publication number CN110169763B discloses a digital medical health parameter monitoring device, which mainly consists of a centralized remote monitoring system and multiple health parameter monitoring node components. The device uses a sensor group, an elastic structure, and a breathable layer to ensure good contact between the sensor and the human body, ensuring the accuracy and stability of data monitoring, and avoiding the adverse effects of sweat on the skin. However, the above-mentioned scheme only focuses on the monitoring of a single body part and cannot fully integrate the health information of multiple parts of the human body, resulting in a monitoring blind area.

[0006] In view of the defects of the prior art, there is an urgent need for a device that can comprehensively, accurately, and real-time monitor the health parameters of multiple parts of the human body to meet the growing demand for health management. SUMMARY

[0007] In order to solve the above problems, the purpose of the present application is to provide a digital medical health parameter monitoring device, which can break through the limitations of traditional single monitoring equipment, effectively integrate multi-site health information, build a more comprehensive and accurate health evaluation system, and provide better health monitoring services for users.

[0008] In order to achieve the above purpose, the technical scheme of the present application is as follows: A digital medical health parameter monitoring device, comprising: a first wearing mechanism and a second wearing mechanism; the first wearing mechanism is worn on the eyes of a user, and the first wearing mechanism is provided with a first acquisition unit, a second acquisition unit and a third acquisition unit, the first acquisition unit is used for acquiring physiological data of the temple region of the user, the second acquisition unit is used for acquiring differential pressure of the alae nasi of the user, and the third acquisition unit is used for acquiring eye movement data of the eyeball of the user, the eye movement data comprising reflection light information and rotation information; the second wearing mechanism is worn on the wrist of the user, and the second wearing mechanism is provided with a fourth acquisition unit, the fourth acquisition unit is used for acquiring pulse condition data of the user; Further comprising a master control hub and an output module; the master control hub is used for spatio-temporal alignment of multi-source data, the multi-source data comprising physiological data, differential pressure, eye movement data and pulse condition data, and multi-modal feature fusion of eye movement temple-breathing alae nasi-spectrum eyeball-pulse condition is performed; based on the fusion result, a preset pathological model library is matched, and a user health evaluation result is output; the output module is used for transmitting the evaluation result to a terminal device for display or cloud storage.

[0009] The above scheme has the following beneficial effects: 1. The existing equipment (such as a single smart bracelet) usually focuses on a single part (such as wrist heart rate), resulting in isolated data and inability to capture the overall related health signals, while the present scheme sets a first wearing mechanism (such as smart glasses) and a second wearing mechanism (such as a smart bracelet or watch), realizing the synchronous acquisition of multi-site health parameters of the human body. For example, the first wearing mechanism can simultaneously monitor the physiological data of the temple, the differential pressure of the alae nasi, and the reflection light and rotation information of the eyeball; the second wearing mechanism is responsible for pulse condition data acquisition, thereby comprehensively covering key parts such as the head, eyes and wrist, effectively eliminating the blind area of single monitoring, and providing a more complete health data atlas for the user.

[0010] 2. The present scheme, through the innovative eye movement-breathing-spectrum-pulse multi-modal feature fusion technology, deeply excavates the potential relationship between the data, and realizes more accurate health status evaluation. For example, combining the rotation information of the eyeball and the pulse condition data, the early signs of heart function abnormalities or nervous system diseases can be more sensitively captured, while traditional single monitoring equipment cannot achieve such accurate comprehensive analysis.

[0011] 3. Unlike traditional static monitoring methods involving scheduled, fixed-location examinations in hospitals, this solution supports continuous wear by users in daily life scenarios. It can collect health parameters in real time and continuously, promptly capturing subtle changes in these parameters and providing dynamic health management data for patients with chronic diseases (such as hypertension and heart disease). For example, continuous monitoring of pulse and respiratory differential pressure can provide early warning of the risk of arrhythmias or respiratory disease attacks, offering users an appropriate time for intervention—something difficult to achieve with traditional monitoring methods.

[0012] Furthermore, the first acquisition unit includes an optical heart rate sensor for acquiring pulse wave and blood oxygen saturation data in the temple area; The second acquisition unit includes a miniature differential pressure sensor for detecting changes in nasal respiratory pressure differential and deriving respiratory rate and depth; The third acquisition unit includes an infrared light source and a CMOS image sensor, used to capture the eye's reflectance spectrum and motion trajectory; The fourth acquisition unit includes a piezoelectric sensor array for acquiring pulse waveforms of the radial artery in the wrist, including pulse rate, pulse pressure, and rhythm characteristics.

[0013] Furthermore, the steps for constructing the pathological model library include: S1. Multi-source medical data acquisition and annotation: Acquire eye movement data, respiratory data, spectral data and pulse data in a clinical setting, and have medical experts annotate them with pathological labels; S2. Cross-modal feature extraction and screening: Extract time-domain, frequency-domain, and spatial features from the data, and screen a subset of features strongly correlated with the target disease through feature importance analysis; S3. Multi-algorithm fusion model training: Based on the selected feature subset, train an integrated model that includes supervised learning and deep learning to generate a pathology classifier; S4. Clinical Validation and Threshold Calibration: Validate model performance through retrospective data and prospective trials, and dynamically adjust warning thresholds based on gold standard diagnostic results; S5. Incremental Learning and Compliance Deployment: Continuously access new case data to update the model and deploy it to the central control unit under a privacy protection framework.

[0014] Furthermore, the main control center is also used to use waveform decomposition algorithms to decompose pulse data into amplitude-time domain feature vectors of main wave, tidal wave, and dicrotic wave; and to identify pathological patterns of slippery pulse, wiry pulse, and hesitant pulse based on convolutional neural networks.

[0015] Furthermore, the main control center performs health assessments based on the following logic: The spatiotemporal alignment is achieved through timestamp synchronization and spatial coordinate correction, ensuring that multi-source data are consistent in time and space. Feature extraction algorithms are used to perform multimodal fusion of eye-tracking data, respiratory data, spectral data, and pulse data to generate multidimensional health indicators; The fusion results are matched with a pre-set pathological model library, which is built based on machine learning algorithms and covers the feature thresholds of hypertension, diabetes, heart disease and nervous system diseases. When assessing anomalies, the current fusion result is compared with the feature threshold. If the indicator deviation exceeds the preset threshold, a health risk is determined.

[0016] Furthermore, the central control unit is also used to perform the following operations when a health assessment fails: Determine if the failure is caused by data interference: If abnormal light spots appear in the spectral data, it is determined to be ambient light interference, and adjustment instructions are generated to optimize the sensitivity of each sensor; To determine if the failure was caused by user activity: if the respiratory data and pulse data show non-physiological fluctuations, it is determined to be motion artifacts, and the data is corrected by filtering through an algorithm.

[0017] Furthermore, the central control unit is also used to predict user health risks based on historical data trend analysis; If the current data is compared with historical benchmarks, and the eye movement data is abnormal or the pulse trend deviates from the linear relationship, an early warning will be issued.

[0018] Furthermore, the first wearing mechanism is also equipped with an image acquisition unit, which is used to collect the user's lip color. The main control center is also used to assist in judging the user's health status based on the user's lip color.

[0019] Furthermore, the first and second wearing mechanisms are made of flexible biocompatible materials.

[0020] Furthermore, it also includes an alarm module, which generates audiovisual alarm signals when the health assessment results exceed a preset safety range. Attached Figure Description

[0021] Fig. 1 This is a three-dimensional structural diagram of the eyeglass frame in the digital medical health parameter monitoring device of the present invention.

[0022] Fig. 2 This is a three-dimensional structural diagram of the smartwatch in the digital medical health parameter monitoring device of the present invention.

[0023] Fig. 3 This is a rear view of the eyeglass frame in the digital medical health parameter monitoring device of the present invention.

[0024] The reference numerals in the accompanying drawings include: 1, eyeglass frame; 2, patch; 3, nose pad; 4, smartwatch; 5, fourth acquisition unit; 101, third acquisition unit. Detailed Implementation

[0025] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0026] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "vertical", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0027] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0028] The following detailed description illustrates the specific implementation method: The basic implementation examples are as follows: Figs. 1-3 The device illustrates a digital medical health parameter monitoring system, primarily comprising a first wearing mechanism and a second wearing mechanism. Preferably, both the first and second wearing mechanisms are made of flexible biocompatible materials. The first wearing mechanism is worn around the user's eyes. In this embodiment, the first wearing mechanism is specifically an eyeglass frame 1. The first wearing mechanism is equipped with a first acquisition unit, a second acquisition unit, and a third acquisition unit 101. The first acquisition unit is used to collect physiological data from the user's temple area. Specifically, a patch 2 is welded and fixed to the inner side of the temple of the eyeglass frame 1. The first acquisition unit uses an optical heart rate sensor (model MAX30102) embedded in the patch 2, fitting snugly against the temple area. The first acquisition unit outputs the following data: pulse wave amplitude, blood oxygen saturation (SpO2), and perfusion index (PI). Pulse wave amplitude reflects the strength of the pressure wave generated when the heart contracts and delivers blood to the peripheral arteries; blood oxygen saturation is an important indicator for measuring the degree of oxygen binding in the blood and can be used to assess the body's respiratory function and blood circulation; the perfusion index reflects the perfusion of tissue blood flow and can be used to monitor microcirculation.

[0029] The second acquisition unit is used to collect the differential pressure of the user's nasal ala. Specifically, a nose pad 3 is provided on the eyeglass frame 1, and the second acquisition unit is a miniature differential pressure sensor (model SDP800) placed on the nose pad 3 in a symmetrical arrangement on both sides. It is used to detect changes in the nasal ala respiratory pressure difference and derive respiratory rate and depth (i.e., respiratory data). Specifically, the opening and closing of the nasal ala (opening and closing frequency) is consistent with the change in respiratory rate. By establishing a mapping relationship between the opening and closing of the nasal ala and the respiratory rate, the opening and closing frequency can be derived from the frequency of respiratory pressure difference changes. Therefore, the respiratory rate can be obtained by collecting changes in the nasal ala respiratory pressure difference. As for respiratory depth, when the user inhales normally, the nasal ala expands outward, and the pressure collected by the second acquisition unit gradually increases to its maximum value. When the user exhales normally, the nasal ala contracts inward, and the pressure collected by the second acquisition unit gradually decreases to its minimum value. The user's respiratory depth can be inferred from the time interval between the maximum and minimum pressure values. Respiratory rate refers to the number of breaths per unit time. In normal adults at rest, it is usually 12-20 breaths per minute. Respiratory depth reflects the amount of air inhaled or exhaled with each breath. It is closely related to lung ventilation and can be used to assess lung function and the health of the respiratory system.

[0030] Preferably, the second acquisition unit also integrates a vibration sensor, which is used to collect vibration information caused by the expansion and contraction of the nasal cavity during ventilation. When air enters and exits the nostrils, it causes periodic vibrations of the nasal alae and surrounding tissues. The vibration sensor detects these vibrations based on the piezoelectric effect (or other highly sensitive sensing principles). During inhalation, the nasal alae expand outward, and the vibration amplitude detected by the sensor gradually increases; during exhalation, the nasal alae contract inward, and the vibration amplitude decreases accordingly. By continuously monitoring parameters such as the amplitude changes and frequency characteristics of the vibration signal, detailed vibration information during nasal ventilation can be obtained. This information is closely related to respiratory rate, respiratory depth, and the smoothness of nasal ventilation. By fusing and analyzing the vibration information from the vibration sensor with the respiratory pressure difference data originally collected by the second acquisition unit: on the one hand, by comparing the frequency of the vibration signal with the frequency of the respiratory pressure difference change, the respiratory frequency can be calculated more accurately, improving the reliability of the measurement; on the other hand, by combining the vibration amplitude and the respiratory pressure difference value, the respiratory depth can be assessed more accurately. For example, in cases of partial nasal obstruction, even if the pressure difference change may not be obvious, the abnormal weakening or irregular change of the vibration signal can still provide an important reference for judging the respiratory depth.

[0031] The third acquisition unit 101 is used to acquire eye movement data of the user's eyes, including reflected light information and rotation information. Specifically, the third acquisition unit 101 includes an infrared light source and a CMOS image sensor to capture the eye's reflectance spectrum (i.e., reflected light information) and movement trajectory (i.e., eye rotation information), and analyze eye movement patterns and pupil diameter changes. The infrared light source uses infrared light of a specific wavelength, which can safely illuminate the eye without causing harm. The CMOS image sensor has the characteristics of high sensitivity and fast response, and can capture the reflectance spectrum of the eye under infrared light illumination and the eye movement trajectory. By analyzing the reflected light spectrum, information about the surface tissues of the eyeball and the tear film can be obtained, such as the stability of the tear film and the blood perfusion of the ocular surface; the analysis of the eye movement trajectory can reveal characteristics such as eye movement patterns and pupil diameter changes, which are closely related to visual function and the state of the nervous system.

[0032] The second wearing mechanism is worn on the user's wrist. In this embodiment, the second wearing mechanism is a smartwatch 4, which is equipped with basic components such as a heart rate sensor, accelerometer, gyroscope, and blood oxygen saturation sensor. A fourth acquisition unit 5 is provided on the second wearing mechanism to collect the user's pulse data. Specifically, the fourth acquisition unit 5 includes a piezoelectric sensor array for collecting the pulse waveform of the radial artery in the wrist. When the pulse wave of the radial artery passes through, it causes a slight deformation on the sensor surface, thereby generating a corresponding electrical signal. By analyzing these electrical signals, pulse rate, pulse pressure, and rhythm characteristics can be extracted. Specifically, the piezoelectric sensor array is distributed at the cun, guan, and chi positions of the wrist, with a piezoelectric matrix (2mm spacing) in the cun, guan, and chi areas. Through the pulse at these three positions, the nature of the condition can be comprehensively judged, such as its superficiality, slowness, rate, strength, etc., thereby inferring the basic pathological characteristics of the disease, such as yin and yang, exterior and interior, cold and heat, deficiency and excess. For example, a superficial and weak pulse may indicate an exterior syndrome or yang deficiency, while a deep and strong pulse may indicate an interior syndrome or yin-cold.

[0033] It also includes a main control center and an output module; the main control center is used for spatiotemporal alignment of multi-source data, including physiological data, differential pressure, eye movement data, and pulse data, and performs multimodal feature fusion of eye movement (temple), respiratory nasal ala, spectral ocular data, and pulse; based on the fusion results, it matches a preset pathological model library and outputs the user's health assessment results. The collaborative acquisition of data from multiple nodes (eye, nose, wrist) overcomes the bottleneck of simultaneous monitoring of multiple human systems.

[0034] The output module transmits the assessment results to a terminal device for display or to cloud storage. It consists of a wireless communication module (such as Bluetooth or Wi-Fi) and a data storage module. The wireless communication module transmits the health assessment results to a terminal device (such as a mobile phone or tablet), allowing users to view their health status in real time through a dedicated application, including values ​​of various physiological parameters, health assessment reports, and warning information. The data storage module stores the collected raw data and health assessment results in the cloud, facilitating users' access to historical data and long-term data analysis and management. Furthermore, cloud storage supports data backup and recovery functions, ensuring data security and reliability.

[0035] Specifically, the steps for constructing a pathological model library include: S1. Multi-source Medical Data Acquisition and Annotation: Eye-tracking, respiratory, spectral, and pulse data were acquired in a clinical setting and annotated with pathological labels by medical experts. Specifically, time-series physiological data from at least 10,000 patients were collected. Annotation dimensions included disease type labels (hypertension, diabetes, arrhythmia), severity grading (Levels I-IV), and temporal-spatial correlation information. Wavelet transform was used to remove motion artifacts and environmental noise interference. The annotation process was conducted by medical experts with extensive clinical experience, using standardized annotation criteria and specifications.

[0036] S2. Cross-modal Feature Extraction and Screening: Temporal, frequency, and spatial features are extracted from the data, and feature importance analysis is used to screen a subset of features strongly correlated with the target disease. For eye-tracking data, temporal features of pupil diameter changes (such as average diameter, rate of diameter change, etc.), frequency features (such as energy distribution of different frequency components, etc.), and spatial features (such as the shape and direction of eye movement trajectories, etc.) can be extracted. For respiratory data, the stability of respiratory frequency, the distribution features of respiratory depth, and the morphological features of respiratory waveforms can be extracted. Feature importance analysis employs feature selection methods from machine learning algorithms (such as random forests, support vector machines, etc.), combined with prior knowledge from medical experts, to screen a subset of features strongly correlated with the target disease, thereby improving the accuracy and efficiency of the model.

[0037] S3. Multi-Algorithm Fusion Model Training: Based on the selected feature subset, an ensemble model incorporating supervised learning and deep learning is trained to generate a pathology classifier. Supervised learning algorithms (such as logistic regression and decision trees) can utilize label information from labeled data to effectively classify and regress features; deep learning algorithms (such as convolutional neural networks and recurrent neural networks) can automatically learn complex features and inherent patterns in the data. By fusing these two types of algorithms to construct an ensemble model, their complementary advantages can be fully utilized to improve the performance and generalization ability of the pathology classifier.

[0038] S4. Clinical Validation and Threshold Calibration: Model performance is validated through retrospective data and prospective trials, and the warning threshold is dynamically adjusted based on the gold standard diagnostic results. When validating model performance through retrospective data and prospective trials, rigorous statistical methods and evaluation metrics (such as accuracy, recall, F1 score, ROC curve, etc.) are employed to quantitatively evaluate the model's classification results. The process of dynamically adjusting the warning threshold based on the gold standard diagnostic results is based on a large amount of clinical data and diagnostic cases. By analyzing the distribution of warning indicators for different disease types, a reasonable threshold range is determined, enabling the model to maintain high warning accuracy while minimizing false positive and false negative rates.

[0039] S5. Incremental Learning and Compliant Deployment: The model is continuously updated with new case data and deployed to the central control unit within a privacy-preserving framework. Incremental learning algorithms are employed during this process, allowing the model to learn from new data without retraining the entire dataset, thus maintaining its timeliness and accuracy. Simultaneously, the deployment process strictly adheres to privacy regulations and standards, employing encrypted transmission and data anonymization techniques to ensure that users' personal health data is not leaked or misused.

[0040] The central control unit also employs waveform decomposition algorithms to decompose pulse data into amplitude-time domain feature vectors of the main wave, tidal wave, and dicrotic wave; and identifies pathological patterns of slippery, wiry, and choppy pulses based on convolutional neural networks. The main wave typically reflects the primary pulse wave generated during cardiac contraction; the tidal wave is formed by the elastic reflection of the aortic arch; and the dicrotic wave is caused by the closure of the atrioventricular valves and atrial contraction during late ventricular diastole. By analyzing the amplitude and time domain characteristics of these waveform components, more information about cardiovascular function can be obtained. The identification of pathological patterns such as slippery, wiry, and choppy pulses based on convolutional neural networks involves training a convolutional neural network model to automatically learn the characteristic patterns of pulse waveforms and associate them with corresponding pathological labels, thereby achieving automatic identification of different pathological pulse patterns.

[0041] The central control unit performs health assessments based on the following logic: The spatiotemporal alignment is achieved through timestamp synchronization and spatial coordinate correction to ensure that multi-source data are consistent in time and space. For example, for eye movement data and respiratory data collected simultaneously, timestamp alignment ensures that the data analyzed is from the same time period. At the same time, spatial coordinate correction is performed based on the spatial relationship between the collection location of eye movement data (eyeball) and the collection location of respiratory data (nose ala) to ensure that the data can accurately reflect the physiological state of the human body.

[0042] Multimodal fusion of eye-tracking, respiratory, spectral, and pulse data was employed using feature extraction algorithms to generate multidimensional health indicators. In terms of feature extraction algorithms, various methods (such as wavelet transform and principal component analysis) were used to fuse the multimodal data, extracting multidimensional health indicators that characterize the user's health status, such as heart rate variability, respiratory-heart rate coupling, and eye-tracking-pulse correlation. These indicators comprehensively reflect the interaction and coordination between multiple physiological systems, providing a more comprehensive basis for health assessment.

[0043] The fusion results are matched with a preset pathological model library, which is built based on machine learning algorithms and covers the feature thresholds of hypertension, diabetes, heart disease and nervous system diseases. The main control center compares the extracted multidimensional health indicators with the feature thresholds in the model library one by one to determine whether the user's health status is normal.

[0044] When assessing anomalies, the central control unit compares the current fusion results with the feature thresholds. If the deviation of the indicators exceeds the preset threshold, it determines that there is a health risk and issues corresponding warning prompts according to the degree of risk.

[0045] Preferably, the main control center is also used to perform the following operations when a health assessment fails: To determine if the failure was caused by data interference: If abnormal light spots appear in the spectral data, it is determined to be ambient light interference, and adjustment instructions are generated to optimize the sensitivity of each sensor. For example, once abnormal light spots are detected in the spectral data, such as irregular spot shapes, abnormal intensity, or spectral wavelengths deviating from the normal range, it is immediately determined to be ambient light interference. At this time, the main control center will generate adjustment instructions based on the characteristic parameters of the spectral data (such as background light intensity, interfering light wavelength, etc.). The instructions include reducing the light intensity of the optical heart rate sensor to reduce ambient light reflection interference; adjusting the sensor's sampling frequency to avoid high-frequency flicker components of ambient light, or invalidating the data collected during that period; and optimizing the sensor's filter parameters to filter out interfering light wavelengths. At the same time, the main control center records information such as the time, type, and intensity of the interference for subsequent data analysis and model optimization.

[0046] To determine if the failure was caused by user activity: If the respiratory and pulse data show non-physiological fluctuations, it is identified as motion artifacts, and the data is corrected through algorithmic filtering. For example, when non-physiological fluctuations in respiratory rate and depth, or pulse waveforms are detected, such as a sudden increase in respiratory rate beyond the normal range (12-20 breaths / minute), or high-frequency jitter or intermittent interruptions in the pulse waveform, data from built-in motion sensors such as accelerometers and gyroscopes is used to determine if the user is in motion. If a motion artifact is identified, the main control center will select an appropriate filtering algorithm to correct the data based on the type (e.g., walking, running, jumping) and intensity of the motion. For example, during low-intensity exercise (e.g., walking), a low-pass filter is used to remove low-frequency motion interference; during high-intensity exercise (e.g., running), an adaptive filtering algorithm is used to dynamically adjust filter parameters and remove motion artifacts in real time.

[0047] While performing data interference and user activity assessments, the central control unit also initiates a data fusion mechanism. This involves fusing data from the first acquisition unit (optical heart rate sensor) on pulse wave amplitude, blood oxygen saturation, and perfusion index with data from the fourth acquisition unit (piezoelectric sensor array) on pulse rate and pulse pressure. If the fused data still shows anomalies, it further combines eye-tracking data from the third acquisition unit (infrared light source and CMOS image sensor) to observe whether eye movements exhibit abnormal trajectory changes due to user activity, and whether lip color monitoring data shows facial congestion due to movement. This comprehensive assessment determines the specific reasons for the failed health assessment, allowing for more precise data correction and reassessment.

[0048] Preferably, the main control center is also used to predict user health risks based on historical data trend analysis; The system compares current data with historical benchmarks, which are generated based on continuous user monitoring data (≥30 days) and cover dynamic physiological indicators such as eye movement trajectory stability, pulse waveform amplitude, and respiratory pressure differential cycle. If eye movement data is abnormal or pulse trend deviates from a linear relationship, an early warning will be issued.

[0049] Preferably, the first wearing mechanism is also provided with an image acquisition unit, which is used to acquire the user's lip color, and the main control center is also used to assist in judging the user's health status based on the user's lip color.

[0050] Specifically, the image acquisition unit includes a high-resolution color camera that can capture high-definition images of the user's lips. The operating frequency of the image acquisition unit can be set according to actual needs, such as capturing one image per minute or capturing images when a specific event is triggered (such as when the user manually starts the system or when the system automatically detects an abnormal situation).

[0051] When analyzing the acquired lip color images, the central control unit first preprocesses the images, including image enhancement, noise reduction, and edge detection, to improve image quality and highlight the features of the lip area. Then, a color analysis algorithm is used to quantify the lip color, extracting color feature parameters such as RGB values, hue, and saturation. Normally, lips have a rosy hue. However, lip color may change when certain diseases or health problems occur. For example, pale lips may be related to anemia, cyanosis may indicate hypoxia, and red lips may be a sign of fever or cardiovascular disease. The central control unit compares the extracted lip color feature parameters with preset normal ranges and combines them with data from other acquisition units (such as pulse wave amplitude, blood oxygen saturation, and respiratory rate) for comprehensive analysis to help determine the user's health status and provide more reference information for health assessment.

[0052] Preferably, it also includes an alarm module for generating an audio-visual alarm signal when the health assessment result exceeds a preset safety range.

[0053] The alarm module includes an audible alarm unit and a visual alarm unit. The audible alarm unit uses a high-fidelity speaker to emit alarm sounds of different frequencies and volumes to attract the user's attention. The visual alarm unit includes LED indicators and a display screen. The LED indicators use different colors and flashing frequencies to display different alarm levels, such as green for normal, yellow for minor abnormalities, and red for serious abnormalities. The display screen shows the alarm information in real time, such as "Abnormal heart rate, please rest immediately" and "Rapid breathing rate, possible risk of respiratory disease." When the health assessment results exceed the preset safety range, the main control center triggers the corresponding alarm unit based on the severity and urgency of the assessment results, generating audible and visual alarm signals. For example, for minor health abnormalities, only the visual alarm unit may be triggered, displaying a yellow warning message; while for serious health risks, such as sudden arrhythmia or respiratory arrest, both the audible and visual alarm units are triggered simultaneously, emitting a loud alarm sound and flashing red lights to ensure that the user can promptly detect the problem and take appropriate measures.

[0054] The classification of alarm levels is based on factors including medical disease severity grading standards, the degree of abnormality in physiological parameters, and the potential harm to human health. For example, for blood oxygen saturation (SpO2), a value between 95% and 100% is within the normal range and does not require an alarm; a value between 90% and 95% is considered slightly abnormal and triggers a low-level alarm; while a value below 90% is considered severely abnormal and triggers a high-level alarm.

[0055] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific structures and / or characteristics in the solutions are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the structure of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A digital medical health parameter monitoring device, comprising: First wearing mechanism and second wearing mechanism; The first wearing mechanism is worn on the user's eyes. The first wearing mechanism is provided with a first acquisition unit, a second acquisition unit, and a third acquisition unit (101). The first acquisition unit is used to collect physiological data of the user's temple area, the second acquisition unit is used to collect the differential pressure of the user's nasal ala, and the third acquisition unit (101) is used to collect eye movement data of the user's eyeballs. The eye movement data includes reflected light information and rotation information. The second wearing mechanism is worn on the user's wrist. The second wearing mechanism is provided with a fourth acquisition unit (5). The fourth acquisition unit (5) is used to collect the user's pulse data. It also includes a main control center and an output module; the main control center is used to perform spatiotemporal alignment of multi-source data, which includes physiological data, differential pressure, eye movement data and pulse data, and to perform multimodal feature fusion of temple-nasal ala-eye-pulse; based on the fusion result, it matches a preset pathological model library and outputs the user's health assessment result; the output module is used to transmit the health assessment result to the terminal device for display or cloud storage.

2. The digital medical health parameter monitoring device according to claim 1, characterized in that: The first acquisition unit includes an optical heart rate sensor for acquiring pulse wave and blood oxygen saturation data in the temple area; The second acquisition unit includes a miniature differential pressure sensor for detecting changes in nasal respiratory pressure differential and deriving respiratory rate and depth; The third acquisition unit (101) includes an infrared light source and a CMOS image sensor. The CMOS image sensor is used to capture the eye's reflectance spectrum and motion trajectory. The fourth acquisition unit (5) includes a piezoelectric sensor array for acquiring the pulse waveform of the radial artery of the wrist, which includes pulse rate, pulse pressure and rhythm characteristics.

3. The digital medical health parameter monitoring device according to claim 2, characterized in that: The steps for constructing a pathological model library include: S1. Multi-source medical data acquisition and annotation: Acquire eye movement data, respiratory data, spectral data and pulse data in a clinical setting, and have medical experts annotate them with pathological labels; S2. Cross-modal feature extraction and screening: Extract time-domain, frequency-domain, and spatial features from the data, and screen a subset of features strongly correlated with the target disease through feature importance analysis; S3. Multi-algorithm fusion model training: Based on the selected feature subset, train an integrated model that includes supervised learning and deep learning to generate a pathology classifier; S4. Clinical Validation and Threshold Calibration: Validate model performance through retrospective data and prospective trials, and dynamically adjust warning thresholds based on gold standard diagnostic results; S5. Incremental Learning and Compliance Deployment: Continuously access new case data to update the model and deploy it to the central control unit under a privacy protection framework.

4. The digital medical health parameter monitoring device according to claim 3, characterized in that: The main control center is also used to use waveform decomposition algorithms to decompose pulse data into amplitude-time domain feature vectors of main wave, tidal wave, and diphtheria wave; and to identify pathological patterns of slippery pulse, wiry pulse, and hesitant pulse based on convolutional neural networks.

5. The digital medical health parameter monitoring device according to claim 2, characterized in that: The central control unit performs health assessments based on the following logic: The spatiotemporal alignment is achieved through timestamp synchronization and spatial coordinate correction, ensuring that multi-source data are consistent in time and space. Feature extraction algorithms are used to perform multimodal fusion of eye-tracking data, respiratory data, spectral data, and pulse data to generate multidimensional health indicators; The fusion results are matched with a pre-set pathological model library, which is built based on machine learning algorithms and covers the feature thresholds of hypertension, diabetes, heart disease and nervous system diseases. When assessing anomalies, the current fusion result is compared with the feature threshold. If the indicator deviation exceeds the preset threshold, a health risk is determined.

6. The digital medical health parameter monitoring device according to claim 5, characterized in that: The central control unit is also used to perform the following operations when a health assessment fails: Determine if the failure is caused by data interference: If abnormal light spots appear in the spectral data, it is determined to be ambient light interference, and adjustment instructions are generated to optimize the sensitivity of each sensor; To determine if the failure was caused by user activity: if the respiratory data and pulse data show non-physiological fluctuations, it is determined to be motion artifacts, and the data is corrected by filtering through an algorithm.

7. The digital medical health parameter monitoring device according to claim 6, characterized in that: The central control unit is also used to predict user health risks based on historical data trend analysis; If the current data is compared with historical benchmarks, and the eye movement data is abnormal or the pulse data trend deviates from the linear relationship, an early warning will be issued.

8. The digital medical health parameter monitoring device according to claim 7, characterized in that: The first wearing mechanism is also equipped with an image acquisition unit, which is used to collect the user's lip color. The main control center is also used to assist in judging the user's health based on the user's lip color.

9. The digital medical health parameter monitoring device according to claim 8, characterized in that: The first and second wearing mechanisms are made of flexible biocompatible materials.

10. The digital medical health parameter monitoring device according to claim 9, characterized in that: It also includes an alarm module, which generates audiovisual alarm signals when the health assessment results exceed the preset safety range.

Images