Physiological index detection method and device, equipment, storage medium and program product
By collecting user pressure data through a pressure-sensing mattress and using polynomial expansion and Lasso model to detect physiological indices, the problem of inconvenience and error in traditional methods is solved, and continuous and accurate detection of special populations is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-04-21
AI Technical Summary
Traditional physiological index detection methods are inconvenient and difficult to conduct continuous tests in special populations such as bedridden patients and elderly people with limited mobility, leading to detection errors.
Pressure sensor mattresses are used to collect pressure data from users on each airbag. Physiological index detection is performed using polynomial expansion and Lasso model. Continuous detection is achieved by matching the airbag positions with key user body parts.
Even if the user does not cooperate with the test, data collection can be easily completed while resting or sleeping, accurately detecting physiological indices and improving the continuity and accuracy of the test.
Smart Images
Figure CN121890942A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, device, storage medium, and program product for detecting physiological indices. Background Technology
[0002] Currently, physiological index testing, such as Body Mass Index (BMI), mainly relies on traditional weighing scales and height measurement tools, requiring users to actively cooperate in completing the measurement. In some medical monitoring and rehabilitation care scenarios, for certain special groups (such as bedridden patients and elderly people with limited mobility), traditional testing methods are inconvenient to use and difficult to conduct continuous testing, and there is also the possibility of non-cooperation, which can lead to errors in the physiological index measurement. Summary of the Invention
[0003] Therefore, it is necessary to provide a method, device, equipment, storage medium, and program product for detecting physiological indices to address the aforementioned technical problems. This method or program can conveniently and continuously detect physiological indices and effectively improve the accuracy of physiological index detection.
[0004] In a first aspect, this application provides a method for detecting a physiological index, the method comprising:
[0005] The system collects pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to a different key part of the user's body.
[0006] The pressure change value of each airbag is determined based on the pressure data and the baseline reference value;
[0007] The pressure change values of each airbag are expanded using a polynomial to obtain the extended feature set corresponding to each airbag;
[0008] The physiological index value of the user is obtained by detecting physiological indexes based on the extended feature set corresponding to each airbag using the Lasso model.
[0009] In one embodiment, before determining the pressure change value of each airbag based on the pressure data and a baseline reference value, the method further includes:
[0010] A multi-channel pressure signal is generated based on the pressure data of each airbag;
[0011] The multi-channel pressure signal is divided into data frames to obtain multiple pressure data frames for each channel;
[0012] For multiple pressure data frames of each channel, the difference between adjacent pressure data frames in the same channel is calculated. After obtaining the difference between adjacent pressure data frames of each channel, the difference between adjacent pressure data frames belonging to the same time in each channel is weighted and summed to obtain the change amplitude used to characterize the intensity of the user activity.
[0013] Determining the pressure change value of each airbag based on the pressure data and the baseline reference value includes:
[0014] If the change amplitude is less than the dynamic threshold, the pressure change value of each airbag is determined based on multiple pressure data frames of each channel and the baseline reference value.
[0015] In one embodiment, before calculating the difference between adjacent pressure data frames in the same channel, the method further includes:
[0016] Multiple pressure data frames of each channel are filtered using a high-pass filter to obtain multiple filtered pressure data frames of each channel.
[0017] In one embodiment, the pressure data includes pressure data over at least two time periods; the step of obtaining the user's physiological index value by detecting physiological indexes based on the extended feature set corresponding to each airbag using the Lasso model includes:
[0018] When obtaining the extended feature set corresponding to each airbag for a time period, the obtained extended feature set corresponding to each airbag is input into the Lasso model so that the Lasso model can perform physiological index detection based on the extended feature set corresponding to each airbag and obtain the corresponding physiological index value.
[0019] When physiological index values are obtained for at least two time periods, the average value of the physiological index values for the at least two time periods is determined; the average value is the user's final physiological index value.
[0020] In one embodiment, the method further includes:
[0021] Pressure data generated by test users on each airbag of the sample mattress is collected to obtain pressure data samples; the sample mattress is the pressure-sensing mattress or other pressure-sensing mattress used in the model training phase.
[0022] Based on the pressure data sample and the baseline reference value, determine the pressure change value of each air bladder in the sample mattress;
[0023] The pressure change values of each air bladder in the sample mattress are expanded using a polynomial to obtain the extended feature set corresponding to each air bladder in the sample mattress; wherein, the extended feature set corresponding to each air bladder in the sample mattress includes a training set.
[0024] The initial Lasso model is trained based on the training set until the model converges, resulting in the trained Lasso model.
[0025] In one embodiment, training the initial Lasso model based on the training set includes:
[0026] The initial Lasso model is trained based on the training set, and the loss value is calculated during the training process. The loss value includes the physiological index prediction loss value and the regularization term loss value.
[0027] The parameters of the initial Lasso model are adjusted based on the loss value.
[0028] Secondly, this application also provides a device for detecting physiological indices, the device comprising:
[0029] The data acquisition module is used to collect pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to different key parts of the user.
[0030] The determination module is used to determine the pressure change value of each airbag based on the pressure data and the baseline reference value;
[0031] The processing module is used to expand the pressure change values of each airbag through a polynomial to obtain the extended feature set corresponding to each airbag;
[0032] The detection module is used to detect physiological indices based on the extended feature set corresponding to each airbag using the Lasso model, and to obtain the physiological index value of the user.
[0033] In one embodiment, the device further includes:
[0034] A generation module is used to generate a multi-channel pressure signal based on the pressure data of each airbag;
[0035] The segmentation module is used to divide the multi-channel pressure signal into data frames to obtain multiple pressure data frames for each channel.
[0036] The calculation module is used to calculate the difference between adjacent pressure data frames in the same channel for multiple pressure data frames in each channel, and after obtaining the difference between adjacent pressure data frames in each channel, to perform a weighted summation of the difference between adjacent pressure data frames in each channel that belong to the same time to obtain the change amplitude used to characterize the intensity of the user activity.
[0037] The determining module is further configured to determine the pressure change value of each airbag based on multiple pressure data frames of each channel and a baseline reference value if the change amplitude is less than a dynamic threshold.
[0038] In one embodiment, the device further includes:
[0039] The filtering module is used to filter multiple pressure data frames of each channel using a high-pass filter to obtain multiple filtered pressure data frames of each channel.
[0040] In one embodiment, the pressure data includes pressure data over at least two time periods;
[0041] The detection module is further configured to input the obtained extended feature sets corresponding to each airbag into the Lasso model each time an extended feature set corresponding to each airbag is obtained, so that the Lasso model performs physiological index detection based on the extended feature sets corresponding to each airbag and obtains the corresponding physiological index value; when physiological index values for at least two time periods are obtained, the average value of the physiological index values for the at least two time periods is determined; the average value is the user's final physiological index value.
[0042] In one embodiment, the device further includes:
[0043] The training module is used to collect pressure data generated by test users on each airbag of the sample mattress to obtain pressure data samples; the sample mattress is the pressure-sensing mattress or other pressure-sensing mattress used in the model training phase; based on the pressure data samples and the baseline reference value, the pressure change value of each airbag of the sample mattress is determined; the pressure change value of each airbag of the sample mattress is expanded by a polynomial to obtain the extended feature set corresponding to each airbag of the sample mattress; wherein, the extended feature set corresponding to each airbag of the sample mattress includes the training set; the initial Lasso model is trained based on the training set until the model converges to obtain the trained Lasso model.
[0044] In one embodiment, the training module is further configured to train an initial Lasso model based on the training set, and calculate a loss value during the training process, the loss value including a physiological index prediction loss value and a regularization term loss value; and adjust the parameters of the initial Lasso model based on the loss value.
[0045] Thirdly, this application also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the physiological index detection method.
[0046] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the physiological index detection method.
[0047] Fifthly, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the method for detecting the physiological index.
[0048] The aforementioned physiological index detection method, device, computer equipment, storage medium, and computer program product collect pressure data generated by the user on each airbag of the pressure-sensing mattress. The position of each airbag corresponds to different key parts of the user, so data collection can be easily completed even if the user does not cooperate with the test, while the user is resting or sleeping. This also helps to continuously detect physiological indices. The pressure change value of each airbag is determined based on the pressure data and baseline reference value. The pressure change value of each airbag is expanded using a polynomial to obtain the extended feature set corresponding to each airbag. This expansion method can fully capture the nonlinear relationship and interaction effect between pressure features and physiological indices. Physiological index detection is performed based on the extended feature set corresponding to each airbag using the Lasso model to obtain the user's physiological index value. Therefore, even if the user does not cooperate with the test, the user's physiological index can be accurately detected. Attached Figure Description
[0049] Figure 1 This is a diagram illustrating the application environment of a physiological index detection method in one embodiment.
[0050] Figure 2 This is a flowchart illustrating a method for detecting physiological indices in one embodiment;
[0051] Figure 3 This is a schematic diagram of a pressure-sensing mattress and the distribution of each airbag in one embodiment;
[0052] Figure 4This is a structural block diagram of a physiological index detection device in one embodiment;
[0053] Figure 5 This is a structural block diagram of a physiological index detection device in another embodiment;
[0054] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0055] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0056] The physiological index detection method provided in this application embodiment can be applied to, for example... Figure 1 The application environment shown is illustrated. Terminal 102, server 104, and pressure-sensing mattress 106 can communicate via a network. A data storage system can be used to store the pressure data collected by the pressure-sensing mattress 106. The data storage system can be configured independently, integrated into server 104, or located in the cloud or on other devices.
[0057] The pressure-sensing mattress 106 can be equipped with sensors and data processing devices for physiological index detection, so the pressure-sensing mattress 106 can independently detect physiological indices; in addition, the server 104 can also detect physiological indices.
[0058] If performed by the pressure-sensing mattress 106, the pressure-sensing mattress 106 collects pressure data generated by the user on each airbag of the pressure-sensing mattress 106, with the position of each airbag corresponding to different key parts of the user; the pressure change value of each airbag is determined based on the pressure data and baseline reference value; the pressure change value of each airbag is expanded by a polynomial to obtain the extended feature set corresponding to each airbag; physiological index detection is performed based on the extended feature set corresponding to each airbag using the Lasso model to obtain the user's physiological index value; and then the physiological index value is transmitted to the terminal 102 of the user, medical staff, or family member.
[0059] If executed by server 104, server 104 acquires the pressure data generated by the user on each airbag of the pressure-sensing mattress 106, collected by the pressure-sensing mattress 106. The position of each airbag corresponds to different key parts of the user. Based on the pressure data and baseline reference values, server 104 determines the pressure change value of each airbag. The pressure change value of each airbag is expanded using a polynomial to obtain the extended feature set corresponding to each airbag. Using the Lasso model, physiological index detection is performed based on the extended feature set corresponding to each airbag to obtain the user's physiological index value. Then, the physiological index value is transmitted to the terminal 102 of the user, medical staff, or family member.
[0060] It should be noted that the physiological index detection method of this application can also be executed by terminal 102.
[0061] The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses.
[0062] Server 104 can be a standalone physical server or a service node in a blockchain system. These service nodes form a peer-to-peer (P2P) network, where the P2P protocol is an application layer protocol running on top of the Transmission Control Protocol (TCP). Furthermore, server 104 can also be a server cluster composed of multiple physical servers, and can be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.
[0063] In one embodiment, such as Figure 2 As shown, a method for detecting physiological indices is provided. This method can be derived from... Figure 1 The pressure-sensing mattress, server, or terminal in the process executes the commands, or the pressure-sensing mattress and server work together to execute the commands, or the pressure-sensing mattress and terminal work together to execute the commands, in accordance with this method. Figure 1 Taking the server execution in [the context of the example] as an example, the steps include:
[0064] S202 collects pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to different key parts of the user.
[0065] Among them, pressure-sensing mattresses can be multi-airbag pressure-sensing mattresses, such as an eight-airbag pressure-sensing mattress. Each airbag contains a high-precision pressure sensor, which can independently collect pressure data from key areas. Key areas could include the user's left shoulder, right shoulder, back, waist, left hip, right hip, thigh, and calf. Figure 3 As shown, pressure data can be collected from the left shoulder, right shoulder, back, waist, left hip, right hip, thigh, and calf.
[0066] Pressure data can be pressure data over a fixed period of time. The location of each airbag corresponds to different key parts of the user, so the pressure data of each airbag can be physiological data of the user's key parts.
[0067] S204, determine the pressure change value of each airbag based on the pressure data and baseline reference value.
[0068] The baseline reference value can be the average of pressure data collected before the user lies on the pressure-sensing mattress. For example, before the user gets into bed, pressure data from each channel is automatically collected within 1 second, and then the average of the pressure data from each channel is calculated and used as the baseline reference value.
[0069] In one embodiment, the server can calculate the difference between the pressure data of each airbag and the baseline reference value, and these differences are the pressure change values of each airbag.
[0070] In one embodiment, the server can generate a multi-channel pressure signal based on the pressure data of each airbag; divide the multi-channel pressure signal into data frames to obtain multiple pressure data frames for each channel; calculate the difference between adjacent pressure data frames in the same channel for the multiple pressure data frames in each channel, and after obtaining the difference between adjacent pressure data frames in each channel, perform a weighted summation of the differences between adjacent pressure data frames in each channel that belong to the same time period to obtain the change amplitude used to characterize the intensity of user activity; if the change amplitude is less than the dynamic threshold, determine the pressure change value of each airbag based on the multiple pressure data frames of each channel and the baseline reference value.
[0071] Each channel employs the aforementioned difference calculation method to calculate the difference between adjacent pressure data frames within the same channel. After obtaining the differences between adjacent pressure data frames in each channel, the differences between adjacent pressure data frames belonging to the same time period in each channel are weighted and summed to obtain multiple amplitude values representing the intensity of user activity. The number of these amplitude values is equal to the number of pressure data frames in the same channel minus 1.
[0072] In one embodiment, before calculating the pressure change value, the server can also use a high-pass filter to filter multiple pressure data frames of each channel to obtain multiple filtered pressure data frames of each channel.
[0073] Through filtering, low-frequency interference caused by the slow deformation and temperature drift of the pressure-sensing mattress can be effectively filtered out, while retaining mid-to-high frequency physiological signals generated by breathing, heartbeat and slight body movements.
[0074] In one embodiment, if the change amplitude is less than the dynamic threshold, the server can determine that the user is in a resting state (such as a resting lying position). At this time, the pressure change value of each airbag can be determined based on multiple pressure data frames from each channel and the baseline reference value. Therefore, by calculating the pressure change value of each airbag relative to the baseline reference value in real time based on the identified pressure data in the resting state, the influence of inherent device characteristics and environmental factors can be effectively eliminated.
[0075] For example, pressure data is collected by built-in pressure sensors, forming an 8-channel 50Hz pressure signal. This signal is divided into data frames at 0.5-second intervals, with each frame containing 25 sampling points. Then, a fourth-order Butterworth high-pass filter with a cutoff frequency of 1.0 Hz is used for filtering. This effectively removes low-frequency interference caused by slow mattress deformation and temperature drift, while retaining mid-to-high-frequency physiological signals generated by breathing, heartbeat, and subtle body movements. Next, the weighted absolute difference in pressure data between adjacent frames is calculated to obtain the amplitude of change in user activity intensity. Finally, a dynamic threshold judgment mechanism based on background noise level is employed. When the amplitude of change in one or more consecutive (e.g., six) data frames falls below the dynamic threshold, it can be determined that the user has entered a resting, lying-down state. This dynamic threshold can be adaptively adjusted based on historical noise data to ensure accurate identification under different environmental conditions.
[0076] S206, the pressure change values of each airbag are expanded using a polynomial to obtain the extended feature set corresponding to each airbag.
[0077] For example, the pressure change values of eight airbags can be expanded using polynomial expansion to obtain an extended feature set. The polynomial expansion consists of three parts: a first-order term, a second-order term, and a cross term. The first-order term retains all original features from the pressure data, the second-order term generates the square of each original feature, and the cross term generates the product of all pairs of features. Starting from eight basic pressure change values, 44 pressure features can be generated after polynomial expansion. This expansion method can effectively capture the nonlinear relationship and interaction effects between pressure features and physiological indices (such as BMI).
[0078] S208 uses the Lasso model to detect physiological indices based on the extended feature set corresponding to each airbag, and obtains the user's physiological index value.
[0079] Among them, the Lasso (Least Absolute Shrinkage and Selection Operator) model is a linear regression model that uses L1 regularization to address the difficulties in variable selection and model overfitting faced by traditional linear regression in high-dimensional data.
[0080] A physiological index value can be a specific value of the Body Mass Index. Body Mass Index (BMI) is a commonly used standard for measuring a person's weight and overall health. Its core logic is to quantify the relationship between height and weight to determine whether a weight is within a healthy range.
[0081] In one embodiment, the stress data may include stress data over at least two time periods. Therefore, when the server obtains the extended feature set corresponding to each airbag for each time period, it inputs the obtained extended feature set corresponding to each airbag into the Lasso model so that the Lasso model can perform physiological index detection based on the extended feature set corresponding to each airbag and obtain the corresponding physiological index value. When obtaining the physiological index values for at least two time periods, the average value of the physiological index values for at least two time periods is determined. The average value is the user's final physiological index value.
[0082] For example, when a user is lying still on an eight-airbag pressure-sensing mattress, if the system detects that the user is in a resting state, pressure data for eight parts of the user's body will be collected every second, for a total of three times. Each time, basic pressure data will be collected, and the pressure change value of the airbags will be calculated. Then, polynomial feature expansion will be performed to obtain the corresponding extended feature set. Finally, the extended feature set will be input into the trained Lasso model to obtain the BMI prediction value. The average of the three predicted BMI values will be used as the user's final BMI value.
[0083] In the above embodiments, pressure data generated by the user on each airbag of the pressure-sensing mattress is collected. The position of each airbag corresponds to different key parts of the user, so that data can be easily collected even if the user does not cooperate with the test, while the user is resting or sleeping. This also helps to continuously detect physiological indices. The pressure change value of each airbag is determined based on the pressure data and the baseline reference value. The pressure change value of each airbag is expanded by a polynomial to obtain the extended feature set corresponding to each airbag. This expansion method can fully capture the nonlinear relationship and interaction effect between pressure features and physiological indices. Physiological index detection is performed based on the extended feature set corresponding to each airbag using the Lasso model to obtain the user's physiological index value. Therefore, even if the user does not cooperate with the test, the user's physiological index can be accurately detected.
[0084] In one embodiment, the server needs to train an initial Lasso model before detecting a user's physiological index. Once the model converges, a Lasso model for detecting physiological indices can be obtained.
[0085] In one embodiment, the server can collect pressure data generated by the test user on each airbag of the sample mattress to obtain pressure data samples; the sample mattress is a pressure-sensing mattress or other pressure-sensing mattress used in the model training phase; based on the pressure data samples and baseline reference values, the pressure change values of each airbag of the sample mattress are determined; the pressure change values of each airbag of the sample mattress are expanded by a polynomial to obtain the extended feature set corresponding to each airbag of the sample mattress; wherein, the extended feature set corresponding to each airbag of the sample mattress includes the training set; the initial Lasso model is trained based on the training set until the model converges to obtain the trained Lasso model.
[0086] For details regarding the acquisition of pressure data, calculation of pressure change values, and polynomial expansion, please refer to the implementation examples in the application stage described above.
[0087] In one embodiment, the server can train an initial Lasso model based on a training set and calculate a loss value during training, including a physiological index prediction loss value and a regularization term loss value; and then adjust the parameters of the initial Lasso model based on the loss value.
[0088] In one embodiment, the extended feature set corresponding to each airbag of the sample mattress also includes a first test set. Therefore, the server can input the first test set into the Lasso model so that the Lasso model can perform physiological index detection based on the first test set to obtain the test physiological index value. If the test physiological index value is consistent with the reference physiological index value of the test user, the Lasso model is deployed. If the test physiological index value is inconsistent with the reference physiological index value of the test user, the Lasso model is further trained.
[0089] The extended feature set corresponding to each airbag of the sample mattress can be divided into multiple subsets, one subset as the first test set, another subset as the second test set (i.e., the validation set), and the remaining subsets as the training set.
[0090] In one embodiment, during model training, the model parameters are solved by minimizing a loss function, which includes prediction error and a regularization term, balancing model complexity and fitting accuracy. During testing, the model performance is evaluated using a validation set under different regularization parameters, and the parameters that perform best in cross-validation are selected as the final model parameters.
[0091] In the above embodiments, pressure data generated by the test user on each airbag of the sample mattress is collected to obtain pressure data samples; based on the pressure data samples and the baseline reference value, the pressure change value of each airbag of the sample mattress is determined; the pressure change value of each airbag of the sample mattress is expanded by a polynomial to obtain the extended feature set corresponding to each airbag of the sample mattress; the initial Lasso model is trained based on the training set in the extended feature set until the model converges, thereby obtaining a Lasso model for physiological index detection. Thus, users can achieve physiological index detection by lying on the pressure-sensing mattress, which can effectively improve the detection accuracy.
[0092] As an example, taking BMI detection as an example, the overall scheme of this application will be described here. The physiological index detection method of this application may include the following main processes, as detailed below:
[0093] 1. Data Collection
[0094] This mattress features a specially designed eight-airbag pressure-sensing mattress. These airbags are precisely distributed across eight key areas of the body, such as the left and right shoulders, back, waist, left and right hips, thighs, and calves. Each airbag contains a high-precision pressure sensor that can independently collect pressure data from each key area, generating an 8-channel 50Hz pressure signal.
[0095] 2. Data Processing
[0096] The 8-channel pressure signal was divided into data frames at 0.5-second intervals, with each data frame containing 25 sampling points. Then, it was filtered by a fourth-order Butterworth high-pass filter with a cutoff frequency of 1.0 Hz. This effectively filtered out low-frequency interference caused by slow mattress deformation and temperature drift, while retaining mid-to-high frequency physiological signals generated by breathing, heartbeat, and minor body movements.
[0097] The amplitude of change in user activity intensity is obtained by calculating the weighted absolute difference of pressure values in each channel between adjacent frames. A dynamic threshold judgment mechanism based on background noise level is employed; when the amplitude of change in six consecutive data frames is below the dynamic threshold, the user can be determined to have entered a resting, lying-down state. This threshold is adaptively adjusted based on historical noise data to ensure accuracy under different environmental conditions.
[0098] In addition, before the user gets into bed, the system automatically collects and saves the average pressure data of each channel within 1 second as a baseline reference value. For the identified resting state data, the system calculates the pressure change of each airbag relative to the baseline reference value in real time, effectively eliminating the influence of inherent equipment characteristics and environmental factors.
[0099] 3. Feature Expansion
[0100] The pressure change values from the eight airbags were expanded using polynomial expansion to generate an extended feature set. This polynomial expansion consisted of three parts: a first-order term, a second-order term, and a cross term. The first-order term retained all original features; the second-order term generated the square of each original feature; and the cross term generated the product of all pairs of features. Specifically, starting from the eight basic pressure data points, the resulting pressure change values were expanded using a second-order polynomial expansion to generate a total of 44 features. This expansion method effectively captures the nonlinear relationship and interaction effects between pressure data and BMI.
[0101] 4. Model Architecture
[0102] A hybrid model architecture (PolyLasso) combining multinomial features and Lasso regression is adopted. This architecture first expands the input features to a high-dimensional feature space through multinomial transformation, and then uses Lasso regression for feature selection and model training. Specifically, Lasso regression adds an L1 regularization term to linear regression. By controlling the regularization strength parameter, it automatically compresses unimportant feature coefficients to zero, achieving feature selection while preventing overfitting.
[0103] 5. Model Training
[0104] The training data is divided into multiple subsets, with one subset used as the validation set and the rest as the training set in turn. The model performance is evaluated under different regularization parameters, and the parameters that perform best in cross-validation are selected as the final model parameters. During model training, the model parameters are solved by minimizing the loss function, which includes both prediction error and regularization term, balancing model complexity and fitting accuracy.
[0105] 6. Feature Importance Analysis
[0106] After training, the polynomial features in the extended feature set are ranked by importance based on the absolute value of the Lasso regression coefficients. Features with larger absolute values contribute more to BMI prediction, and these features often correspond to physically meaningful stress distribution patterns. By analyzing the types and signs of important features, the relationship between different stress distribution features and BMI can be understood, providing a physical interpretation for the method. Typical important features include combinations of features with clear biomechanical significance, such as the squared term of shoulder stress and the cross term of hip-shoulder stress.
[0107] 7. BMI Forecast
[0108] After the Lasso model is trained, when the user is lying on the eight-airbag pressure-sensing mattress, the pressure data generated by the user on each airbag of the pressure-sensing mattress can be collected. When the user is identified as being at rest, pressure data can continue to be collected, and three sets of pressure data at different time periods can be obtained. Based on these pressure data, the pressure change value is calculated, and then multinomial feature expansion is performed. Finally, the data is input into the trained Lasso model to obtain the BMI prediction value. The average of the three predictions is used as the final BMI value.
[0109] The solution described above allows users to measure their BMI simply by lying on an eight-airbag pressure-sensing mattress, requiring no additional steps and addressing the limitations of traditional methods. Furthermore, the multinomial feature effectively captures the nonlinear relationship between pressure and BMI, achieving significantly higher prediction accuracy than linear methods, with a BMI error within ±1.5 kg / m². Lasso regression automatically filters important features, avoiding redundant feature interference and improving model generalization ability; a single prediction time of less than 0.1 seconds meets real-time requirements. Feature importance ranking provides a physical explanation of the relationship between pressure distribution and BMI, aiding in understanding the body pressure distribution characteristics of obese individuals. This solution can be applied to long-term, continuous BMI trend monitoring, enabling timely detection of changes in body composition and providing data support for clinical nutritional interventions and rehabilitation assessments.
[0110] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0111] Based on the same inventive concept, this application also provides a physiological index detection device for implementing the physiological index detection method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations of one or more physiological index detection device embodiments provided below can be found in the limitations of the physiological index detection method described above, and will not be repeated here.
[0112] In one embodiment, such as Figure 4 As shown, a physiological index detection device is provided, comprising: a data acquisition module 402, a determination module 404, a processing module 406, and a detection module 408, wherein:
[0113] The data acquisition module 402 is used to collect pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to different key parts of the user.
[0114] Module 404 is used to determine the pressure change value of each airbag based on the pressure data and baseline reference value;
[0115] Processing module 406 is used to expand the pressure change values of each airbag through polynomial expansion to obtain the extended feature set corresponding to each airbag;
[0116] The detection module 408 is used to detect physiological indices based on the extended feature set corresponding to each airbag using the Lasso model, and obtain the user's physiological index value.
[0117] In one embodiment, such as Figure 5 As shown, the device also includes:
[0118] The generation module 410 is used to generate a multi-channel pressure signal based on the pressure data of each airbag.
[0119] The segmentation module 412 is used to divide the multi-channel pressure signal into data frames to obtain multiple pressure data frames for each channel.
[0120] The calculation module 414 is used to calculate the difference between adjacent pressure data frames in the same channel for multiple pressure data frames in each channel, and after obtaining the difference between adjacent pressure data frames in each channel, to perform a weighted summation of the difference between adjacent pressure data frames in the same time in each channel to obtain the change amplitude used to characterize the intensity of user activity.
[0121] The determination module 404 is also used to determine the pressure change value of each airbag based on multiple pressure data frames of each channel and the baseline reference value if the change amplitude is less than the dynamic threshold.
[0122] In one embodiment, such as Figure 5As shown, the device also includes:
[0123] The filtering module 416 is used to filter multiple pressure data frames of each channel through a high-pass filter to obtain multiple pressure data frames of each channel after filtering.
[0124] In one embodiment, the pressure data includes pressure data over at least two time periods;
[0125] The detection module 408 is also used to input the obtained extended feature set corresponding to each airbag into the Lasso model each time the extended feature set corresponding to each airbag is obtained, so that the Lasso model can perform physiological index detection based on the extended feature set corresponding to each airbag and obtain the corresponding physiological index value; when the physiological index value of at least two time periods is obtained, the average value of the physiological index value of at least two time periods is determined; the average value is the user's final physiological index value.
[0126] In the above embodiments, pressure data generated by the user on each airbag of the pressure-sensing mattress is collected. The position of each airbag corresponds to different key parts of the user, so that data can be easily collected even if the user does not cooperate with the test, while the user is resting or sleeping. This also helps to continuously detect physiological indices. The pressure change value of each airbag is determined based on the pressure data and the baseline reference value. The pressure change value of each airbag is expanded by a polynomial to obtain the extended feature set corresponding to each airbag. This expansion method can fully capture the nonlinear relationship and interaction effect between pressure features and physiological indices. Physiological index detection is performed based on the extended feature set corresponding to each airbag using the Lasso model to obtain the user's physiological index value. Therefore, even if the user does not cooperate with the test, the user's physiological index can be accurately detected.
[0127] In one embodiment, such as Figure 5 As shown, the device also includes:
[0128] Training module 418 is used to collect pressure data generated by test users on each airbag of the sample mattress to obtain pressure data samples; the sample mattress is a pressure-sensing mattress or other pressure-sensing mattress used in the model training phase; based on the pressure data samples and baseline reference values, the pressure change values of each airbag of the sample mattress are determined; the pressure change values of each airbag of the sample mattress are expanded by a polynomial to obtain the extended feature set corresponding to each airbag of the sample mattress; wherein, the extended feature set corresponding to each airbag of the sample mattress includes the training set; the initial Lasso model is trained based on the training set until the model converges to obtain the trained Lasso model.
[0129] In one embodiment, the training module is further configured to train an initial Lasso model based on a training set and calculate a loss value during the training process, including a physiological index prediction loss value and a regularization term loss value; and adjust the parameters of the initial Lasso model based on the loss value.
[0130] In the above embodiments, pressure data generated by the test user on each airbag of the sample mattress is collected to obtain pressure data samples; based on the pressure data samples and the baseline reference value, the pressure change value of each airbag of the sample mattress is determined; the pressure change value of each airbag of the sample mattress is expanded by a polynomial to obtain the extended feature set corresponding to each airbag of the sample mattress; the initial Lasso model is trained based on the training set in the extended feature set until the model converges, thereby obtaining a Lasso model for physiological index detection. Thus, users can achieve physiological index detection by lying on the pressure-sensing mattress, which can effectively improve the detection accuracy.
[0131] Each module in the aforementioned physiological index detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.
[0132] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores stress data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for detecting physiological indices.
[0133] In one embodiment, a computer device is provided, which may be a pressure-sensing mattress, including a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for detecting physiological indices. The display unit of the computer device is used to form a visually visible image. It can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.
[0134] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0135] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for detecting physiological indices.
[0136] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for detecting physiological indices.
[0137] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described method for detecting physiological indices.
[0138] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.
[0139] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, and other data processing logic devices, and are not limited to these.
[0140] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0141] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for detecting a physiological index, characterized in that, The method includes: The system collects pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to a different key part of the user's body. The pressure change value of each airbag is determined based on the pressure data and the baseline reference value; The pressure change values of each airbag are expanded using a polynomial to obtain the extended feature set corresponding to each airbag; The physiological index value of the user is obtained by detecting physiological indexes based on the extended feature set corresponding to each airbag using the Lasso model.
2. The method according to claim 1, characterized in that, Before determining the pressure change value of each airbag based on the pressure data and the baseline reference value, the method further includes: A multi-channel pressure signal is generated based on the pressure data of each airbag; The multi-channel pressure signal is divided into data frames to obtain multiple pressure data frames for each channel; For multiple pressure data frames of each channel, the difference between adjacent pressure data frames in the same channel is calculated. After obtaining the difference between adjacent pressure data frames of each channel, the difference between adjacent pressure data frames belonging to the same time in each channel is weighted and summed to obtain the change amplitude used to characterize the intensity of the user activity. Determining the pressure change value of each airbag based on the pressure data and the baseline reference value includes: If the change amplitude is less than the dynamic threshold, the pressure change value of each airbag is determined based on multiple pressure data frames of each channel and the baseline reference value.
3. The method according to claim 2, characterized in that, Before calculating the difference between adjacent pressure data frames in the same channel, the method further includes: Multiple pressure data frames of each channel are filtered using a high-pass filter to obtain multiple filtered pressure data frames of each channel.
4. The method according to claim 1, characterized in that, The pressure data includes pressure data from at least two time periods; the physiological index value of the user is obtained by detecting physiological indexes using the Lasso model based on the extended feature set corresponding to each airbag, including: When obtaining the extended feature set corresponding to each airbag for a time period, the obtained extended feature set corresponding to each airbag is input into the Lasso model so that the Lasso model can perform physiological index detection based on the extended feature set corresponding to each airbag and obtain the corresponding physiological index value. When physiological index values are obtained for at least two time periods, the average value of the physiological index values for the at least two time periods is determined; the average value is the user's final physiological index value.
5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Pressure data generated by test users on each airbag of the sample mattress is collected to obtain pressure data samples; the sample mattress is the pressure-sensing mattress or other pressure-sensing mattress used in the model training phase. Based on the pressure data sample and the baseline reference value, determine the pressure change value of each air bladder in the sample mattress; The pressure change values of each air bladder in the sample mattress are expanded using a polynomial to obtain the extended feature set corresponding to each air bladder in the sample mattress; wherein, the extended feature set corresponding to each air bladder in the sample mattress includes a training set. The initial Lasso model is trained based on the training set until the model converges, resulting in the trained Lasso model.
6. The method according to claim 5, characterized in that, The training of the initial Lasso model based on the training set includes: The initial Lasso model is trained based on the training set, and the loss value is calculated during the training process. The loss value includes the physiological index prediction loss value and the regularization term loss value. The parameters of the initial Lasso model are adjusted based on the loss value.
7. A device for detecting a physiological index, characterized in that, The device includes: The data acquisition module is used to collect pressure data generated by the user on each airbag of the pressure-sensing mattress; the position of each airbag corresponds to different key parts of the user. The determination module is used to determine the pressure change value of each airbag based on the pressure data and the baseline reference value; The processing module is used to expand the pressure change values of each airbag through a polynomial to obtain the extended feature set corresponding to each airbag; The detection module is used to detect physiological indices based on the extended feature set corresponding to each airbag using the Lasso model, and to obtain the physiological index value of the user.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.