Pulse analysis method based on pulse diagnosis instrument and pulse diagnosis monitoring system

By acquiring pressure change signals on a pulse diagnostic instrument, and using wavelet decomposition and noise removal techniques, the pulse cycle characteristic values ​​are identified, a pulse image is generated, and input into a standard model. This solves the problem of low detection accuracy in existing pulse diagnostic instruments, realizes accurate pulse analysis and automated diagnosis, and improves the accuracy and efficiency of TCM diagnosis.

CN121400798APending Publication Date: 2026-01-27LIULIU (GUANGZHOU) CLOUD TECH CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202511601848.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2024-11-25
Filing Date
2025-11-04
Publication Date
2026-01-27

AI Technical Summary

Technical Problem

Existing pulse diagnosis instruments have low detection accuracy and cannot provide accurate pulse data, thus failing to meet the needs of traditional Chinese medicine diagnosis.

Method used

The pressure change signal is obtained by the pressure detection component of the pulse diagnosis instrument. Waveform decomposition is performed using wavelet basis functions and decomposition scale parameters to remove noise, identify pulse cycle feature values, generate pulse image, and input it into a standard model to determine the pulse type. Combined with multinomial fitting and sentiment analysis, intelligent diagnosis is achieved.

Benefits of technology

It improves the accuracy and completeness of pulse data, realizes intelligent and automated pulse analysis, enhances the accuracy and efficiency of TCM diagnosis, and provides support for users' health management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121400798A_ABST
    Figure CN121400798A_ABST
Patent Text Reader

Abstract

The embodiment of the invention relates to the technical field of pulse diagnosis, and discloses a pulse analysis method based on a pulse diagnosis instrument, which comprises the following steps: processing a pressure change signal within a set time to obtain an original pulse waveform signal associated with a user; performing high-frequency denoising processing on the first pulse frequency signal according to a set noise threshold value; acquiring all wavelet coefficients associated with the high-frequency detail signal, traversing all the wavelet coefficients, determining that the absolute value of the wavelet coefficient at a certain point is greater than the absolute value of the wavelet coefficient adjacent to the wavelet coefficient, and taking the wavelet coefficient as a signal abrupt change point; generating a corresponding pulse image of the user based on the pulse condition data of the user within a set time range; and inputting the obtained pulse condition graph of the user into a pre-constructed standard pulse condition model to determine a pulse condition type result of the user, and outputting the pulse condition type result of the user. According to the scheme provided by the embodiment of the invention, the user can conveniently, quickly and accurately know the pulse condition and the physique state of the user, and the use convenience is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of electronic equipment technology, and specifically to a pulse analysis method and pulse monitoring system based on a pulse diagnostic instrument. Background Technology

[0002] Currently, pulse diagnosis is one of the four diagnostic methods in Traditional Chinese Medicine (TCM), a unique diagnostic approach. It primarily utilizes the sensation of the fingers to analyze the pulse's "location, rate, shape, and force," thereby assessing the functional state of the internal organs and achieving non-invasive diagnosis, which is of positive significance for disease diagnosis and treatment. Although some pulse diagnosis devices are available on the market for pulse detection, their accuracy is not high, failing to provide adequate data. Therefore, designing a more accurate detection method has become a pressing technical problem for those skilled in the art. Summary of the Invention

[0003] To address the aforementioned shortcomings, this invention discloses a pulse analysis method based on a pulse diagnostic instrument, which can achieve accurate detection of pulse type.

[0004] The first aspect of this invention discloses a pulse analysis method based on a pulse diagnostic instrument, comprising: The pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period is acquired, and the pressure change signal within the set time period is processed to obtain the original pulse waveform signal associated with the user. The original pulse waveform signal is decomposed according to the selected wavelet basis function and decomposition scale parameter to obtain the first pulse frequency signal and the second pulse frequency signal under different decomposition scales; wherein, the first pulse frequency signal is used to characterize the high-frequency detail signal, the high-frequency detail signal includes high-frequency white noise and power frequency noise, and the second pulse frequency noise is used to characterize the low-frequency approximation signal. The first pulse frequency signal is subjected to high-frequency denoising processing according to the set noise threshold to remove high-frequency white noise and power frequency noise in the high-frequency detail signal; all wavelet coefficients associated with the high-frequency detail signal are obtained, all wavelet coefficients are traversed and it is determined that the absolute value of the wavelet coefficient at a certain point is greater than the absolute value of its adjacent wavelet coefficients, and this is taken as the signal change point. The pulse cycle characteristic value is determined based on all signal abrupt change points. The user pulse data within a set time range is determined based on the pulse cycle characteristic value. The corresponding user pulse image is generated based on the user pulse data within the set time range. The obtained user pulse image is input into a pre-built standard pulse image model to determine the user's pulse image type result, and the user's pulse image type result is output.

[0005] As an optional implementation, in a first aspect of the present invention, generating a corresponding user pulse image based on user pulse image data within a set time range includes: Acquire user pulse data within a set time range, and construct a cubic polynomial equation for each user pulse data, wherein the cubic polynomial equation is the same at each data point and the first and second derivatives are continuous between adjacent intervals. The obtained cubic polynomial equation is calculated to obtain the parameter values ​​of each coefficient in the cubic polynomial equation. Based on the calculated parameter values ​​of each coefficient in the cubic polynomial equation, a pulse fitting curve is constructed on the interval of each data point. Based on the error values ​​between the user's pulse data and the pulse fitting curve, if all error values ​​are within the error threshold, the pulse fitting curve is determined as the user's pulse map.

[0006] As an optional implementation, in a first aspect of the present invention, after obtaining all wavelet coefficients associated with the high-frequency detail signal, the method further includes: All wavelet coefficients are filtered according to a pre-set extreme value filtering threshold. The filtered wavelet coefficients are then traversed to determine the absolute value of a wavelet coefficient at a certain point that is greater than the absolute value of its adjacent wavelet coefficients. And / or, receive detection attribute information input by the user, the detection attribute information including user gender, detection time and detection season.

[0007] As an optional implementation, in the first aspect of the present invention, the pressure detection component includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure monitoring points arranged horizontally in sequence; The pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period includes: By setting the pressure detection grid on the pulse diagnostic instrument, surface pressure distribution data after the user's wrist is placed is collected synchronously. The pressure gradient field of the entire pressure detection grid area is determined based on the surface pressure distribution data. Identify characteristic regions in the pressure gradient field where pressure gradient vectors converge or change. Determine an optimal initial measurement point that meets preset conditions from the feature region.

[0008] As an optional implementation, in a first aspect of the present invention, the logic for determining the optimal initial measurement point includes: Calculate the pressure gradient parameter for each grid point and identify the set of candidate points whose pressure gradient parameters exceed a first threshold; In the candidate point set, find the point with the minimum divergence of the pressure gradient direction in its surrounding local region; The point with the smallest divergence in the direction of the pressure gradient is determined as the optimal initial measurement point; After determining the point with the minimum divergence in the pressure gradient direction as the optimal initial measurement point, the method further includes: The detection component with vertical pressurization capability is controlled to apply pressure at different pressure levels in the vertical direction and acquire the corresponding pulse waveform signals; Based on the signal quality assessment algorithm, the pulse waveforms acquired at different depths are scored; The point corresponding to the highest signal quality score is determined as the final accurate pulse measurement point.

[0009] As an optional implementation, in the first aspect of the present invention, after outputting the pulse type result of the user, the following step is further included: Retrieve a general question library associated with pulse type; each question in the general question library has pre-configured attribute information; Based on the pulse type results, a targeted set of questions is selected from the general question library and generated, and then presented to the user. Receive user feedback answers to the set of questions; The feedback answer and the pulse type result are fused to generate a comprehensive diagnostic report; And / or, following the output user's pulse type result, it also includes: A moxibustion therapy sequence is generated based on the pulse type results; wherein, the moxibustion therapy sequence includes at least: one or more target acupoints, and a suggested therapy time corresponding to each acupoint; The generated moxibustion treatment sequence is presented to the user in the form of a visual chart, which includes the location of acupoints, the order of moxibustion, and the treatment time for each acupoint.

[0010] As an optional implementation, in the first aspect of the present invention, the pulse analysis method further includes: Acquire user facial images at various time points captured by the camera; User facial images at each time detection point are input into a pre-built facial analysis model for identification to determine the user's emotion type at each time detection point; wherein, the facial analysis model is constructed using a convolutional neural network; The user's emotion type at each time point is input into a waveform modification model based on the relationship between emotion and pulse to optimize pulse parameters and correct the resulting user pulse map.

[0011] A second aspect of this invention discloses a pulse monitoring system, comprising: A housing assembly having a receiving cavity for accommodating the wrist to be measured. A first inflation assembly, comprising a first air pump and a first airbag, wherein the first airbag is disposed at the top of the accommodating cavity, and the first air pump is used to inflate the first airbag; The second inflation assembly includes a second air pump and a second airbag. The second airbag is disposed at the bottom of the accommodating cavity, and the second air pump is used to inflate the first airbag. A pressure detection component is disposed on the first airbag and is used to detect the pressure on the wrist to be measured placed in the accommodating cavity. The control module is electrically connected to the first air pump, the second air pump, and the pressure detection component. The control module is used to control the first air pump and the second air pump to work according to the received measurement control signal so that the first airbag and the second airbag inflate, thereby enabling the pressure detection component located at the first airbag to detect the pulse of the user's wrist.

[0012] As an optional implementation, in a second aspect of the present invention, the number of the first airbags is two, and the two first airbags are symmetrically arranged on the top of the accommodating cavity; the pressure detection component is a flexible pressure sensor, and the two ends of the flexible pressure sensor are respectively connected to the two first airbags. The flexible pressure sensor is provided with a first pressure sensor group and a second pressure sensor group. The first pressure sensor group includes a first pressure detection area, a second pressure detection area and a third pressure detection area arranged longitudinally in sequence. The second pressure sensor group includes a fourth pressure detection area, a fifth pressure detection area and a sixth pressure detection area arranged longitudinally in sequence. Each pressure detection area includes multiple pressure detection points arranged laterally in sequence. During the first pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the first pressure detection area, the second pressure detection area, and the third pressure detection area, respectively. During the second pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the fourth pressure detection area, the fifth pressure detection area, and the sixth pressure detection area, respectively.

[0013] As an optional implementation, in a second aspect of the present invention, the number of the first airbags is two, and the two first airbags are symmetrically arranged on the top of the accommodating cavity; the pressure detection component is a flexible pressure sensor, and the two ends of the flexible pressure sensor are respectively connected to the two first airbags. The flexible pressure sensor is provided with a first pressure sensor group and a second pressure sensor group. The first pressure sensor group includes a first pressure detection area, a second pressure detection area and a third pressure detection area arranged longitudinally in sequence. The second pressure sensor group includes a fourth pressure detection area, a fifth pressure detection area and a sixth pressure detection area arranged longitudinally in sequence. Each pressure detection area includes multiple pressure detection points arranged laterally in sequence. During the first pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the first pressure detection area, the second pressure detection area, and the third pressure detection area, respectively. During the second pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the fourth pressure detection area, the fifth pressure detection area, and the sixth pressure detection area, respectively.

[0014] As an optional implementation, in the second aspect of the present invention, the pulse monitoring system further includes a backend server, a communication module, and a camera module. The camera module is used to transmit the monitored user facial expressions and camera attributes to the backend server through the communication module. The camera attributes include camera number information and camera timestamp. The control module or the backend server is used to execute the pulse analysis method based on the pulse diagnostic instrument as described in any one of the first aspects of the present invention. The backend server is used to determine the corresponding pulse diagnostic instrument number based on the camera attributes and the camera-pulse diagnostic instrument mapping table. The backend server is also used to recognize the user's facial expressions to determine the user's current emotional state, and update the user's pulse data based on the current emotional state. The control module is used to control the first air pump and the second air pump to work according to the set three-level control mode, and to work according to the first control parameter, the second control parameter and the third control parameter respectively. And / or, the flexible pressure sensor includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure detection points arranged laterally in sequence. The length of the first pressure detection area, the second pressure detection area, the third pressure detection area, the fourth pressure detection area, the fifth pressure detection area and the sixth pressure detection area is any value between 26mm and 30mm, and the width is any value between 6mm and 8mm.

[0015] A third aspect of the present invention discloses an electronic device, comprising: a memory storing executable program code; a processor coupled to the memory; the processor calling the executable program code stored in the memory to execute the pulse analysis method based on a pulse diagnostic instrument disclosed in the first aspect of the present invention.

[0016] A fourth aspect of the present invention discloses a computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the pulse analysis method based on a pulse diagnostic instrument disclosed in the first aspect of the present invention.

[0017] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: In this embodiment of the invention, the pressure detection component on the pulse diagnostic instrument can accurately capture the pressure change signal of the user's pulse within a set time, and then process it to obtain the raw pulse waveform signal. This ensures the accuracy and integrity of the pulse data, providing a reliable foundation for subsequent analysis. By inputting the user's pulse image into a pre-built standard pulse image model, the system can automatically determine the user's pulse type, realizing intelligent and automated pulse image analysis. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of the pulse analysis method based on a pulse diagnostic instrument disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram of the process for pulse curve fitting disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the detected floating pulse pattern disclosed in an embodiment of the present invention; Figure 4 This is a schematic diagram of the detected central vein pulse pattern disclosed in an embodiment of the present invention. Figure 5 This is a schematic diagram of the detected deep pulse pattern disclosed in an embodiment of the present invention; Figure 6 This is a schematic diagram of a normal pulse width disclosed in an embodiment of the present invention; Figure 7 This is a schematic diagram of the surging pulse pattern disclosed in an embodiment of the present invention; Figure 8 This is a schematic diagram of the pulse pattern of a fine pulse as disclosed in an embodiment of the present invention; Figure 9 This is a schematic diagram showing the questionnaire disclosed in an embodiment of the present invention; Figure 10 This is a schematic diagram showing the generation of the conditioning scheme disclosed in the embodiments of the present invention.

[0020] Figure 11 This is a schematic diagram of the pulse diagnosis instrument disclosed in an embodiment of the present invention; Figure 12 This is an exploded structural diagram of the pulse diagnosis instrument disclosed in an embodiment of the present invention; Figure 13 This is a cross-sectional view of the pulse diagnosis instrument disclosed in an embodiment of the present invention; Figure 14 This is a schematic diagram of the structure of the pressure detection component disclosed in an embodiment of the present invention.

[0021] Reference numerals: 1. Housing assembly; 11. Upper housing; 12. Main housing; 13. Lower housing; 2. Receiving cavity; 3. First air pump; 4. Second air pump; 5. First airbag; 6. Second airbag; 7. Pressure detection assembly; 71. First pressure detection area; 72. Second pressure detection area; 73. Third pressure detection area; 74. Fourth pressure detection area; 75. Fifth pressure detection area; 76. Sixth pressure detection area; 8. Control module. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It should be noted that the terms first, second, third, fourth, etc., in the specification and claims of this invention are used to distinguish different objects, not to describe a specific order. The terms used in the embodiments of this invention include and have, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0024] While some pulse diagnostic devices are available on the market for detecting a user's pulse, their accuracy is not high, failing to provide adequate data. Therefore, this invention discloses a pulse analysis method, system, electronic device, and storage medium based on a pulse diagnostic device. Through a pressure detection component on the pulse diagnostic device, it can accurately capture the pressure change signal of the user's pulse within a set time, and then process it to obtain the raw pulse waveform signal. This ensures the accuracy and integrity of the pulse data, providing a reliable foundation for subsequent analysis. By inputting the user's pulse image into a pre-constructed standard pulse image model, the system can automatically determine the user's pulse type, achieving intelligent and automated pulse analysis.

[0025] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating the pulse analysis method based on a pulse diagnostic instrument disclosed in an embodiment of the present invention. The execution entity of the method described in this embodiment is an execution entity composed of software and / or hardware. This execution entity can receive relevant information via wired or / or wireless means and can send certain instructions. It may also have certain processing and storage functions. This execution entity can control multiple devices, such as remote physical servers or cloud servers and related software, or local hosts or servers and related software that perform related operations on devices located in a certain location. In some scenarios, multiple storage devices can also be controlled; these storage devices may be placed in the same location as the device or in different locations. Figure 1 As shown, the pulse analysis method based on a pulse diagnostic instrument includes the following steps: S101: Obtain the pressure change signal detected by the pressure detection component set at the pulse diagnostic instrument within a set time period, and process the pressure change signal within the set time period to obtain the original pulse waveform signal associated with the user. S102: The original pulse waveform signal is decomposed according to the selected wavelet basis function and decomposition scale parameter to obtain the first pulse frequency signal and the second pulse frequency signal under different decomposition scales; wherein, the first pulse frequency signal is used to characterize the high-frequency detail signal, the high-frequency detail signal includes high-frequency white noise and power frequency noise, and the second pulse frequency noise is used to characterize the low-frequency approximation signal. S103: Perform high-frequency denoising processing on the first pulse frequency signal according to the set noise threshold to remove high-frequency white noise and power frequency noise in the high-frequency detail signal; obtain all wavelet coefficients associated with the high-frequency detail signal, traverse all wavelet coefficients and determine that the absolute value of the wavelet coefficient at a certain point is greater than the absolute value of its adjacent wavelet coefficients, and take it as the signal mutation point. S104: Determine the pulse cycle characteristic value based on all signal abrupt change points, determine the user pulse data within a set time range based on the pulse cycle characteristic value, and generate the corresponding user pulse image based on the user pulse data within the set time range. S105: Input the obtained user pulse image into the pre-built standard pulse image model to determine the user's pulse image type result, and output the user's pulse image type result.

[0026] The method of this invention first acquires the pressure change signal within a set time period through the pressure detection component of the pulse diagnostic instrument, and processes it to obtain the raw pulse waveform signal. This step ensures the accurate capture of the pulse signal, providing a reliable basis for subsequent analysis.

[0027] Wavelet basis functions and decomposition scale parameters are used to decompose the original pulse waveform signal, obtaining the first pulse frequency signal (high-frequency detail signal) and the second pulse frequency signal (low-frequency approximation signal) at different decomposition scales. This method can effectively separate the different components in the pulse signal.

[0028] Next, the high-frequency detail signal is denoised according to the set noise threshold to remove high-frequency white noise and power frequency noise. This step improves the signal-to-noise ratio of the pulse signal, making subsequent analysis more accurate. Based on the preprocessing, high-frequency and low-frequency noise are further removed. High-frequency noise may originate from the equipment itself or interference from the external environment, while low-frequency noise may originate from interference from physiological activities such as breathing and heartbeat. The purpose of denoising is to improve the signal-to-noise ratio, making subsequent feature extraction and classification more accurate. In the denoised pulse signal, pulse cycle feature values ​​are extracted using algorithms such as the modulus maxima principle. These feature values ​​include the amplitude of the main peak, the amplitude of the main straight ... Low-frequency detail signals contain the main low-frequency components of the pulse signal, which reflect the overall trend of the pulse signal, such as the periodic changes in the pulse and the average heart rate.

[0029] This method identifies pulse cycle characteristic values ​​by determining signal abrupt change points. It accurately captures periodic changes in pulse signals, thus accurately reflecting the user's pulse characteristics. Based on these pulse cycle characteristic values, user pulse data within a set time range is determined, and a corresponding user pulse diagram is generated. This provides users with intuitive pulse information, facilitating subsequent health analysis. The obtained user pulse diagram is input into a pre-constructed standard pulse model to determine the user's pulse type. This step achieves intelligent and automated pulse diagnosis, improving accuracy and efficiency. The user's pulse type result is then output. This pulse analysis method based on a pulse diagnostic instrument achieves accurate, comprehensive, and intelligent analysis of the user's pulse through steps such as precise pulse signal capture, effective decomposition and denoising, accurate pulse characteristic identification, and intelligent pulse diagnosis. This not only improves the accuracy and efficiency of traditional Chinese medicine diagnosis but also provides strong support for user health management. Specifically, when the autocorrelation threshold between the waveform expression and the 28 common pulse types reaches 0.6 or higher, the pulse expression is considered valid.

[0030] More preferably, such as Figure 2As shown, the step of generating a corresponding user pulse map based on user pulse data within a set time range includes: S1041: Acquire user pulse data within a set time range, and construct a cubic polynomial equation for each user pulse data, wherein the cubic polynomial equation is the same at each data point and the first and second derivatives between adjacent intervals are continuous. S1042: Calculate the obtained cubic polynomial equation to obtain the parameter values ​​of each coefficient in the cubic polynomial equation, and construct a pulse fitting curve on the interval of each data point based on the calculated parameter values ​​of each coefficient in the cubic polynomial equation. S1043: Based on the error value between the user's pulse data and the pulse fitting curve, if all error values ​​are within the error threshold, then the pulse fitting curve is determined as the user's pulse map.

[0031] This invention constructs a cubic polynomial equation for each user's pulse data, ensuring that the equation is identical at every data point and that the first and second derivatives are continuous between adjacent intervals. This continuity guarantees a smooth transition of pulse data, avoiding analytical errors caused by abrupt changes or jumps in data. By calculating the parameter values ​​of the coefficients in the cubic polynomial equation and constructing a pulse fitting curve for each data point interval, accurate fitting of the pulse waveform can be achieved. This fitting method considers not only the accuracy of the data points but also the changing trends between data points, thereby improving the accuracy and reliability of the fitting.

[0032] The accuracy of the fit can be assessed by comparing the error values ​​between the user's pulse data and the fitted pulse curve. When all error values ​​are within the error threshold, the fitted pulse curve is determined as the user's pulse map, ensuring its accuracy and reliability. This validation step helps reduce errors caused by improper fitting and improves the accuracy of pulse analysis. Using cubic polynomial equations for fitting and validation can automatically generate user pulse maps without manual intervention. This not only improves the efficiency of pulse analysis but also makes the process more intelligent and automated.

[0033] More preferably, after obtaining all wavelet coefficients associated with the high-frequency detail signal, the method further includes: All wavelet coefficients are filtered according to a pre-set extreme value filtering threshold. The filtered wavelet coefficients are then traversed to determine the absolute value of a wavelet coefficient at a certain point that is greater than the absolute value of its adjacent wavelet coefficients. And / or, receive detection attribute information input by the user, the detection attribute information including user gender, detection time and detection season.

[0034] The solution in this embodiment of the invention, after obtaining all wavelet coefficients associated with the high-frequency detail signal, filters the wavelet coefficients using a pre-set extreme value screening threshold to remove extreme points that clearly do not conform to the characteristics of the pulse signal. This reduces the impact of noise and outliers on the identification of signal abrupt changes, thereby improving the accuracy of identification. The filtered wavelet coefficients are then iterated through to determine which wavelet coefficient has an absolute value greater than the absolute values ​​of its adjacent wavelet coefficients; this is identified as a signal abrupt change point. This step further refines the identification of abrupt changes, ensuring that the abrupt changes accurately reflect the periodic changes of the pulse signal.

[0035] This invention enables more personalized pulse analysis by receiving user-inputted detection attribute information (such as user gender, detection time, and detection season). Pulse characteristics may differ across genders, times, and seasons. By considering these attributes, pulse data can be interpreted more accurately, improving the accuracy and relevance of the analysis.

[0036] In this embodiment of the invention, the explanation of special cases mainly addresses the following aspects: gender differences, with women exhibiting slippery, soft, and weak pulses, which are more common than in men and therefore often considered normal. Seasonal differences. Adjustments are made based on the pulse characteristics of the four seasons. A wiry pulse is common in spring, a surging pulse in summer, a floating pulse in autumn, and a deep pulse in winter. Compared to other seasons, these pulse characteristics appear in their corresponding seasons, and the indicators tend to be normal.

[0037] More preferably, before acquiring the pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period, the method further includes: The system receives an arm measurement signal input by the user. If the arm measurement signal is a left-hand signal, the system activates the first pressure sensor group to detect pressure in the corresponding area of ​​the user's wrist. If the arm measurement signal is a right-hand signal, the system activates the second pressure sensor group to detect pressure in the corresponding area of ​​the user's wrist. The method of acquiring the pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period, and processing the pressure change signal within the set time period to obtain the original pulse waveform signal associated with the user, further includes: When the inflation assembly reaches the first preset state, the pressure change signals detected by the corresponding three pressure detection areas in the pressure detection assembly set at the pulse diagnostic instrument within a set time period are acquired. If the pressure change signals meet the set requirements, the pressure change signals detected by the corresponding pressure detection areas are processed to obtain the original pulse waveform signals associated with the user. The number of the original pulse waveform signals is three. If the pressure change signals do not meet the set requirements, the inflation assembly is controlled to work according to the set adjustment state to acquire waveform signals. The pressure value in the set adjustment state is greater than the pressure value in the first preset state. When the inflation component reaches the second preset state, the pressure change signals detected by the corresponding three pressure detection areas in the pressure detection component set at the pulse diagnostic instrument within a set time period are acquired, and the pressure change signals detected by the corresponding pressure detection areas are processed to obtain the original pulse waveform signals associated with the user; the number of the original pulse waveform signals is three. When the inflation component reaches the third set state, the pressure change signals detected by the corresponding three pressure detection areas in the pressure detection component set at the pulse diagnostic instrument within a set time are acquired, and the pressure change signals detected by the corresponding pressure detection areas are processed to obtain the original pulse waveform signals associated with the user; the number of the original pulse waveform signals is three.

[0038] Specifically, when the pulse diagnosis instrument is activated, it will indicate whether the instrument is for left or right hand testing. 48 monitoring points are selected on each side, and the pressure changes are observed simultaneously. Specifically: the Cun pulse is located on the left hand [1-16], and on the right hand [81-96]; the Guan pulse is located on the left hand [17-32], and on the right hand [65-80]; the Chi pulse is located on the left hand [33-48], and on the right hand [49-64]. The upper airbag is pressurized to simulate the superficial, middle, and deep pulse diagnosis techniques used in Traditional Chinese Medicine. Through multiple experiments and analyses, the optimal pressure is determined to be 12000 Pa for superficial pulse diagnosis, 15000 Pa for middle pulse diagnosis, and 18000 Pa for deep pulse diagnosis. For pulses that are deep or indistinct, the optimal pressure is 15000 Pa for superficial pulse diagnosis, 18000 Pa for middle pulse diagnosis, and 20000 Pa for deep pulse diagnosis.

[0039] The pulse at the cun, guan, and chi positions, each with the highest relevant threshold, is used as the observable pulse map. Under the three pressure levels of floating, middle, and deep, the three selected points at the cun, guan, and chi positions are sampled simultaneously for 15 seconds to obtain three raw data segments. Wavelet transform is then repeatedly used to obtain the pulse map.

[0040] When the inflation component reaches different preset states (first preset state, second preset state, and third preset state), the system acquires the pressure change signals detected by the corresponding three pressure detection areas in the pressure detection component set at the pulse diagnostic instrument. This staged, pressure-based detection method can more comprehensively capture subtle changes in the user's pulse and improve the accuracy of pulse analysis.

[0041] Simultaneously, the system determines whether to acquire waveform signals or adjust the working state of the inflation component based on whether the pressure change signal meets the set requirements. This intelligent control strategy ensures the stability and reliability of the measurement process.

[0042] In each inflation state, the system acquires pressure change signals detected in three pressure detection zones and processes them to obtain three raw pulse waveform signals. This not only increases the amount of raw data but also provides richer information for subsequent pulse analysis and diagnosis. These three zones correspond to the user's Cun, Guan, and Chi positions, and dividing these zones facilitates further in-depth analysis and determination by the user.

[0043] More preferably, the pressure detection component includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure monitoring points arranged horizontally in sequence; The pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period includes: By setting the pressure detection grid on the pulse diagnostic instrument, surface pressure distribution data after the user's wrist is placed is collected synchronously. The pressure gradient field of the entire pressure detection grid area is determined based on the surface pressure distribution data. Identify characteristic regions in the pressure gradient field where pressure gradient vectors converge or change. Determine an optimal initial measurement point that meets preset conditions from the feature region.

[0044] The solution of this invention uses a pressure detection grid composed of 96 pressure monitoring points (divided into 16 horizontally arranged detection areas) to comprehensively cover the pulse-related area of ​​the user's wrist, and simultaneously collect surface pressure distribution data. This avoids information omissions that may be caused by single-point or few-point detection, and provides a more comprehensive raw data foundation for subsequent analysis.

[0045] In practice, the pressure detection grid consisting of 96 pressure monitoring points can comprehensively cover the pulse-related area of ​​the user's wrist, and simultaneously collect surface pressure distribution data, avoiding information omissions that may be caused by single-point or few-point detection, and providing a more comprehensive raw data foundation for subsequent analysis.

[0046] Specifically, the optimal initial measurement point that meets preset conditions is selected from the feature region, avoiding the blindness of random point selection. This optimal initial point can more accurately correspond to the region where the pulse signal is strongest and most stable, laying the foundation for obtaining high-quality raw pulse waveform signals, reducing the interference of invalid signals on subsequent analysis, and improving the reliability of the overall pulse analysis.

[0047] The solution of this invention, through the synergistic effect of grid layout, gradient field analysis and feature region recognition, enables the pulse diagnosis instrument to automatically adapt to the differences in wrist shape and pulse position of different users, without relying on manual adjustment or experience judgment, thereby improving the device's adaptability to individual differences and promoting the intelligence and standardization of the pulse diagnosis process.

[0048] Specifically, the spatial coordinates of the 96 pressure detection points are first calibrated: Assuming the pressure detection grid is a two-dimensional plane (x-axis is horizontal, corresponding to the wrist width direction; y-axis is vertical, corresponding to the wrist length direction), each detection point is assigned unique coordinates. For example, the x-coordinates of the 16 horizontal points are x1, x2, ..., x16, and the vertical coordinates are y1, y2, ..., y6, divided by region. The pressure value of each detection point is associated with its coordinates to form a two-dimensional pressure matrix, meaning the pressure value at any point within the grid can be represented by known detection point data. The pressure gradient is a vector containing partial derivatives in the x-direction (horizontal) and y-direction (vertical). The gradient directions of all detection points in the pressure gradient field are traversed, and the angle between the gradient vectors within a local region (e.g., a 3×3 grid neighborhood) is calculated. If the angle between the directions of most gradient vectors within a certain region is less than a set threshold (e.g., 30°), and the vectors point towards the center of the region (i.e., the central region is the gradient convergence point), then the region is determined to be a gradient vector convergence area.

[0049] When an artery pulsates, pressure diffuses outwards, and the gradient vector points from the pulsation center outwards (or from the outwards to the center, depending on the pressure change trend), forming a locally consistent convergent feature, which is significantly different from the chaotic gradient directions in non-pulsating areas. The gradient magnitude of each detection point is calculated, and the standard deviation of the gradient magnitude within the local region is statistically analyzed. If the standard deviation of the gradient magnitude in a certain region exceeds a set threshold, and there are one or more detection points with a significantly larger gradient than surrounding points (e.g., more than twice the mean), then that region is determined to be a region of drastic gradient change. Spatial continuity verification is performed on the convergent and drastically changing regions identified in the above steps: if adjacent detection points both belong to candidate feature regions, they are merged into a single connected region to avoid misjudgments caused by single-point noise.

[0050] More preferably, the logic for determining the optimal initial measurement point includes: Calculate the pressure gradient parameter for each grid point and identify the set of candidate points whose pressure gradient parameters exceed a first threshold; In the candidate point set, find the point with the minimum divergence of the pressure gradient direction in its surrounding local region; The point with the smallest divergence in the direction of the pressure gradient is determined as the optimal initial measurement point; After determining the point with the minimum divergence in the pressure gradient direction as the optimal initial measurement point, the method further includes: The detection component with vertical pressurization capability is controlled to apply pressure at different pressure levels in the vertical direction and acquire the corresponding pulse waveform signals; Based on the signal quality assessment algorithm, the pulse waveforms acquired at different depths are scored; The point corresponding to the highest signal quality score is determined as the final accurate pulse measurement point.

[0051] The solution in this invention calculates pressure gradient parameters and filters out a set of candidate points exceeding a first threshold. This allows for the rapid identification of regions strongly correlated with pulse pulsations, eliminating meaningless low-gradient regions and narrowing the effective measurement range. Furthermore, the optimal initial measurement point is selected from the candidate point set based on the point with the smallest divergence in the surrounding local pressure gradient direction. This accurately locates the core region where the pulse signal is most stable and directionally consistent. This region typically corresponds to the center of the arterial pulsation, avoiding the subjectivity of random or experience-based point selection and laying a high-quality foundation for subsequent measurements.

[0052] Based on the initial measurement point, by applying different pressure levels and acquiring pulse waveforms at multiple depths, and combining this with a signal quality assessment algorithm to score the waveforms, the point with the highest signal quality is ultimately determined as the precise measurement point. This process fully considers the differences in pulse signals at different depths (vertical direction), overcoming the limitations of two-dimensional planar point selection. It ensures that the measurement point not only accurately corresponds to the artery location in the horizontal direction but also matches the depth with the clearest signal and most significant features in the vertical direction, greatly improving the effectiveness of the pulse signal.

[0053] The wrist anatomy, artery location, and depth vary among different users, making traditional fixed-point measurements susceptible to individual differences. This proposed solution dynamically determines the initial point through gradient parameter selection and directional divergence analysis, and optimizes the final point by combining multi-depth compression testing. It can adapt to the physiological characteristics of different users, ensuring that high-quality pulse signals can still be captured even when the artery position is offset or the depth is different, thus improving the universality of the pulse diagnostic instrument.

[0054] Determining the optimal initial measurement point and the final precise measurement point directly reduces interference from noise signals and invalid data. High-quality pulse waveform signals (with the highest signal quality score) can more realistically reflect the user's pulse characteristics, making subsequent wavelet decomposition, denoising, and periodic feature extraction steps more reliable, ultimately improving the accuracy of pulse type judgment.

[0055] The entire point selection process is based on quantified gradient parameters, divergence indices, and signal quality scores. It optimizes the pulse diagnosis from the initial point to the precise point without human intervention, avoiding errors and subjectivity inherent in manual operation. This standardized point selection logic ensures consistency in measurements taken by different users at different times, providing a unified standard for comparative analysis and clinical application of pulse data, and contributing to the objectification and standardization of traditional Chinese medicine pulse diagnosis.

[0056] More preferably, after outputting the pulse type result for the user, the method further includes the following step: A general question library associated with pulse type is retrieved; each question in the general question library is pre-configured with attribute information; wherein the attribute information is determined based on at least one of the following: the importance of the question, the closeness of its association with a specific pulse type, or the symptom level to which it belongs; Based on the pulse type results, a targeted set of questions is selected from the general question library and generated, and then presented to the user. Receive user feedback answers to the set of questions; The feedback answer and the pulse type result are fused to generate a comprehensive diagnostic report; And / or, following the output user's pulse type result, it also includes: A moxibustion therapy sequence is generated based on the pulse type results; wherein, the moxibustion therapy sequence includes at least: one or more target acupoints, and a suggested therapy time corresponding to each acupoint; The generated moxibustion treatment sequence is presented to the user in the form of a visual chart, which includes the location of acupoints, the order of moxibustion, and the treatment time for each acupoint.

[0057] The solution in this invention, by retrieving a general question database associated with pulse types, can combine pulse information with the user's subjective symptom feedback, overcoming the limitations of single pulse data. For example, a certain pulse may correspond to multiple potential symptoms; targeted questioning can further pinpoint the specific cause, expanding the diagnosis from a single pulse dimension to a multi-dimensional analysis of pulse and symptoms, significantly improving the reliability of the comprehensive diagnostic report. Figure 9 The diagram shown illustrates the generation and display of the corresponding questions. Dynamically generating a set of questions based on pulse type avoids interference from irrelevant questions, reduces the user's burden of answering, and improves interaction efficiency. Simultaneously, differentiated configuration of question attributes (such as prioritizing high-importance and highly relevant questions) ensures that core information is obtained first, making the inquiry process more targeted and logical, and enhancing the user experience.

[0058] The structured design of the general question bank (including attribute information) and the standardized collection of feedback answers enable the quantitative integration of subjective symptom information with objective pulse data to form a standardized comprehensive diagnostic report.

[0059] Specifically, a moxibustion sequence containing target acupoints and treatment time is generated based on the pulse type, directly transforming the diagnosis into an executable treatment plan. For example, when a certain pulse type corresponds to qi and blood deficiency, acupoints such as Guanyuan and Zusanli can be recommended with matching moxibustion durations, ensuring that the treatment plan accurately matches the user's constitution and pulse characteristics, avoiding blind moxibustion.

[0060] Visual charts intuitively present the location of acupoints, the order of moxibustion application, and the timing, reducing users' reliance on professional acupoint knowledge. Even non-professional users can accurately locate acupoints and follow the procedures using the charts, solving the problem of traditional Chinese medicine treatment plans being highly specialized and difficult to implement, thus promoting the popularization of traditional Chinese medicine health preservation methods. Figure 10 This is a display diagram generated from the corresponding treatment plan.

[0061] The seamless integration from pulse diagnosis to treatment plan generation forms a complete closed loop of detection-diagnosis-intervention. Compared to simply outputting diagnostic results, this plan provides users with direct guidance for follow-up actions, enhancing the practical value of the pulse diagnosis device, and enabling the concept of TCM syndrome differentiation and treatment to be implemented through technological means, thereby improving users' compliance with TCM treatment.

[0062] The two technical solutions described above together expand the functional boundaries of pulse diagnosis instruments, upgrading them from simple pulse detection to a comprehensive service system encompassing diagnosis, inquiry, reporting, and treatment. By integrating objective pulse data with subjective symptom feedback, and transforming diagnostic results into personalized treatment plans, this approach not only enhances the scientific rigor and standardization of TCM diagnosis but also strengthens user engagement and practicality in health management, propelling the intelligentization of TCM from auxiliary diagnosis to full-process health services.

[0063] Specifically, indexed by pulse type, typical symptom questions highly correlated with that pulse type are stored. For example: a floating and weak pulse (corresponding to Qi deficiency): related questions include whether you often feel weak, whether you are prone to shortness of breath after activity, and whether you sweat easily. (Attribute information: high importance, 90% correlation, symptom level: core symptom). A deep and slow pulse (corresponding to cold stagnation): related questions include whether you are afraid of cold, whether you have a cold pain in your abdomen, and whether your stool is loose. (Attribute information: medium importance, 80% correlation, symptom level: accompanying symptom). The attribute information of each question is annotated with expert experience and clinical data to form a weight value (e.g., importance weight 0.8-1.0, correlation strength 0-100%), which is used for subsequent priority filtering.

[0064] The pulse diagnosis-moxibustion sequence association library uses pulse diagnosis type as the core and binds it to the corresponding standard moxibustion scheme. The location coordinates of each acupoint (based on the human standard acupoint model), moxibustion duration, and sequence rules are all stored in a structured manner and support dynamic retrieval.

[0065] More preferably, the standard pulse model is obtained through the following steps: Obtain a sample training set, which includes pulse image samples; and perform a partitioning operation on the sample training set to obtain a training set, a validation set, and a test set. The pulse sample images are preprocessed. The preprocessed pulse image sample is input into the standard pulse model with configured parameters for training and parameter adjustment until the standard pulse model meets the training requirements, and the current best model parameters are saved; the standard pulse model is a model built based on a convolutional neural network. The trained standard pulse model is used to predict the pulse test images in the test set to obtain the pulse type of each pulse test image. Based on the prediction results and actual annotations of the standard pulse model, a category confusion matrix is ​​generated; the rows of the category confusion matrix represent the actual type and the columns represent the predicted type. The overall performance metrics of the standard pulse model are calculated based on the aforementioned confusion matrix. These overall performance metrics include accuracy, precision, recall, and F1 score. The standard pulse model is then adjusted and optimized based on the analysis results of the aforementioned confusion matrix. The step of inputting the obtained user pulse map into a pre-constructed standard pulse model to determine the user's pulse type includes: The similarity values ​​between the obtained user pulse image and the pre-built standard pulse image model are calculated by using the Euclidean distance formula. The maximum value among the similarity values ​​of each standard pulse image is determined. If the maximum value among the similarity values ​​of each standard pulse image model is greater than the similarity threshold, the user's pulse image type is determined.

[0066] This invention, through obtaining a sample training set containing pulse image samples and dividing it into training, validation, and test sets, ensures data diversity and independence during model training. This helps improve the model's accuracy and generalization ability. Preprocessing the pulse image samples, such as denoising and enhancement, further improves image quality, thereby enhancing model training effectiveness. Training with a standard pulse model based on a convolutional neural network (CNN) and using the training and validation sets for parameter configuration adjustments and model optimization ensures the model can learn deep features in pulse images, improving classification accuracy.

[0067] By using pulse test images from the test set to predict the pulse type of a trained standard pulse model, the pulse type of each test image can be determined. By comparing the prediction results with the ground truth annotations, a confusion matrix can be generated, allowing for the calculation of the model's overall performance metrics (accuracy, precision, recall, and F1 score). This helps to comprehensively evaluate the model's performance and provides direction for subsequent optimization.

[0068] By calculating the similarity score between a user's pulse image and a pre-built standard pulse image model using the Euclidean distance formula, and determining the maximum similarity score among all standard pulse images, the user's pulse type can be more accurately determined. This method considers not only the overall features of the pulse image but also the similarity between pulse images, improving matching accuracy. A similarity threshold is set; the user's pulse type is only determined when the maximum similarity score among all standard pulse image models exceeds the threshold. This strategy further reduces the possibility of misjudgment and improves the accuracy of pulse analysis. By optimizing the acquisition process of the standard pulse image model and the matching process between the user's pulse image and the standard pulse image model, pulse analysis can be made more efficient and practical. This not only helps improve the accuracy and efficiency of TCM diagnosis but also provides more convenient and personalized support for users' health management.

[0069] In this embodiment of the invention, waveform similarity analysis can effectively distinguish pulse types such as normal pulse, wiry pulse, slippery pulse, hesitant pulse, surging pulse, thready pulse, weak pulse, strong pulse, tight pulse, and soft pulse.

[0070] When constructing a specific standard model, pulse data is sampled using a traditional Chinese medicine pulse diagnosis training instrument. The data is then manually labeled, classified, and organized to provide training and learning materials for the AI ​​model. The system provides a data labeling backend for manually labeling the recorded standard pulse data to obtain the manually labeled results.

[0071] More preferably, the pulse analysis method further includes: The pressure change signals detected by the three corresponding pressure detection areas in the pressure detection component set at the pulse diagnostic instrument within a set time period are obtained, and the corresponding pulse position data are determined based on the pressure change signals detected by the three corresponding pressure detection areas within the set time period. The waveform recognition algorithm is used to identify the user's pulse image to determine the corresponding pulse pulsation cycle, amplitude of the cun, guan, and chi waveforms, pulse width data, time interval between each pulsation, strength of pulse fluctuation, smoothness of pulse fluctuation, tension of pulse, and uniformity of waveform. The corresponding pulse type is determined based on the obtained pulse pulsation cycle, amplitude of the cun, guan, and chi waveforms, pulse width data, time interval between each pulsation, strength of pulse fluctuation, smoothness of pulse fluctuation, tension of pulse, and uniformity of waveform. In this embodiment of the invention, a waveform recognition algorithm is used to identify the pulse pattern, which can extract detailed data such as pulse cycle, amplitude, pulse width, pulse interval time, pulse strength, smoothness, tension, and waveform uniformity. These detailed pulse features provide rich information for subsequent pulse type determination and constitution analysis.

[0072] Specifically, pulse rate refers to the pulse's pulsation cycle. It applies to pulses with a fixed cycle, and the cycle frequency can be inferred from the pulse's pulsation frequency. This allows for the differentiation of rapid, normal, slow, moderate, and rapid pulses. Rapid pulse: 90-120 beats per minute; Normal pulse: 70-90 beats per minute; Slow pulse: less than 60 beats per minute; Moderate pulse: 60-70 beats per minute; Rapid pulse: more than 120 beats per minute.

[0073] Pulse length refers to the amplitude of the pulse waves measured by the three fingers corresponding to the cun, guan, and chi positions, which is used to distinguish between long and short pulses. Long pulse: The pulse amplitude at the cun and chi positions is close to that at the guan position (using 10% as the judgment threshold), thus it is judged as a long pulse. Short pulse: The pulse amplitude at the cun or chi positions is low, or cannot be detected upon superficial palpation, or the amplitude is less than 40% of that at the guan position, thus it is judged as a short pulse.

[0074] Pulse width, primarily referring to the amplitude of the pulse wave, is generally used to distinguish related pulse characteristics such as a bounding pulse and a thready pulse. For example... Figures 6 to 8 As shown, under normal circumstances: the pulse width corresponds to a pulse wave amplitude between 150 and 350; a bounding pulse: the pulse width corresponds to a pulse wave amplitude greater than 350; a thready pulse: the pulse width corresponds to a pulse wave amplitude less than 150.

[0075] Based on the uniformity of the waveform and the time interval between each pulse, differentiate between knotted pulse, rapid pulse, and intermittent pulse. Knotted pulse: 60-70 pulses per minute, irregular rhythm, and inconsistent durations of no-pulse periods. Rapid pulse: 90-120 pulses per minute, irregular rhythm, and inconsistent durations of no-pulse periods. Intermittent pulse: 60-70 pulses per minute, irregular rhythm, but with regular pauses.

[0076] The 28 pulse patterns in Traditional Chinese Medicine represent the first comprehensive manifestation of all characteristics; a single characteristic cannot effectively diagnose the pulse's features. Therefore, it is necessary to integrate multiple parameters for specific analysis.

[0077] In practice, the pulse data reported by the pulse diagnosis instrument can be used to observe the pulse conditions in the three segments: superficial, middle, and deep. Specifically, as shown below... Figures 3 to 5 As shown, based on the depth of the pulse, the locational characteristics of floating pulse, deep pulse, hollow pulse, weak pulse, surging pulse, firm pulse, soft pulse, leathery pulse, weak pulse, scattered pulse, and hidden pulse can be distinguished. (Reference) Figures 3 to 5 The pulse recorder records pulse patterns from floating, middle, and deep. The pulse is prominent when the pulse is touched superficially and weak when it is touched deeply, which can be used to determine that it is a floating pulse.

[0078] More preferably, after determining the pulse cycle characteristic value based on all signal abrupt change points, determining the user pulse data within a set time range based on the pulse cycle characteristic value, and generating the corresponding user pulse image based on the user pulse data within the set time range, the method further includes: Key pulse characteristic parameters are extracted from the collected pulse waveform data. These pulse characteristic parameters include the shape and position of the main wave, dicrotic wave, and descending isthmus, as well as the pulse period, frequency, amplitude, peak-to-peak value, rise time, and fall time. The extracted feature parameters are preprocessed, including data cleaning, normalization, and dimensionality reduction. The pulse feature parameters of the newly acquired pulse waveform data are used as observation data and input into the Bayesian model. The Bayesian model calculates the posterior probability of various pulse types based on the prior knowledge base and the observation data. Based on the calculated posterior probability, the most likely pulse type is output as the recognition result. Following the determination of the user's pulse type, the following is also included: By combining the recognition results of the Bayesian model and the waveform similarity recognition results, a comprehensive judgment is made. If the conclusions drawn from the two are consistent or similar, the final pulse type can be determined.

[0079] In this embodiment of the invention, a prior knowledge base of pulse type feature parameters needs to be constructed based on Traditional Chinese Medicine (TCM) theory and clinical practice. This knowledge base contains the feature parameter ranges and distributions of various TCM pulse types (such as normal pulse, wiry pulse, slippery pulse, etc.). The feature parameters of newly acquired pulse waveform data are input into a Bayesian model as observation data. The Bayesian model calculates the posterior probability of various pulse types based on the prior knowledge base and the observation data. During this process, the model considers the degree of matching between the observation data and the prior knowledge, and updates the probability distribution of various pulse types accordingly.

[0080] This invention extracts key pulse characteristic parameters from the acquired pulse waveform data, such as the morphology and location of the main wave, dicrotic wave, and descending isthmus, as well as the pulse period, frequency, amplitude, peak-to-peak value, rise time, and fall time. These characteristic parameters can comprehensively reflect subtle changes in the pulse waveform, providing accurate basic data for subsequent pulse type identification.

[0081] Preprocessing operations are performed on the extracted feature parameters, including data cleaning, normalization, and dimensionality reduction. Data cleaning removes noise and outliers, improving data accuracy; normalization transforms the data to a uniform scale, facilitating subsequent processing; and dimensionality reduction reduces redundant information, improving computational efficiency. These preprocessing operations significantly enhance the recognition performance and stability of the Bayesian model.

[0082] The pulse feature parameters of newly acquired pulse waveform data are input into a Bayesian model as observation data. Based on a prior knowledge base and the observation data, the Bayesian model calculates the posterior probability of various pulse types and outputs the most probable pulse type as the recognition result. This method combines prior knowledge and observation data to achieve efficient and accurate pulse type recognition.

[0083] The results of Bayesian model recognition and waveform similarity recognition are combined for a comprehensive judgment. If the conclusions reached by the two methods are consistent or similar, the final pulse type can be determined. This comprehensive judgment method can make full use of the advantages of different recognition methods, improving the reliability and accuracy of the recognition results.

[0084] More preferably, the pulse analysis method further includes: Acquire user facial images at various time points captured by the camera; User facial images at each time detection point are input into a pre-built facial analysis model for identification to determine the user's emotion type at each time detection point; wherein, the facial analysis model is constructed using a convolutional neural network; The user's emotion type at each time point is input into a waveform modification model based on the relationship between emotion and pulse to optimize pulse parameters and correct the resulting user pulse map.

[0085] Since human emotions do indeed affect pulse wave measurements, this invention incorporates a camera to monitor the user's emotional state during implementation. This emotional state is then used to correct the pulse wave waveform. In this embodiment, the camera captures facial images of the user at various time points, and a pre-built facial analysis model is used for emotion recognition. This step integrates emotional factors into the pulse analysis process. Emotion is a significant factor influencing human physiological state, and its changes are often accompanied by pulse fluctuations. Therefore, considering emotional factors makes pulse analysis more comprehensive and accurate. The facial analysis model is constructed using a convolutional neural network, which has powerful image data processing capabilities and can accurately identify the user's emotion type. By automatically learning features in images, the convolutional neural network can capture subtle changes in facial expressions, thereby achieving accurate emotion recognition.

[0086] By inputting the user's emotion type at various time points into a waveform modification model based on the relationship between emotion and pulse, pulse parameters can be optimized, thereby correcting the resulting user pulse map. This step considers the influence of emotion on the pulse, adjusting pulse parameters to reflect emotional changes, thus improving the accuracy and reliability of the pulse map.

[0087] By combining emotion recognition and pulse analysis, more personalized health assessments can be provided to users. The body's physiological responses differ depending on its emotional state. Therefore, considering emotional factors when assessing a user's health status can make the assessment results more closely reflect the user's actual situation, providing more targeted health advice.

[0088] In this embodiment of the invention, machine learning algorithms are used for correction. For example, a regression model can be used to establish a relationship model between emotion and pulse, and this model can be used to correct the pulse measurement data. Alternatively, a classification algorithm can be used to identify emotion types, and different corrections can be made to the pulse measurement data based on different emotion types. A large dataset containing emotion labels and pulse measurement data is collected; a machine learning algorithm is used to train a model to establish the relationship between emotion and pulse; the trained model is then applied to new pulse measurement data for correction calculations.

[0089] In this embodiment of the invention, a pressure sensor array can also be used to achieve automatic pulse finding; the pressure sensor array can be arranged in the radial artery area of ​​the wrist, which is a commonly used pulse finding location in traditional Chinese medicine pulse diagnosis.

[0090] Each sensor in the array can independently detect pressure changes at its location, enabling comprehensive capture of pulse signals. When a user places their wrist on the sensor array, the sensors begin acquiring pulse signals. The acquired signals undergo preprocessing operations such as filtering and amplification to improve the signal-to-noise ratio and accuracy. Based on the acquired pulse signals, the system can use an automatic pulse-finding algorithm to determine the precise location of the pulse. This algorithm identifies the pulse signal and determines its location by analyzing its waveform, frequency, amplitude, and other characteristics. The system can adjust the sensor array layout or acquisition parameters in real time based on the results of the automatic pulse-finding algorithm to ensure accurate capture of the pulse signal. For example, if the algorithm determines that the pulse location is offset, the system can automatically adjust the position of the sensor array or reassign acquisition tasks.

[0091] This invention employs advanced facial emotion recognition technology, such as deep learning models (e.g., Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), etc.), to recognize users' facial emotions in real time. During model training, a facial image dataset containing various emotion labels is used to ensure the model can accurately recognize a wide range of emotions. Post-processing of the recognition results, such as smoothing filtering and thresholding, is performed to improve the stability and accuracy of the recognition.

[0092] In addition to the features mentioned above, other facial features can be identified to assist in subsequent processing. These other facial features refer to those helpful in Traditional Chinese Medicine (TCM) diagnosis, such as facial color: Yellow: may reflect poor spleen and stomach function; White: may indicate problems with the lungs and respiratory system; Red: may be related to cardiovascular abnormalities; Blue: may indicate liver, gallbladder, and immune system dysfunction; Black: may reflect poor kidney and bladder function.

[0093] Facial luster and shape: A bright and lustrous complexion indicates that the essence and energy of the internal organs are not yet depleted; a withered and dull complexion indicates that the essence and energy of the internal organs are severely depleted. Changes in facial shape, such as swelling or emaciation, may also reflect internal organ diseases.

[0094] Facial reflex zones: Traditional Chinese medicine believes that different areas of the face correspond to internal organs. For example, the forehead is related to the brain, throat, lungs, and heart; the philtrum is related to the uterus and bladder; the tip of the nose reflects the condition of the spleen and stomach; and the area below the cheekbones is related to the large intestine and kidneys. By observing changes in the color and shape of these reflex zones, one can gain further insight into the health of internal organs. These characteristics provide more comprehensive data.

[0095] In this embodiment of the invention, the pressure detection component on the pulse diagnostic instrument can accurately capture the pressure change signal of the user's pulse within a set time, and then process it to obtain the raw pulse waveform signal. This ensures the accuracy and integrity of the pulse data, providing a reliable foundation for subsequent analysis. By inputting the user's pulse image into a pre-built standard pulse image model, the system can automatically determine the user's pulse type, realizing intelligent and automated pulse image analysis.

[0096] Example 2 like Figures 11 to 14 As shown, an embodiment of the present invention discloses a pulse diagnosis monitoring system, comprising: The housing assembly 1 has a receiving cavity 2 for accommodating the wrist to be measured; specifically, the housing assembly 1 includes an upper housing 11, a main housing 12, and a lower housing 13 that are stacked in sequence; the main housing 12 has the receiving cavity 2. The first inflation assembly includes a first air pump 3 and a first airbag 5. The first airbag 5 is disposed at the top of the accommodating cavity 2, and the first air pump 3 is used to inflate the first airbag 5. The second inflation assembly includes a second air pump 4 and a second airbag 6. The second airbag 6 is disposed at the bottom of the accommodating cavity 2. The second air pump 4 is used to inflate the first airbag 5. Pressure detection component 7, the detection component is disposed on the first airbag 5, the detection component is used to perform pressure detection on the wrist to be measured placed in the accommodating cavity 2; The control module 8 is electrically connected to the first air pump 3, the second air pump 4, and the pressure detection component 7. The control module 8 is used to control the first air pump 3 and the second air pump 4 to work according to the received measurement control signal so that the first airbag 5 and the second airbag 6 inflate, thereby enabling the pressure detection component 7 located at the first airbag 5 to detect the pulse of the user's wrist.

[0097] Specifically, the user places the wrist to be measured into the cavity 2 of the pulse monitoring system. The cavity 2 consists of an upper shell 11, a main shell 12, and a lower shell 13, forming a closed space to facilitate pulse detection.

[0098] When the user initiates the measurement, the control module 8 receives a measurement control signal. The control module 8 then activates the first air pump 3 and the second air pump 4. The first air pump 3 inflates the first airbag 5 located at the top of the receiving cavity 2, causing it to expand and fit snugly against the user's wrist. The second air pump 4 inflates the second airbag 6 located at the bottom of the receiving cavity 2, causing it to expand and support the area below the user's wrist, ensuring stability during the measurement process.

[0099] As the first airbag 5 and the second airbag 6 inflate, they together provide a stable and uniform contact surface for the user's wrist. The pressure detection component 7, located on the first airbag 5, begins to operate, detecting pressure on the user's wrist. During this process, the pressure detection component 7 can capture minute pressure changes caused by the pulse. The pressure detection component 7 transmits the detected pressure signal to the control module 8. The control module 8 processes and analyzes the received signal, extracting relevant pulse information (such as heart rate, pulse intensity, etc.).

[0100] Finally, the control module 8 can display this information on the user interface or transmit the data to an external device via other means (such as wireless transmission).

[0101] In this embodiment of the invention, the inflation of the first airbag 5 and the second airbag 6 provides a stable and uniform contact surface for the user's wrist, reducing the influence of external factors on the measurement results. The pressure detection component 7 can capture minute pressure changes, thereby improving the accuracy of pulse detection. The design of the receiving cavity 2 keeps the user's wrist stable during measurement, reducing errors caused by wrist movement. The system is easy to operate; the user simply places their wrist in the receiving cavity 2 and starts the measurement. The system is suitable for people of different ages and body types because the inflation levels of the first airbag 5 and the second airbag 6 can be adjusted as needed.

[0102] More preferably, there are two first airbags 5, which are symmetrically arranged on the top of the accommodating cavity 2; the pressure detection component 7 is a flexible pressure sensor, and the two ends of the flexible pressure sensor are respectively connected to the two first airbags 5. The flexible pressure sensor is provided with a first pressure sensor group and a second pressure sensor group. The first pressure sensor group includes a first pressure detection area 71, a second pressure detection area 72 and a third pressure detection area 73 arranged longitudinally in sequence. The second pressure sensor group includes a fourth pressure detection area 74, a fifth pressure detection area 75 and a sixth pressure detection area 76 arranged longitudinally in sequence. Each pressure detection area includes multiple pressure detection points arranged laterally in sequence. During the first pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the first pressure detection area 71, the second pressure detection area 72, and the third pressure detection area 73, respectively. During the second pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the fourth pressure detection area 74, the fifth pressure detection area 75, and the sixth pressure detection area 76, respectively.

[0103] More preferably, there are two first airbags 5, which are symmetrically arranged on the top of the accommodating cavity 2; the pressure detection component 7 is a flexible pressure sensor, and the two ends of the flexible pressure sensor are respectively connected to the two first airbags 5. The flexible pressure sensor is provided with a first pressure sensor group and a second pressure sensor group. The first pressure sensor group includes a first pressure detection area 71, a second pressure detection area 72 and a third pressure detection area 73 arranged longitudinally in sequence. The second pressure sensor group includes a fourth pressure detection area 74, a fifth pressure detection area 75 and a sixth pressure detection area 76 arranged longitudinally in sequence. Each pressure detection area includes multiple pressure detection points arranged laterally in sequence. During the first pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the first pressure detection area 71, the second pressure detection area 72, and the third pressure detection area 73, respectively. During the second pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the fourth pressure detection area 74, the fifth pressure detection area 75, and the sixth pressure detection area 76, respectively.

[0104] In this embodiment of the invention, by arranging two first airbags 5 symmetrically positioned at the top of the accommodating cavity 2, and connecting them to a flexible pressure sensor, the system can more comprehensively cover the area above the user's wrist. This design helps to capture pressure changes in different parts of the wrist, thereby more accurately reflecting the detailed characteristics of the pulse.

[0105] The flexible pressure sensor features a first and second pressure sensor group, each containing multiple longitudinally arranged pressure detection areas, and each area further includes multiple laterally arranged pressure detection points. This layout allows the system to capture pressure changes at multiple points on the wrist, further improving the accuracy and comprehensiveness of the measurement. Due to its flexible design, the sensor adapts to users with different wrist shapes and sizes, ensuring close contact throughout the measurement process. This adaptability helps reduce measurement errors caused by differences in wrist shape. By setting multiple pressure detection areas and points, the system can adjust the measurement range and accuracy as needed. For example, when more refined pulse characteristics are required, more or denser detection points can be used; when broader pulse information is needed, data from multiple detection areas can be combined. The use of the flexible pressure sensor makes the measurement process more comfortable because it conforms to the curve of the wrist, reducing discomfort caused by contact with hard objects.

[0106] Because the system can capture pressure changes at multiple points, it provides a wealth of information for subsequent data analysis. For example, by analyzing data from different detection areas and points, the pulse characteristics of different parts of the wrist can be assessed, thus providing a more comprehensive basis for traditional Chinese medicine diagnosis.

[0107] Utilizing a flexible sensor, the design ensures the sensor fits as closely as possible to the skin, increasing effective contact points. The average adult wrist width is 5.5cm-7.5cm. Simulating pulse diagnosis, the middle position is selected as the Guan pulse, placed against the inner side of the high bone on the back of the palm, simulating the width of three fingers (approximately 5cm) of an average adult. A three-segment sensor (cun, guan, chi) is designed, each segment measuring 28.5mm x 7mm. Each segment employs 16 contact points, representing the maximum number of contact points possible within the current wrist width.

[0108] In this embodiment of the invention, a three-layered airbag is designed at the bottom. An air pump is used to inflate the airbags, and when the wrist is raised, it can fully contact the sensor above. A split double-layered airbag is designed above the sensor to solve two problems: the sensor can move downwards to a certain distance, allowing the wrist to be fully clamped; the airbags on both sides allow the sensor to fit against the radial artery of the wrist, ensuring that the pulse position is fully aligned with the sensor.

[0109] More preferably, the pulse monitoring system further includes a backend server, a communication module, and a camera module. The camera module is used to transmit the monitored user facial expressions and camera attributes to the backend server through the communication module. The camera attributes include camera number information and camera timestamp. The control module 8 or the backend server is used to execute the pulse analysis method based on the pulse diagnostic instrument as described in any one of the first aspects of the present invention. The backend server is used to determine the corresponding pulse diagnostic instrument number based on the camera attributes and the camera-pulse diagnostic instrument mapping table. The backend server is also used to recognize the user's facial expressions to determine the user's current emotional state, and update the user's pulse data based on the current emotional state. The control module 8 is used to control the first air pump 3 and the second air pump 4 to work according to the set three-level control mode, and to work according to the first control parameter, the second control parameter and the third control parameter respectively. And / or, the flexible pressure sensor includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure detection points arranged laterally in sequence. The length of the first pressure detection area 71, the second pressure detection area 72, the third pressure detection area 73, the fourth pressure detection area 74, the fifth pressure detection area 75 and the sixth pressure detection area 76 is any value between 26mm and 30mm, and the width is any value between 6mm and 8mm.

[0110] The solution in this invention, by introducing a backend server, a communication module, and a camera module, enables the system to achieve remote monitoring and data transmission. This not only improves the convenience of monitoring but also makes data analysis and processing more intelligent.

[0111] The backend server can quickly and accurately determine the corresponding pulse diagnostic device number based on camera attributes (such as camera ID and timestamp) and the camera-pulse diagnostic device mapping table, thereby achieving precise association and management of user pulse data. The camera module can capture the user's facial expressions and perform emotion recognition through the backend server. This function allows the system to comprehensively consider the impact of the user's emotional state on pulse data, thus providing more personalized health analysis and suggestions. Based on the user's current emotional state, the system can update the pulse data to reflect the potential impact of emotional changes on the pulse. This helps doctors or users gain a more comprehensive understanding of their own health status.

[0112] The control module 8 of this embodiment provides a three-level control mode, allowing different control parameters (such as the first control parameter, the second control parameter, and the third control parameter) to control the operation of the first air pump 3 and the second air pump 4 according to actual needs. This design improves the flexibility and adaptability of the system, meeting the needs of different users or different application scenarios. The flexible pressure sensor contains 96 pressure monitoring points, with each pressure detection area including 16 laterally arranged pressure detection points. This refined layout enables the system to capture more subtle pulse changes, improving the accuracy and sensitivity of the measurement. At the same time, the size design of each pressure detection area is reasonable (length from 26mm to 30mm, width from 6mm to 8mm), ensuring both the measurement coverage area and avoiding measurement inaccuracies caused by excessive size.

[0113] The pulse diagnosis monitoring system described above integrates a backend server, communication module, and camera module to achieve remote monitoring, intelligent data processing, and comprehensive analysis of emotional state and pulse data. Simultaneously, the system provides a flexible three-level control mode and a refined layout of pressure monitoring points, further improving the accuracy and adaptability of measurements. These enhanced and expanded technological effects enable the pulse diagnosis monitoring system to better serve the fields of traditional Chinese medicine diagnosis and health analysis.

[0114] In this embodiment of the invention, the pressure detection component on the pulse diagnostic instrument can accurately capture the pressure change signal of the user's pulse within a set time, and then process it to obtain the raw pulse waveform signal. This ensures the accuracy and integrity of the pulse data, providing a reliable foundation for subsequent analysis. By inputting the user's pulse image into a pre-built standard pulse image model, the system can automatically determine the user's pulse type, realizing intelligent and automated pulse image analysis.

[0115] The pulse analysis method, system, electronic device, and storage medium based on a pulse diagnostic instrument disclosed in the embodiments of the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A pulse analysis method based on a pulse diagnostic instrument, characterized in that, include: The pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period is acquired, and the pressure change signal within the set time period is processed to obtain the original pulse waveform signal associated with the user. The original pulse waveform signal is decomposed according to the selected wavelet basis function and decomposition scale parameter to obtain the first pulse frequency signal and the second pulse frequency signal under different decomposition scales; wherein, the first pulse frequency signal is used to characterize the high-frequency detail signal, the high-frequency detail signal includes high-frequency white noise and power frequency noise, and the second pulse frequency noise is used to characterize the low-frequency approximation signal. The first pulse frequency signal is subjected to high-frequency denoising processing according to the set noise threshold to remove high-frequency white noise and power frequency noise in the high-frequency detail signal; all wavelet coefficients associated with the high-frequency detail signal are obtained, all wavelet coefficients are traversed and it is determined that the absolute value of the wavelet coefficient at a certain point is greater than the absolute value of its adjacent wavelet coefficients, and this is taken as the signal change point. The pulse cycle characteristic value is determined based on all signal abrupt change points. The user pulse data within a set time range is determined based on the pulse cycle characteristic value. The corresponding user pulse image is generated based on the user pulse data within the set time range. The obtained user pulse image is input into a pre-built standard pulse image model to determine the user's pulse image type result, and the user's pulse image type result is output.

2. The pulse analysis method based on a pulse diagnostic instrument as described in claim 1, characterized in that, The process of generating a corresponding user pulse map based on user pulse data within a set time range includes: Acquire user pulse data within a set time range, and construct a cubic polynomial equation for each user pulse data, wherein the cubic polynomial equation is the same at each data point and the first and second derivatives are continuous between adjacent intervals. The obtained cubic polynomial equation is calculated to obtain the parameter values ​​of each coefficient in the cubic polynomial equation. Based on the calculated parameter values ​​of each coefficient in the cubic polynomial equation, a pulse fitting curve is constructed on the interval of each data point. Based on the error values ​​between the user's pulse data and the pulse fitting curve, if all error values ​​are within the error threshold, the pulse fitting curve is determined as the user's pulse map.

3. The pulse analysis method based on a pulse diagnostic instrument as described in claim 1, characterized in that, After obtaining all the wavelet coefficients associated with the high-frequency detail signal, the process further includes: All wavelet coefficients are filtered according to a pre-set extreme value filtering threshold. The filtered wavelet coefficients are then traversed to determine the absolute value of a wavelet coefficient at a certain point that is greater than the absolute value of its adjacent wavelet coefficients. And / or, receive detection attribute information input by the user, the detection attribute information including user gender, detection time and detection season.

4. The pulse analysis method based on a pulse diagnostic instrument as described in claim 1, characterized in that, The pressure detection component includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure monitoring points arranged horizontally in sequence; The pressure change signal detected by the pressure detection component at the pulse diagnostic instrument within a set time period includes: By setting the pressure detection grid on the pulse diagnostic instrument, surface pressure distribution data after the user's wrist is placed is collected synchronously. The pressure gradient field of the entire pressure detection grid area is determined based on the surface pressure distribution data. Identify characteristic regions in the pressure gradient field where pressure gradient vectors converge or change. Determine an optimal initial measurement point that meets preset conditions from the feature region.

5. The pulse analysis method based on a pulse diagnostic instrument as described in claim 4, characterized in that, The logic for determining the optimal initial measurement point includes: Calculate the pressure gradient parameter for each grid point and identify the set of candidate points whose pressure gradient parameters exceed a first threshold; In the candidate point set, find the point with the minimum divergence of the pressure gradient direction in its surrounding local region; The point with the smallest divergence in the direction of the pressure gradient is determined as the optimal initial measurement point; After determining the point with the minimum divergence in the pressure gradient direction as the optimal initial measurement point, the method further includes: The detection component with vertical pressurization capability is controlled to apply pressure at different pressure levels in the vertical direction and acquire the corresponding pulse waveform signals; Based on the signal quality assessment algorithm, the pulse waveforms acquired at different depths are scored; The point corresponding to the highest signal quality score is determined as the final accurate pulse measurement point.

6. The pulse analysis method based on a pulse diagnostic instrument as described in claim 1, characterized in that, Following the output of the user's pulse type result, the following step is also included: Retrieve a general question library associated with pulse type; each question in the general question library has pre-configured attribute information; Based on the pulse type results, a targeted set of questions is selected from the general question library and generated, and then presented to the user. Receive user feedback answers to the set of questions; The feedback answer and the pulse type result are fused to generate a comprehensive diagnostic report; And / or, following the output user's pulse type result, it also includes: A moxibustion therapy sequence is generated based on the pulse type results; wherein, the moxibustion therapy sequence includes at least: one or more target acupoints, and a suggested therapy time corresponding to each acupoint; The generated moxibustion treatment sequence is presented to the user in the form of a visual chart, which includes the location of acupoints, the order of moxibustion, and the treatment time for each acupoint.

7. The pulse analysis method based on a pulse diagnostic instrument as described in claim 1, characterized in that, The pulse analysis method further includes: Acquire user facial images at various time points captured by the camera; User facial images at each time detection point are input into a pre-built facial analysis model for identification to determine the user's emotion type at each time detection point; wherein, the facial analysis model is constructed using a convolutional neural network; The user's emotion type at each time point is input into a waveform modification model based on the relationship between emotion and pulse to optimize pulse parameters and correct the resulting user pulse map.

8. A pulse diagnosis monitoring system, characterized in that, include: A housing assembly having a receiving cavity for accommodating the wrist to be measured. A first inflation assembly, comprising a first air pump and a first airbag, wherein the first airbag is disposed at the top of the accommodating cavity, and the first air pump is used to inflate the first airbag; The second inflation assembly includes a second air pump and a second airbag. The second airbag is disposed at the bottom of the accommodating cavity, and the second air pump is used to inflate the first airbag. A pressure detection component is disposed on the first airbag and is used to detect the pressure on the wrist to be measured placed in the accommodating cavity. The control module is electrically connected to the first air pump, the second air pump, and the pressure detection component. The control module is used to control the first air pump and the second air pump to work according to the received measurement control signal so that the first airbag and the second airbag inflate, thereby enabling the pressure detection component located at the first airbag to detect the pulse of the user's wrist.

9. The pulse diagnosis monitoring system as described in claim 8, characterized in that, The number of first airbags is two, and the two first airbags are symmetrically arranged on the top of the accommodating cavity; the pressure detection component is a flexible pressure sensor, and the two ends of the flexible pressure sensor are respectively connected to the two first airbags; The flexible pressure sensor is provided with a first pressure sensor group and a second pressure sensor group. The first pressure sensor group includes a first pressure detection area, a second pressure detection area and a third pressure detection area arranged longitudinally in sequence. The second pressure sensor group includes a fourth pressure detection area, a fifth pressure detection area and a sixth pressure detection area arranged longitudinally in sequence. Each pressure detection area includes multiple pressure detection points arranged laterally in sequence. During the first pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the first pressure detection area, the second pressure detection area, and the third pressure detection area, respectively. During the second pressure sensor detection process, pressure changes are detected at corresponding positions on the user's wrist through the fourth pressure detection area, the fifth pressure detection area, and the sixth pressure detection area, respectively.

10. The pulse diagnosis monitoring system as described in claim 8, characterized in that, The pulse monitoring system further includes a backend server, a communication module, and a camera module. The camera module is used to transmit the monitored user facial expressions and camera attributes to the backend server through the communication module. The camera attributes include camera ID information and camera timestamp. The control module or the backend server is used to execute the pulse analysis method based on the pulse diagnostic instrument as described in any one of claims 1 to 7. The backend server is used to determine the corresponding pulse diagnostic instrument number based on the camera attributes and the camera-pulse diagnostic instrument mapping table. The backend server is also used to recognize the user's facial expressions to determine the user's current emotional state, and update the user's pulse data based on the current emotional state. The control module is used to control the first air pump and the second air pump to work according to the set three-level control mode, and to work according to the first control parameter, the second control parameter and the third control parameter respectively. And / or, the flexible pressure sensor includes 96 pressure monitoring points, and each pressure detection area includes 16 pressure detection points arranged laterally in sequence. The length of the first pressure detection area, the second pressure detection area, the third pressure detection area, the fourth pressure detection area, the fifth pressure detection area and the sixth pressure detection area is any value between 26mm and 30mm, and the width is any value between 6mm and 8mm.

Citation Information

Patent Citations

  • Automatic three-portion and nine-pulse-taking information acquisition and recognition device and method of MEMS hydraulic transmission touch

    CN103330550A

  • Quick pulse condition collecting device and control method

    CN103845043A

  • Pulse condition recognition method based on double-channel convolutional neural network fusion

    CN112487945A

  • Pulse feeling device

    CN112674733A

  • Pulse strength judgment method and device, electronic equipment and storage medium

    CN115251856A