Identification processing method for flood pulse waves

Abstract

The invention relates to a flood pulse wave recognition method based on a signal processing technology, which comprises the following steps of: collecting pulse wave original data of a patient, including pulse condition information of three parts, namely cun, guan and chi; performing periodic feature extraction and analysis on the data, and dividing the pulse wave into a plurality of periodic signals; performing low-pass filtering processing on the periodic signal data by using a low-pass filtering technology so as to eliminate noise interference and highlight flood pulse characteristics; in the data identification stage, key features including but not limited to dominant wave rise time, dominant wave amplitude, waveform rising branch steep degree, waveform area, pulse wave fullness and the like are extracted from filtered signals, and feature matching is carried out; and the extracted features are compared with a pre-constructed mathematical model, and when the features meet specific conditions of the flood pulse, the condition of the flood pulse can be confirmed to be the condition of the flood pulse. According to the method, the accuracy and efficiency of flood pulse condition recognition are improved, the dependence on subjective experience of doctors is reduced, and a feasible scheme is provided for modern diagnosis and treatment of traditional Chinese medicine.

Description

Technical Field

[0001] The present invention relates to the field of signal processing, and in particular to a method for identifying and processing flood pulse waves. Background Art

[0002] In the practice of pulse diagnosis in traditional Chinese medicine, the Hong pulse, as an important type of pulse, is highly valued for its unique pathological and physiological significance. The typical characteristics of the Hong pulse are that the pulse is large, strong and powerful, full when it comes and fast when it goes, which is figuratively compared to the surging of rivers, lakes and seas. In the clinical diagnosis of traditional Chinese medicine, the Hong pulse is often regarded as a direct reflection of the pathological state of the body, such as hyperactivity of yang heat, vigorous qi and blood, or the existence of evil and the absence of deficiency of the body. It is closely related to the occurrence and development of a series of diseases, such as fever and excess syndrome, so the accurate identification and in-depth analysis of the Hong pulse is crucial.

[0003] However, current pulse diagnosis technology faces many challenges in identifying the surging pulse. Traditional pulse diagnosis methods rely heavily on the physician's experience and individual perception, lacking unified, objective quantitative indicators and standardized operating procedures, which undoubtedly increases the subjectivity and uncertainty of surging pulse identification. In view of this, the existing pulse diagnosis technology urgently needs to be improved and perfected, so as to capture and analyze surging pulse signals in a more accurate and reproducible way, thereby improving the objectivity and accuracy of traditional Chinese medicine pulse diagnosis and better serving clinical practice and scientific research.

[0004] In recent years, with the rapid progress of science and technology, especially the innovation of new materials and electronic computer technology, these technologies have been increasingly widely used in the medical field. In particular, a series of remarkable achievements have been made in the acquisition, processing and visualization of pulse information, laying a solid foundation for the theoretical research of pulse signals. Breakthroughs in new material technology have greatly improved the accuracy and efficiency of pulse diagnosis equipment, so that patients' pulse data can be collected more accurately. By using advanced signal processing algorithms, key feature information of pulse can be extracted from these data. At the same time, the development of electronic computer technology has provided powerful tools for big data analysis and machine learning, making it possible to deeply understand and automatically identify pulse. In the paper "Analysis and Recognition of Pulse Signals Based on Residual Networks and Attention Mechanisms", Zhu Jiajian et al. used an improved one-dimensional residual network to extract features from pulse signals and combined with an improved attention mechanism to improve the classification performance of the network, which is conducive to its deployment on actual pulse diagnosis equipment and its application in real pulse detection. In the article "Traditional Chinese Medicine Pulse Recognition Method Based on Extreme Learning Machine", Chen Xingchi et al. proposed a pulse recognition method based on extreme learning machine (ELM) in view of the strong fuzziness, wide variety and complex characteristics of traditional Chinese medicine pulse, as well as the shortcomings of traditional fuzzy clustering method and BP neural network recognition method. This method extracts the feature vector of pulse signal and then uses ELM to train and classify the feature vector. Summary of the invention

[0005] The present invention provides a method for identifying and processing a hong pulse wave, which is dedicated to significantly reducing the diagnosis time by accurately identifying the characteristics of the hong pulse, and effectively coping with the contradiction between large-scale data sets and high-performance computing requirements. The characteristics of the hong pulse are characterized by a wide, fluent and grand pulse wave, like a rolling wave, with a significantly full pulse position, a strong pulsation force and a rapid castration.

[0006] A method for identifying and processing a flood pulse wave comprises the following steps: S1. Using a pulse diagnosis device to collect raw pulse wave data of the subject to be tested, including but not limited to changes in pulse intensity at each pulse position within a certain period of time; S2. extracting the period of the collected raw pulse wave data and analyzing the periodic characteristics of the pulse wave; S3. Filtering the periodic signal data, preferably using a low-pass filter to reduce noise interference and highlight the characteristics of the Hong pulse wave; S4. Perform data recognition steps on the filtered signal, including but not limited to extracting effective features unique to the Hong pulse wave, such as the main wave rise time, main wave amplitude, main wave fall speed and waveform structure features unique to the Hong pulse, and performing feature matching.

[0007] Furthermore, in step S1, it is further clarified that the pulse wave data collected by the pulse pillow device covers the pulse beating conditions of the three parts of Cun, Guan and Chi.

[0008] Furthermore, in the feature identification of step S4, effective features include but are not limited to main wave width, main wave height, tidal wave height, dicrotic wave height, descending gorge height, main peak angle and ascending limb angle.

[0009] Furthermore, feature matching also includes: The pulse wave signal's pulsation starting point (X 1 ,Y 1 ), the main wave peak (X 2 ,Y 2 ), the main wave end point (X 3 ,Y 3 ) and the end point of the beat (X 4 ,Y 4 ); S4.1 calculates the slope k of the ascending branch of the retained pulse wave signal, , remove the signal with k<11.43; S4.2 The effective features include but are not limited to the time of the main wave rising branch, that is, the time value from the starting point of the pulse wave graph to the peak of the main wave, recorded as t 1 =X 2 -X 1 , requiring 0.07s≤t 1 ≤0.11s, and the main wave height, that is, the height from the peak of the main wave to the pulse wave baseline, is recorded as h, and h>10mm; S4.3 Calculate the depth of the descending branch and the depth of the trough of the descending middle gorge, that is, the height of the descending middle gorge. The ratio of the height of the descending middle gorge to the height of the main wave must satisfy Y 3 / Y 2 <0.3; S4.4 Calculate the distance L from the end point of the main wave to the ascending branch of the pulse wave in the retained pulse wave signal 1 , , the distance L from the end point of the main wave to the peak point of the main wave 2 , , the effective features also include the main wave shape parameters, where the angle formed by the line connecting the main wave vertex and the starting point is recorded as the main wave angle θ, , remove the pulse signals that do not satisfy θ>20°.

[0010] Furthermore, clear pulse signals can be obtained under three different pulse taking intensities: floating, middle and deep.

[0011] Furthermore, in the filtering processing stage, a Butterworth filter is specifically used as a low-pass filtering tool, and its parameter configuration is designed to ensure that the high-frequency noise components in the pulse wave signal are effectively filtered out, while keeping the core characteristics of the Hong pulse or other specific pulse patterns such as the Hong pulse wave intact; this process ensures that the signal after filtering can fully display the typical characteristics of the target pulse pattern such as the Hong pulse wave, so that in the subsequent data recognition stage, by comparing and matching the extracted effective features with the pre-constructed mathematical model, the pulse wave morphology represented by the filtered signal can be accurately determined, thereby achieving accurate identification of the Hong pulse wave or other pulse types.

[0012] Furthermore, the waveform is required to have floating pulse characteristics, that is, the angle of the rising part of the main wave must be at least 84°, and there must be no weak sub-wave phenomenon in the rising stage; in addition, the ratio of the height of the descending gorge to the height of the main wave must be lower than 0.3. Only when all these specific geometric and morphological characteristics are met can it be concluded that the pulse wave morphology meets the characteristic description of the flood pulse wave.

[0013] In the pulse diagnosis technology of the present invention, the pulse diagnosis device successfully captures the original pulse data from the three key positions of the subject, namely, Cun, Guan and Chi, which fully reflect the comprehensive information of the pulse activity. In order to deepen the insight and analysis of these initial data, the present invention uses the period extraction technology to reasonably divide the continuous pulse wave original data into multiple periodic segments.

[0014] In order to further optimize the data quality and highlight the core characteristics of the Hong pulse wave, the present invention chooses to use low-pass filtering technology for data processing. This method aims to effectively filter out the high-frequency noise components in the pulse wave signal and ensure that the low-frequency part, that is, the part closely related to the Hong pulse wave morphology, is preserved to the greatest extent possible.

[0015] Compared with the deep learning method commonly used today, the proposed method for identifying the Hong pulse wave is simple and efficient. Although the identification process is simplified, it can still maintain a relatively high accuracy, which means that even while reducing the complexity of the algorithm, the identification effect can still be ensured to be accurate. The advantage of this method is that it is more convenient to implement and has lower economic costs, thus showing stronger feasibility and practicality in practical applications.

[0016] In summary, the Hong pulse wave identification technology introduced in this study is very likely to be widely used and promoted in future pulse diagnosis practice and research fields due to its simple operation, high accuracy and low cost investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to ensure that readers have an accurate understanding of the content of the present invention, this article will discuss and explain it in more detail based on the specific embodiments and drawings of the present invention, in which: Figure 1 It is a schematic diagram of a single cycle of the Hongmai pulse condition of the present invention; Figure 2 It is the pulse signal identification flow chart of the present invention as a whole; Figure 3 It is a schematic diagram of a normal pulse condition identified by the present invention; Figure 4 This is a pulse diagram that does not belong to the Hong pulse wave. DETAILED DESCRIPTION

[0018] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments so that the present invention can be better understood and implemented by those skilled in the art.

[0019] Embodiment 1: The present invention provides a method for identifying a Hongmai pulse condition, which uses a custom-designed pulse condition sensing device to accurately capture the Hongmai characteristic signals of the three pulse positions of Cun, Guan, and Chi. Through signal processing, including pre-processing means such as period refinement, filtering and denoising, the original data of the Hongmai pulse condition is effectively separated and optimized. Finally, a mathematical model is used to deeply mine the pre-processed pulse signal, extract effective features unique to the Hongmai such as the main wave rising speed, peak amplitude, pulse wave duration, etc., and perform detailed feature matching analysis, thereby achieving accurate identification and classification of the Hongmai pulse condition.

[0020] Reference Figure 1 As shown, the filtered signal should include a main wave and a dicrotic wave, and in the Hongmai pulse diagram provided in this embodiment, the amplitude of the main wave fluctuates more.

[0021] like Figure 2 The figure shows a flow chart of pulse recognition, which includes the following steps.

[0022] First, use pulse diagnosis equipment such as a pulse diagnostic instrument to collect raw pulse wave data containing Hongmai characteristics at the Cun, Guan, and Chi positions of the subject. The sampling period at each position is fixed to ensure that complete dynamic change information of the Hongmai pulse in different time periods is obtained.

[0023] Subsequently, the collected raw data of Hongmai pulse wave are periodically processed through periodic feature extraction technology, and decomposed into multiple independent periodic signal units, so as to facilitate in-depth exploration of the intrinsic periodic law of the pulse wave.

[0024] In the data preprocessing stage, with the help of low-pass filtering technology, especially selecting Butterworth filter as the filtering tool, low-pass filtering operation is performed on the periodic signal data, which effectively filters out the high-frequency noise in the pulse wave signal, highlights the core characteristics of the pulse wave, and ensures the high quality and accuracy of the signal.

[0025] Entering the feature recognition stage, the key features of the filtered signal are extracted, including but not limited to the main wave rise time, main wave amplitude, main wave fall rate and the waveform structure attributes of the Hong pulse wave, such as the main wave width, main wave height, descending gorge height, main peak angle and ascending branch angle. At the same time, feature matching also includes the precise identification of the pulse wave signal's pulsation start point, main wave peak point, main wave end point and pulsation end point.

[0026] In order to illustrate the specific process of matching and identifying the characteristics of the flood pulse wave, refer to Figure 1 , mark the pulsation starting point, main wave peak point, main wave end point and pulsation end point of the Hongmai pulse wave signal as coordinates (X 1 ,Y 1 ), (X 2 ,Y 2 ), (X 3 ,Y 3 ), (X 4 ,Y 4 ).

[0027] First, calculate the slope k of the retained flood pulse signal, which can be expressed as , and remove the pulse wave signals whose slope k does not reach the normal threshold, that is, the signals with k < 11.43, to ensure the significance of the flood pulse wave characteristics.

[0028] The time from the starting point of the pulse wave to the peak of the main wave is recorded as t 1 =X 2 -X 1 , requiring 0.07s≤t 1 ≤0.11s. At the same time, the main wave height, that is, the height from the peak of the main wave to the pulse wave baseline, is recorded as h, and h>10mm is required.

[0029] Calculate the distance L from the end point of the main wave to the ascending branch of the pulse wave in the retained pulse wave signal 1 , the distance L from the end point of the main wave to the peak point of the main wave 2 ,in, , , the effective features also include the main wave shape parameters, where the angle formed by the line connecting the main wave vertex and the starting point is recorded as the main wave angle θ, , remove the pulse signals that do not satisfy θ>20°.

[0030] Next, the peak height Y of the main wave is determined according to the A / D conversion circuit parameters of the pulse diagnosis instrument used. 2 and the height of the gorge Y 3 If Y 3 / Y 2 <0.3, then the pulse wave signals that do not conform to the characteristics of the large pulse wave will be removed.

[0031] After the above identification steps and calculations, the identified pulse data must show the typical characteristics of a surging pulse, that is, the angle of the rising part of the main wave must be at least 84°, and there must be no weak sub-wave phenomenon in the rising stage. In addition, it is emphasized that the ratio of the height of the descending gorge to the height of the main wave must be less than 0.3. The pulse wave shape that meets these strict conditions is determined to be a surging pulse wave.

[0032] Embodiment 2: This embodiment will introduce in detail how to use the surging pulse wave recognition method provided by the present invention to accurately recognize the actual surging pulse condition.

[0033] First, the pulse diagnosis device is used to collect the raw data of the pulse wave of the subject to be tested. These data record the characteristics of the three pulses of Cun, Guan and Chi in detail. The collected data is periodically extracted to divide the pulse wave into multiple signal segments with periodic characteristics, which is the basic step to identify the pulse wave.

[0034] Next, the periodic signal data is low-pass filtered, and the Butterworth filter is used to suppress high-frequency noise, so that the key features of the flood pulse wave can be revealed. Then, the feature extraction stage is entered to calculate and analyze the key features.

[0035] Calculate the slope of the rising branch of the main wave to determine the steepness of the rising branch, and remove signals with a slope less than 11.43.

[0036] According to the main wave peak height Y 2 and the height of the gorge Y 3 The ratio Y 3 / Y 2 , calculate and filter out Y 3 / Y 2 Then the main wave angle is calculated through the pulsation starting point, main wave peak point and main wave ending point, and the signal with an angle greater than 20° is selected.

[0037] Finally, it is determined that the angle of the main wave rising branch is greater than 84°, and there is no small wave in the rising branch. The ratio of the height of the descending gorge to the height of the main wave is less than 0.3. The waveform has a deep trough of the descending gorge, and the peak of the dicrotic wave is obvious. The peak becomes higher and sharper, and its net height is ≥1 / 4 of the main wave height. We get the desired Hong pulse waveform. The specific waveform obtained by identification is as follows Figure 3 shown.

[0038] Embodiment 3: In this embodiment, we will demonstrate how to apply the pulse recognition method of the present invention to identify and exclude pulses that do not conform to the typical characteristics of a flood pulse wave.

[0039] First, the pulse diagnosis equipment is still used to collect the original pulse wave data of the object to be tested at the three positions of Cun, Guan and Chi, and the pulse wave is also divided into multiple periodic signal segments through the period extraction process.

[0040] After low-pass filtering the periodic signal, focus on those pulses that do not meet the key characteristics of the flood pulse wave. For such signals, perform the following identification steps: Calculate the slope of the rising branch of the main wave. If the slope is greater than or equal to 11.43, or if the slope is small but not enough to reflect the steep rising characteristics unique to flood pulse waves, it is preliminarily identified as a non-flood pulse wave signal.

[0041] Check the ratio Y of the main wave peak height to the mid-fall height 3 / Y 2 If the ratio is higher than 0.3, it does not meet the characteristics of the low-lying mid-gorge of the flood pulse wave, and is further excluded as a flood pulse wave, such as Figure 4 shown.

[0042] Calculate the main wave angle. If the angle is less than 20° or does not have the open rising trend that a flood pulse wave should have, it cannot be classified as a typical flood pulse wave.

[0043] Furthermore, if the angle of the rising branch of the main wave in the pulse does not reach above 84°, or if there is a particularly obvious small wave in the rising branch, this pulse cannot be identified as a surging pulse wave.

[0044] In this embodiment, we successfully identify and output pulse signals that do not meet the characteristics of the flood pulse wave through a series of feature screening and matching, and prove the effectiveness of the method of the present invention in excluding atypical flood pulse waves through actual operation. The pulse recognition results of this part of atypical or non-flood pulse waves can be obtained through specific pulse diagrams (such as Figure 4 The patient's condition can be intuitively presented as shown in the figure to facilitate clinicians and other professionals to conduct further disease analysis and diagnostic decisions.

Claims

1. A method for identifying and processing a flood pulse wave, characterized in that: The steps include: S1. Use the pulse diagnosis device to collect the original pulse wave data of the subject to be tested, including but not limited to the changes in the pulse intensity of each pulse position within a certain period of time; S2. extracting the period of the collected raw pulse wave data and analyzing the periodic characteristics of the pulse wave; S3. Filter the periodic signal data, preferably using a low-pass filter to reduce noise interference and highlight the characteristics of the Hong pulse wave; S4. Perform data recognition steps on the filtered signal, including but not limited to extracting effective features unique to the Hong pulse wave, such as the main wave rise time, main wave amplitude, main wave fall speed and waveform structure features unique to the Hong pulse, and performing feature matching.

2. The method for identifying and processing a flood pulse wave according to claim 1, characterized in that: In step S1, it is further clarified that the pulse wave data collected by the pulse pillow device covers the pulse beating conditions of the three parts of Cun, Guan and Chi.

3. The method for identifying and processing a flood pulse wave according to claim 1, characterized in that: In the feature identification of step S4, effective features include but are not limited to main wave width, main wave height, tidal wave height, dicrotic wave height, descending gorge height, main peak angle and ascending branch angle.

4. The method for identifying and processing a flood pulse wave according to claim 3, characterized in that: Feature matching also includes: The pulse wave signal's pulsation start point (X1, Y1), main wave peak point (X2, Y2), main wave end point (X3, Y3) and pulsation end point (X4, Y4); S4.1 calculates the slope k of the ascending branch of the retained pulse wave signal, , remove the signal with k<11.43; S4.2 The effective features include but are not limited to the time of the main wave rising branch, that is, the time value from the starting point of the pulse wave graph to the peak of the main wave, denoted as t1=X2-X1, requiring 0.07s≤t1≤0.11s, and the main wave height, that is, the height from the peak of the main wave to the baseline of the pulse wave graph, denoted as h, requiring h>10mm; S4.3 Calculate the depth of the descending branch and the depth of the trough of the descending middle gorge, that is, the height of the descending middle gorge, where the ratio of the height of the descending middle gorge to the height of the main wave must satisfy Y3 / Y2<0.3; S4.4 calculates the distance L1 from the end point of the main wave in the retained pulse wave signal to the ascending branch of the pulse wave, , the distance from the end point of the main wave to the peak point of the main wave is L2, , the effective features also include the main wave shape parameters, where the angle formed by the line connecting the main wave vertex and the starting point is recorded as the main wave angle θ, , remove the pulse signals that do not satisfy θ>20°.

5. The method for identifying and processing a flood pulse wave according to claim 1, characterized in that: Clear pulse signals can be obtained under three different pulse taking intensities: floating, middle and deep.

6. The method for identifying and processing a flood pulse wave according to claim 1, characterized in that: In the filtering processing stage, a Butterworth filter is specifically used as a low-pass filtering tool. Its parameter configuration is designed to ensure that the high-frequency noise components in the pulse wave signal are effectively filtered out, while keeping the core characteristics of the Hong pulse or other specific pulse patterns such as the Hong pulse wave intact; this process ensures that the signal after filtering can fully display the typical characteristics of the target pulse pattern such as the Hong pulse wave, so that in the subsequent data recognition stage, by comparing and matching the extracted effective features with the pre-constructed mathematical model, the pulse wave morphology represented by the filtered signal can be accurately determined, thereby achieving accurate identification of the Hong pulse wave or other pulse types.

7. The method for identifying and processing a flood pulse wave according to claim 1, characterized in that: The waveform is required to have floating pulse characteristics, that is, the angle of the rising part of the main wave must be at least 84°, and there must be no weak sub-wave phenomenon in the rising stage; in addition, the ratio of the height of the descending gorge to the height of the main wave must be lower than 0.

3. Only when all these specific geometric and morphological characteristics are met can it be concluded that the pulse wave morphology meets the characteristic description of a flood pulse wave.

Images