A pulse sampling method and an AI pulse diagnosis system for traditional Chinese medicine auxiliary diagnosis
By adaptively adjusting the spacing and position of the signal acquisition terminals, and combining signal feedback to adjust the acquisition height, the pulse waveform is acquired and filtered, solving the problem that the sensor cannot adapt to different patients and improving the accuracy of TCM diagnosis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- DR ZHONG (HEBEI) INTELLIGENT TECH CO LTD
- Filing Date
- 2026-06-05
- Publication Date
- 2026-07-14
AI Technical Summary
Current technology cannot effectively utilize sensors to sample the pulse of different patients, resulting in data acquisition that cannot use fixed parameters, thus affecting diagnostic accuracy.
By adaptively adjusting the spacing and position of the signal acquisition terminals and adjusting the acquisition height in conjunction with signal feedback, the waveforms of floating pulse, middle pulse, and deep pulse are acquired, and then filtered and aligned to obtain accurate pulse samples.
It enables adaptive pulse sampling based on the individual characteristics of different patients, improving the accuracy and reliability of TCM diagnosis.
Smart Images
Figure CN122392902A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data analysis technology, and in particular to a pulse sampling method and an AI pulse diagnosis system for auxiliary diagnosis in traditional Chinese medicine. Background Technology
[0002] The AI pulse diagnosis system aims to transform the subjective tactile experience of traditional Chinese medicine pulse diagnosis into objective, quantifiable, and repeatable digital signals. Its core logic is to "simulate" the three-finger pulse diagnosis of experienced Chinese medicine practitioners through modern sensing technology, and then use artificial intelligence algorithms for in-depth interpretation.
[0003] The specific solution involves using sensors to simulate the "three-finger placement" and "lifting, pressing, and searching" finger techniques, acquiring the three-dimensional waveforms of the cun, guan, and chi positions, quantifying the data using feature extraction, and finally using AI for feature extraction, comparison, and analysis to obtain the final result.
[0004] The advantage of using sensors for data acquisition is that it can quantify pulse data. However, different patients have different conditions, so it is not possible to use fixed parameters for data acquisition. Further research is needed to solve this problem. Summary of the Invention
[0005] This application provides a pulse sampling method and an AI pulse diagnosis system for auxiliary diagnosis in traditional Chinese medicine. The method uses an adaptive adjustment approach to sample pulses from different objects using different parameter sets, thereby obtaining more accurate pulse sampling data.
[0006] The above-mentioned objective of this application is achieved through the following technical solution:
[0007] Firstly, this application provides a pulse sampling method for auxiliary diagnosis in Traditional Chinese Medicine, including:
[0008] Obtain hand samples of the sampling object and calculate the converted size ratio based on the hand samples;
[0009] In response to the received position adjustment command, the spacing of the signal acquisition terminals is adjusted and the detection area is determined;
[0010] Adjust the position of the signal acquisition terminal in the detection area based on the signal feedback from the signal acquisition terminal in the detection area to determine the acquisition position of the signal acquisition terminal;
[0011] The drive signal acquisition terminal determines the acquisition height at the acquisition position, which includes the surface height, elastic height, and rigid height.
[0012] Signal acquisition is performed at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms.
[0013] Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
[0014] In one possible implementation of the first aspect, adjusting the position of the signal acquisition terminal in the detection area based on signal feedback from the signal acquisition terminal in the detection area includes:
[0015] Create a straight line trajectory along the length of the detection area;
[0016] Select collection points on the straight line of the movement trajectory. There are multiple collection points, and the collection points are set at intervals on the straight line of the movement trajectory.
[0017] Signals are acquired at the acquisition points to obtain the acquired waveforms;
[0018] Calculate the peak value of the acquired waveform and calculate the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform;
[0019] The sampling depths at all sampling points were the same.
[0020] In one possible implementation of the first aspect, calculating the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform includes:
[0021] Construct discrete points in a coordinate system. The horizontal coordinate of the discrete points is the position coordinate of the acquisition point, and the vertical coordinate of the discrete points is the wave peak value at the discrete points.
[0022] Construct numerical curves of wave crests using discrete points in a coordinate system;
[0023] Calculate the peak of the peak numerical curve and use the corresponding position of the peak numerical curve as the acquisition position of the signal acquisition terminal.
[0024] In one possible implementation of the first aspect, determining the acquisition height at the acquisition position by the drive signal acquisition end includes:
[0025] The drive signal acquisition end moves linearly at the acquisition position, synchronously recording the moving distance and deformation.
[0026] Create a movement curve using the movement distance as the horizontal axis and the deformation as the vertical axis.
[0027] Analyze and identify abrupt change points on the moving curve and record the corresponding moving distances at these points;
[0028] The sampling height range is determined by the movement distance corresponding to the abrupt change point, and there are multiple sampling height ranges.
[0029] Select a value as the acquisition height within each acquisition height range.
[0030] In one possible implementation of the first aspect, the number of height ranges collected is three;
[0031] The surface height is within the first acquisition height range, the elastic height is within the second acquisition height range, and the rigid height is within the third acquisition height range;
[0032] The first data collection showed a positive correlation between the distance traveled and the deformation within the height range.
[0033] The distance traveled and the deformation within the second acquisition height range are positively correlated, and the positive correlation coefficient between the distance traveled and the deformation within the second acquisition height range is greater than that between the distance traveled and the deformation within the first acquisition height range.
[0034] The third finding shows a linear correlation between the distance traveled and the deformation within the height range.
[0035] In one possible implementation of the first aspect, aligning the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform includes:
[0036] The floating pulse waveform, middle pulse waveform, and deep pulse waveform are filtered and grouped according to the acquisition location. The floating pulse waveform, middle pulse waveform, and deep pulse waveform in the same group are generated at the same acquisition location.
[0037] Calculate the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform respectively;
[0038] Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform according to their peak points.
[0039] Among them, the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are located at the same time point.
[0040] In one possible implementation of the first aspect, filtering the floating pulse waveform includes:
[0041] The floating pulse waveform is decomposed in the frequency domain to obtain a low-frequency decomposed waveform group and a high-frequency decomposed waveform group.
[0042] Calculate the residual between the waveform composed of a low-frequency decomposition waveform group and the floating pulse waveform;
[0043] Based on the residuals, the low-frequency decomposition waveform group and the high-frequency decomposition waveform group are re-divided, and the residuals of the waveforms composed of the low-frequency decomposition waveform group and the floating pulse waveform are calculated again until the residuals are less than or equal to the reference value.
[0044] The low-frequency decomposition waveform group is processed by grouping and residual processing to obtain the second low-frequency decomposition waveform group and the second high-frequency decomposition waveform group, and so on, to obtain the Nth low-frequency decomposition waveform group and the Nth high-frequency decomposition waveform group.
[0045] Filter the high-frequency decomposition waveform groups from the first to the Nth high-frequency decomposition waveform groups to remove high-frequency decomposition waveform groups containing noise.
[0046] The floating pulse waveform is formed by combining the low-frequency decomposition waveform group from one to N times with the remaining high-frequency decomposition waveform.
[0047] The filtering process for the middle pulse waveform and the deep pulse waveform is the same as that for the floating pulse waveform.
[0048] N is a natural number greater than or equal to 2.
[0049] Secondly, this application provides a pulse sampling device for auxiliary diagnosis in traditional Chinese medicine, comprising:
[0050] The data acquisition unit is used to acquire hand samples of the sampling object and calculate the converted size ratio based on the hand samples;
[0051] The first position adjustment unit is used to adjust the spacing of the signal acquisition terminals and determine the detection area in response to the received position adjustment command;
[0052] The second position adjustment unit is used to adjust the position of the signal acquisition terminal in the detection area according to the signal feedback from the signal acquisition terminal in the detection area, and to determine the acquisition position of the signal acquisition terminal.
[0053] The third position adjustment unit is used to drive the signal acquisition end to determine the acquisition height at the acquisition position. The acquisition height includes the surface height, the elastic height, and the rigid height.
[0054] The signal acquisition unit is used to acquire signals at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms.
[0055] The sample processing unit is used to align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
[0056] Thirdly, this application provides an AI pulse diagnosis system for auxiliary diagnosis in Traditional Chinese Medicine, the system comprising:
[0057] One or more memories for storing instructions; and
[0058] One or more processors are configured to call and execute the instructions from the memory to perform the methods described in the first aspect and any possible implementation thereof.
[0059] Fourthly, this application provides a computer-readable storage medium, the computer-readable storage medium comprising:
[0060] The program, when run by a processor, is executed as described in the first aspect and any possible implementation thereof.
[0061] Fifthly, this application provides a computer program product, including program instructions that, when run by a computing device, execute the method described in the first aspect and any possible implementation thereof.
[0062] Sixthly, this application provides a chip system including a processor for implementing the functions involved in the foregoing aspects, such as generating, receiving, transmitting, or processing the data and / or information involved in the foregoing methods.
[0063] This chip system can consist of chips or include chips and other discrete components.
[0064] In one possible design, the chip system also includes a memory for storing necessary program instructions and data. The processor and the memory can be decoupled and located on different devices, connected via wired or wireless means, or the processor and the memory can be coupled to the same device. Attached Figure Description
[0065] Figure 1 This is a flowchart illustrating the steps of a pulse sampling method for auxiliary diagnosis in Traditional Chinese Medicine, as provided in this application.
[0066] Figure 2 This is a schematic diagram of a body size provided in this application.
[0067] Figure 3 This is a schematic diagram of a method for determining the spacing between three signal acquisition terminals provided in this application.
[0068] Figure 4 This is a schematic diagram of the location of a detection area provided in this application.
[0069] Figure 5 This is a schematic diagram illustrating the correspondence between three sampling heights and the superficial, middle, and deep pulses, as provided in this application.
[0070] Figure 6 This is a schematic diagram of a process for obtaining the collection location provided in this application.
[0071] Figure 7 This is a schematic diagram illustrating the relationship between travel distance and deformation provided in this application.
[0072] Figure 8 This is a schematic diagram of the acquisition height within the acquisition height range provided in this application.
[0073] Figure 9 This is a schematic diagram of the process of successively decomposing the floating pulse waveform provided in this application.
[0074] Figure 10 This is a schematic diagram of a process for synthesizing a new floating pulse waveform provided in this application. The dashed lines in the diagram indicate that waveform groups are discarded. Detailed Implementation
[0075] The technical solutions in this application will be further described in detail below with reference to the accompanying drawings.
[0076] This application discloses a pulse sampling method for auxiliary diagnosis in Traditional Chinese Medicine. Please refer to [link / reference]. Figure 1 In some examples, the pulse sampling method for auxiliary diagnosis in traditional Chinese medicine disclosed in this application includes the following steps:
[0077] S101, Obtain a hand sample of the sampling object and calculate the converted size ratio based on the hand sample;
[0078] S102, in response to the received position adjustment command, adjusts the spacing of the signal acquisition terminals and determines the detection area;
[0079] S103, adjust the position of the signal acquisition terminal in the detection area according to the signal feedback of the signal acquisition terminal in the detection area, and determine the acquisition position of the signal acquisition terminal;
[0080] S104, the drive signal acquisition end determines the acquisition height at the acquisition position, the acquisition height includes the surface height, elastic height and rigid height;
[0081] S105, signal acquisition is performed at the acquisition height to obtain pulse signal waveforms, including floating pulse waveforms, middle pulse waveforms and deep pulse waveforms;
[0082] S106, align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
[0083] First, it should be noted that the pulse sampling method for auxiliary diagnosis in traditional Chinese medicine disclosed in this application is applied to an AI pulse diagnosis system for auxiliary diagnosis in traditional Chinese medicine. The AI pulse diagnosis system consists of a main body of equipment, sensors, actuators, and data processing units. The following detailed explanation of the method will further describe the AI pulse diagnosis system.
[0084] In step S101, a hand sample of the sampling object is first obtained and the converted size ratio is calculated based on the hand sample. Here, the concept of "body inch" needs to be introduced. "Body inch" refers to the unit of length for measuring acupoints or pulse diagnosis locations, which is determined by certain parts of the patient's body surface, rather than using a fixed metric ruler (such as a centimeter ruler). Its core logic is "different from person to person": "1 inch" is longer for tall people and shorter for short people.
[0085] In this application, hand samples of the sampling objects are used as the conversion basis for body measurements, and the width of four fingers held together is defined as 3 inches. Figure 2 As shown.
[0086] The total width of the three fingers (index, middle, and ring fingers) held together when taking a pulse is defined as 2 cun in traditional Chinese medicine theory. Based on the calculated conversion ratio, the specific length (in centimeters) of 2 cun can be determined.
[0087] At this point, the three signal acquisition terminals on the AI pulse diagnosis system can adjust the spacing. Specifically, if the sampled object is tall, its actual length of "2 inches" may reach 4-5 centimeters or even more; while if the sampled object is short, its "2 inches" may only be 2-3 centimeters.
[0088] In other words, the main purpose of step S101 is to determine the spacing between the three signal acquisition terminals, such as... Figure 3 As shown, S1 and S2 (which equal S1) need to be adjusted based on the 3 inches obtained from the above content.
[0089] In step S102, after receiving the position adjustment command, the AI pulse diagnosis system adjusts the spacing of the signal acquisition terminals and determines the detection area. Adjusting the spacing of the signal acquisition terminals here refers to adjusting the spacing according to the body measurements (step S101). After the spacing adjustment is completed, the detection area is determined in the direction perpendicular to the arm of the sampled object. The shape of the detection area is a rectangle, such as... Figure 4 As shown, in the horizontal direction, the signal acquisition end moves in a straight line within the signal detection area; in the vertical direction, the signal acquisition end moves towards the arm of the sampling object and away from the arm of the sampling object.
[0090] Of course, several specific issues also need to be addressed at this point, as follows:
[0091] To determine the position of the sampled object's arm, a suitable approach is for the AI pulse diagnosis system to provide a platform with an arc-shaped groove. This groove guides the sampled object to place its arm, and the image sensor on the AI pulse diagnosis system then determines the edge contour of the sampled object's arm. This method can determine the position of the sampled object's arm.
[0092] One way to determine the location of the radial styloid process is by manual contact by the operator. Another way is to add a pressure detection arm. The detection end of the pressure detection arm moves horizontally and contacts the outer side of the sampled object's arm. The location of the radial styloid process is determined by the pressure change.
[0093] In step S103, the position of the signal acquisition terminal in the detection area is adjusted based on the signal feedback from the signal acquisition terminal in the detection area to determine the acquisition position of the signal acquisition terminal. The acquisition position here refers to the position directly above the radial artery of the sampled object, because the signal obtained here is more accurate. The specific reasons are as follows:
[0094] Superficial location: This blood vessel is very close to the skin surface, covered only by skin, fascia and very thin muscle, without a thick layer of fat to block it, and its pulsation is easily felt;
[0095] Bone support: The radial artery is located directly below the radius bone. When the blood vessel pulsates, it is rigidly supported by the bone and will not spread to the surrounding area, making the pulsation clearer and stronger.
[0096] Fixed course: This segment of the blood vessel is relatively straight and fixed, not easy to slide, and easy to apply pressure stably.
[0097] In step S104, the signal acquisition terminal is driven to determine the acquisition height at the acquisition position. The acquisition height includes the surface height, elastic height, and rigid height, such as... Figure 5 As shown, surface height, elastic height, and rigidity height correspond to floating pulse, middle pulse, and deep pulse, respectively, or can be described as corresponding to light palpation, middle palpation, and deep palpation, respectively.
[0098] In step S105, signal acquisition is performed at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms. Finally, in step S106, the floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms are aligned to obtain a pulse sample.
[0099] The reason for alignment is that the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are obtained separately, but alignment is required during diagnosis to determine the corresponding changes of the floating pulse, the middle pulse, and the deep pulse within the same time period.
[0100] For example, a characteristic of a deep pulse is that it only contains the waveforms of a medium pulse and a deep pulse; a characteristic of a slow pulse is that it only contains the waveform of a medium pulse; and a characteristic of a weak pulse is that it only contains the waveform of a floating pulse. These examples demonstrate that only by aligning the waveforms of a floating pulse, a medium pulse, and a deep pulse can a pulse sample truly reflect the situation at a given sampling location.
[0101] In some examples, such as Figure 6As shown, the specific method for adjusting the position of the signal acquisition terminal in the detection area based on the signal feedback from the signal acquisition terminal in the detection area is as follows:
[0102] Create a straight line trajectory along the length of the detection area;
[0103] Select collection points on the straight line of the movement trajectory. There are multiple collection points, and the collection points are set at intervals on the straight line of the movement trajectory.
[0104] Signals are acquired at the acquisition points to obtain the acquired waveforms;
[0105] Calculate the peak value of the acquired waveform and calculate the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform;
[0106] The sampling depths at all sampling points were the same.
[0107] This method determines the acquisition location by comparing multiple intensity points. The radial artery pulsation intensity is different at different locations in the detection area. By quantifying the pulsation intensity and then using a fitting method, the peak intensity location can be obtained, which is the acquisition location of the signal acquisition end.
[0108] The specific method for calculating the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform is as follows:
[0109] Construct discrete points in a coordinate system. The horizontal coordinate of the discrete points is the position coordinate of the acquisition point, and the vertical coordinate of the discrete points is the wave peak value at the discrete points.
[0110] Construct numerical curves of wave crests using discrete points in a coordinate system;
[0111] Calculate the peak of the peak numerical curve and use the corresponding position of the peak numerical curve as the acquisition position of the signal acquisition terminal.
[0112] In the above method, after obtaining the acquisition waveform at the acquisition point, only the peak value (maximum intensity value) of the acquisition waveform is used. At this time, the position of the acquisition point is used as the abscissa and the peak value at the acquisition point position is used as the ordinate to obtain a set of discrete points. The curve obtained by fitting this set of discrete points has a change process of first rising and then falling. The maximum value corresponding to this curve is calculated, and the position corresponding to the maximum value is the acquisition position.
[0113] In some possible implementations, the curve obtained by fitting discrete points is performed using a smooth spline fitting method, as follows:
[0114] Define a mathematical objective function (also called a loss function) that needs to be minimized. This function consists of two parts:
[0115] Fitting error term: measures how close the curve is to the original discrete points (i.e., the sum of squared residuals).
[0116] Smoothing penalty term: measures the “curvature” or “roughness” of the curve (usually the integral of the curve’s second derivative).
[0117] Total cost = fitting error + λ × smoothing penalty. The value of λ is generally determined based on the noise level. When the data noise is large, the value of λ ranges from 0.01 to 0.05. When the data noise is small, the value of λ ranges from 0.1 to 1.
[0118] The entire horizontal coordinate interval is divided into many small segments, each represented by a cubic polynomial. At the same time, these piecewise polynomials are required to have equal function values at the connection points (called "nodes"), and their first derivative (slope) and second derivative (curvature) must also remain continuous.
[0119] cubic polynomial: y = a + bx + cx 2 +dx 3 Calculate a, b, c, and d in the cubic polynomial corresponding to each segment. There are multiple sets of a, b, c, and d. Here, it is required that the function values, the first derivative, and the second derivative are equal.
[0120] In some cases, the drive signal acquisition terminal determines the acquisition height at the acquisition location in the following way:
[0121] The drive signal acquisition end moves linearly at the acquisition position, synchronously recording the moving distance and deformation.
[0122] Create a movement curve using the movement distance as the horizontal axis and the deformation as the vertical axis.
[0123] Analyze and identify abrupt change points on the moving curve and record the corresponding moving distances at these points;
[0124] The sampling height range is determined by the movement distance corresponding to the abrupt change point, and there are multiple sampling height ranges.
[0125] Select a value as the acquisition height within each acquisition height range.
[0126] The above method determines the sampling height by measuring the distance traveled and the deformation, such as... Figure 7 As shown, specifically, there are three sampling height ranges, with one sampling height within each range. That is, there are three sampling heights: surface height, elastic height, and rigid height. The surface height is located within the first sampling height range, the elastic height is located within the second sampling height range, and the rigid height is located within the third sampling height range.
[0127] The specific way to use the movement distance and deformation is to create a movement curve by using the movement distance as the horizontal axis and the deformation as the vertical axis, then analyze and determine the abrupt change points on the movement curve and record the movement distance corresponding to the abrupt change points, and finally use the movement distance corresponding to the abrupt change points to determine the acquisition height range.
[0128] There are three cases regarding the distance the deformation travels, as detailed below:
[0129] The first data collection showed a positive correlation between the distance traveled and the deformation within the height range.
[0130] The distance traveled and the deformation within the second acquisition height range are positively correlated, and the positive correlation coefficient between the distance traveled and the deformation within the second acquisition height range is greater than that between the distance traveled and the deformation within the first acquisition height range.
[0131] The third finding shows a linear correlation between the distance traveled and the deformation within the height range.
[0132] To further explain this in the context of the wrist, the human tissue corresponding to the surface height (the first sampling height range) is the skin, subcutaneous superficial fascia, and superficial fat. Because the tissue is soft, it will deform significantly with slight force, so the correlation coefficient (slope) is relatively small.
[0133] At the elastic height (the second sampling height range), the corresponding human tissues are the deep fascia, muscle layer, and radial artery wall, where the tissue resistance begins to increase significantly. At this point, as the distance traveled continues to increase, the deformation still increases (positively correlated), but because the tissue has hardened, the deformation resistance generated by the same displacement is greater.
[0134] At the rigid height (the third acquisition height range), the corresponding human tissue is the radius (bone). At this point, the moving distance and deformation show a strict linear correlation (or the deformation hardly changes with the moving distance, and the slope tends to infinity or becomes a constant value), indicating that the probe has reached the bottom.
[0135] In the above content, the method for analyzing and determining the abrupt change points on the moving curve is to use the second derivative detection. The specific process is to first smooth the moving curve (smoothing spline fitting, Savitzky-Golay filtering, or moving average), and then calculate the first derivative, which represents the slope of the smoothed curve.
[0136] Next, calculate the second derivative. The x-coordinate corresponding to the peak (maximum) or trough (minimum) of the second derivative is the point of sudden change on the curve.
[0137] The above method yields two abrupt change points, which are used to divide the data into three acquisition height ranges. Within each range, a value is selected as the acquisition height, where (S3, S4, S5) is 0.7-0.8 times the corresponding acquisition height range (height). Figure 8 As shown.
[0138] In some examples, the specific methods for aligning the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are as follows:
[0139] The floating pulse waveform, middle pulse waveform, and deep pulse waveform are filtered and grouped according to the acquisition location. The floating pulse waveform, middle pulse waveform, and deep pulse waveform in the same group are generated at the same acquisition location.
[0140] Calculate the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform respectively;
[0141] Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform according to their peak points.
[0142] Among them, the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are located at the same time point.
[0143] The above method is obtained through peak points. If the peak points of multiple sets of floating pulse waveforms, middle pulse waveforms, and deep pulse waveforms can be directly aligned, then alignment can be performed accordingly. If the peak points of multiple sets of floating pulse waveforms, middle pulse waveforms, and deep pulse waveforms cannot be directly aligned, then the distance difference needs to be calculated to minimize the cumulative value of the distance difference.
[0144] The distance difference here refers to the horizontal distance between the peak points of a set of corresponding floating pulse waveforms, middle pulse waveforms, and deep pulse waveforms.
[0145] In some cases, the floating pulse waveform is filtered in the following ways:
[0146] The floating pulse waveform is decomposed in the frequency domain to obtain a low-frequency decomposed waveform group and a high-frequency decomposed waveform group.
[0147] Calculate the residual between the waveform composed of a low-frequency decomposition waveform group and the floating pulse waveform;
[0148] Based on the residuals, the low-frequency decomposition waveform group and the high-frequency decomposition waveform group are re-divided, and the residuals of the waveforms composed of the low-frequency decomposition waveform group and the floating pulse waveform are calculated again until the residuals are less than or equal to the reference value.
[0149] The first low-frequency decomposition waveform group is processed using grouping and residual methods to obtain the second low-frequency decomposition waveform group and the second high-frequency decomposition waveform group, and so on, to obtain the Nth low-frequency decomposition waveform group and the Nth high-frequency decomposition waveform group. Please refer to [link to relevant documentation]. Figure 9 ;
[0150] Filter the high-frequency decomposition waveform groups from the first to the Nth high-frequency decomposition waveform groups to remove high-frequency decomposition waveform groups containing noise.
[0151] A floating pulse waveform is formed by combining the first to Nth low-frequency decomposition waveform groups with the remaining high-frequency decomposition waveforms, such as... Figure 10 As shown;
[0152] The filtering process for the middle pulse waveform and the deep pulse waveform is the same as that for the floating pulse waveform.
[0153] N is a natural number greater than or equal to 2.
[0154] The purpose of the above method is to remove noise from the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform. Specifically, the waveform obtained by pulse diagnosis contains noise, which are various noises and non-target physiological signals that interfere with the true pulse signal.
[0155] The method used in this application is to decompose the floating pulse waveform in the frequency domain to obtain a low-frequency decomposed waveform group and a high-frequency decomposed waveform group. That is, the floating pulse waveform is grouped according to frequency, which will result in a low-frequency decomposed waveform group and a high-frequency decomposed waveform group.
[0156] Here, a low-frequency decomposition waveform group represents the overall profile, baseline, and main trends of the signal, while a high-frequency decomposition waveform group represents the details, abrupt changes, and noise of the signal.
[0157] Next, calculate the residual between the waveform composed of the low-frequency decomposition waveform group and the floating pulse waveform. If the residual is less than or equal to the reference value, it indicates that the division of the low-frequency decomposition waveform group and the high-frequency decomposition waveform group is reasonable; otherwise, the division is unreasonable and needs to be readjusted.
[0158] The residual specifically refers to the difference between two waveforms, and is calculated as follows:
[0159] MSE=(1 / N)×Σ[Soriginal(i)-Scomponent(i)]²;
[0160] N: The total number of sampling points for the signal;
[0161] Σ: Sum the squared differences of all sampling points (i from 1 to N).
[0162] When the calculated MSE residual is less than or equal to the reference value, it indicates that the current low-frequency decomposition waveform is very close to the main outline of the original signal, or that enough high-frequency components have been stripped away, and the current decomposition iteration can be stopped.
[0163] Here, in the first grouping process, the ratio of the number of waveforms in the low-frequency decomposed waveform group to the number of waveforms in the high-frequency decomposed waveform group is generally controlled at 9:1 or 8:2.
[0164] The reference range for residuals is when the energy of the residuals accounts for less than or equal to 1%-3% of the total energy of the original signal. In this case, the grouping method of the first low-frequency decomposition waveform group and the first high-frequency decomposition waveform group is considered appropriate.
[0165] In general, the value of N in the above method is in the range of 4-5.
[0166] The specific method for filtering waveform groups from one high-frequency decomposition to N high-frequency decomposition is as follows:
[0167] The high-frequency decomposed waveform group and the original waveform are sampled as follows:
[0168] X = (x1, x2, …, x n ),
[0169] Y = (y1, y2, …, y n )
[0170] Calculate the average of the two sets of data using the following formula:
[0171] ,
[0172] ,
[0173] Substitute into the formula to calculate:
[0174] ,
[0175] The calculated result of r is as follows:
[0176] r close to 1: This indicates that the two sets of waveforms are highly positively correlated and have completely consistent trends.
[0177] r close to 0: indicates that the two sets of waveforms have almost no linear relationship.
[0178] r close to -1: indicates that the two sets of waveforms are completely negatively correlated and have completely opposite trends.
[0179] Based on the calculation results, only the high-frequency decomposed waveform groups with r close to 1 are retained.
[0180] This application also provides a pulse sampling device for auxiliary diagnosis in traditional Chinese medicine, comprising:
[0181] The data acquisition unit is used to acquire hand samples of the sampling object and calculate the converted size ratio based on the hand samples;
[0182] The first position adjustment unit is used to adjust the spacing of the signal acquisition terminals and determine the detection area in response to the received position adjustment command;
[0183] The second position adjustment unit is used to adjust the position of the signal acquisition terminal in the detection area according to the signal feedback from the signal acquisition terminal in the detection area, and to determine the acquisition position of the signal acquisition terminal.
[0184] The third position adjustment unit is used to drive the signal acquisition end to determine the acquisition height at the acquisition position. The acquisition height includes the surface height, the elastic height, and the rigid height.
[0185] The signal acquisition unit is used to acquire signals at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms.
[0186] The sample processing unit is used to align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
[0187] Furthermore, adjusting the position of the signal acquisition terminal in the detection area based on the signal feedback from the signal acquisition terminal in the detection area includes:
[0188] Create a straight line trajectory along the length of the detection area;
[0189] Select collection points on the straight line of the movement trajectory. There are multiple collection points, and the collection points are set at intervals on the straight line of the movement trajectory.
[0190] Signals are acquired at the acquisition points to obtain the acquired waveforms;
[0191] Calculate the peak value of the acquired waveform and calculate the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform;
[0192] The sampling depths at all sampling points were the same.
[0193] Furthermore, the acquisition position of the signal acquisition end is calculated based on the peak values of the acquired waveform, including:
[0194] Construct discrete points in a coordinate system. The horizontal coordinate of the discrete points is the position coordinate of the acquisition point, and the vertical coordinate of the discrete points is the wave peak value at the discrete points.
[0195] Construct numerical curves of wave crests using discrete points in a coordinate system;
[0196] Calculate the peak of the peak numerical curve and use the corresponding position of the peak numerical curve as the acquisition position of the signal acquisition terminal.
[0197] Furthermore, the process of determining the acquisition height at the acquisition location by the drive signal acquisition terminal includes:
[0198] The drive signal acquisition end moves linearly at the acquisition position, synchronously recording the moving distance and deformation.
[0199] Create a movement curve using the movement distance as the horizontal axis and the deformation as the vertical axis.
[0200] Analyze and identify abrupt change points on the moving curve and record the corresponding moving distances at these points;
[0201] The sampling height range is determined by the movement distance corresponding to the abrupt change point, and there are multiple sampling height ranges.
[0202] Select a value as the acquisition height within each acquisition height range.
[0203] Furthermore, the number of height ranges collected is three;
[0204] The surface height is within the first acquisition height range, the elastic height is within the second acquisition height range, and the rigid height is within the third acquisition height range;
[0205] The first data collection showed a positive correlation between the distance traveled and the deformation within the height range.
[0206] The distance traveled and the deformation within the second acquisition height range are positively correlated, and the positive correlation coefficient between the distance traveled and the deformation within the second acquisition height range is greater than that between the distance traveled and the deformation within the first acquisition height range.
[0207] The third finding shows a linear correlation between the distance traveled and the deformation within the height range.
[0208] Furthermore, aligning the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform includes:
[0209] The floating pulse waveform, middle pulse waveform, and deep pulse waveform are filtered and grouped according to the acquisition location. The floating pulse waveform, middle pulse waveform, and deep pulse waveform in the same group are generated at the same acquisition location.
[0210] Calculate the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform respectively;
[0211] Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform according to their peak points.
[0212] Among them, the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are located at the same time point.
[0213] Furthermore, filtering the floating pulse waveform includes:
[0214] The floating pulse waveform is decomposed in the frequency domain to obtain a low-frequency decomposed waveform group and a high-frequency decomposed waveform group.
[0215] Calculate the residual between the waveform composed of a low-frequency decomposition waveform group and the floating pulse waveform;
[0216] Based on the residuals, the low-frequency decomposition waveform group and the high-frequency decomposition waveform group are re-divided, and the residuals of the waveforms composed of the low-frequency decomposition waveform group and the floating pulse waveform are calculated again until the residuals are less than or equal to the reference value.
[0217] The low-frequency decomposition waveform group is processed by grouping and residual processing to obtain the second low-frequency decomposition waveform group and the second high-frequency decomposition waveform group, and so on, to obtain the Nth low-frequency decomposition waveform group and the Nth high-frequency decomposition waveform group.
[0218] Filter the high-frequency decomposition waveform groups from the first to the Nth high-frequency decomposition waveform groups to remove high-frequency decomposition waveform groups containing noise.
[0219] The floating pulse waveform is formed by combining the low-frequency decomposition waveform group from one to N times with the remaining high-frequency decomposition waveform.
[0220] The filtering process for the middle pulse waveform and the deep pulse waveform is the same as that for the floating pulse waveform.
[0221] N is a natural number greater than or equal to 2.
[0222] In one example, the unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.
[0223] For example, when the units in the device can be implemented through a processing element scheduler, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling programs. Alternatively, these units can be integrated together to form a system-on-a-chip (SOC).
[0224] In this application, various objects such as messages / information / devices / network elements / systems / apparatus / actions / operations / processes / concepts may be named. It is understood that these specific names do not constitute a limitation on the relevant objects. The names may be changed depending on the scenario, context, or usage habits. The understanding of the technical meaning of the technical terms in this application should be mainly determined from their functions and technical effects embodied / performed in the technical solution.
[0225] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0226] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0227] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0228] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0229] It should also be understood that in the various embodiments of this application, "first," "second," etc., are merely used to indicate that multiple objects are different. For example, a first time window and a second time window are only used to indicate different time windows. They should not have any effect on the time windows themselves, and the aforementioned "first," "second," etc., should not impose any limitations on the embodiments of this application.
[0230] It should also be understood that, in the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.
[0231] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0232] This application also provides an AI pulse diagnosis system for auxiliary diagnosis in traditional Chinese medicine, the system comprising:
[0233] One or more memories for storing instructions; and
[0234] One or more processors are configured to retrieve and execute the instructions from the memory, performing the methods described above.
[0235] This application also provides a computer program product including instructions that, when executed, cause the terminal device and the network device to perform operations corresponding to the methods described above.
[0236] This application also provides a chip system including a processor for implementing the functions involved in the above description, such as generating, receiving, transmitting, or processing the data and / or information involved in the above methods.
[0237] This chip system can consist of chips or include chips and other discrete components.
[0238] The processor mentioned above can be a CPU, a microprocessor, an ASIC, or one or more integrated circuits that execute a program to control the method of transmitting the feedback information described above.
[0239] In one possible design, the chip system also includes a memory for storing necessary program instructions and data. The processor and the memory can be decoupled and located on different devices, connected via wired or wireless means to support the chip system in implementing the various functions described in the above embodiments. Alternatively, the processor and the memory can also be coupled to the same device.
[0240] Optionally, the computer instructions are stored in memory.
[0241] Optionally, the memory can be a storage unit within the chip, such as a register or cache. Alternatively, the memory can be a storage unit located outside the chip within the terminal, such as a ROM or other types of static storage devices that can store static information and instructions, such as RAM.
[0242] It is understood that the memory in this application may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.
[0243] Non-volatile memory can be ROM, programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.
[0244] Volatile memory can be RAM, which is used as an external cache. There are many different types of RAM, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory.
[0245] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.
Claims
1. A pulse sampling method for auxiliary diagnosis in Traditional Chinese Medicine, characterized in that, include: Obtain hand samples of the sampling object and calculate the converted size ratio based on the hand samples; In response to the received position adjustment command, the spacing of the signal acquisition terminals is adjusted and the detection area is determined; Adjust the position of the signal acquisition terminal in the detection area based on the signal feedback from the signal acquisition terminal in the detection area to determine the acquisition position of the signal acquisition terminal; The drive signal acquisition terminal determines the acquisition height at the acquisition position, which includes the surface height, elastic height, and rigid height. Signal acquisition is performed at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms. Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
2. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 1, characterized in that, Adjusting the position of the signal acquisition terminal within the detection area based on signal feedback from the signal acquisition terminal includes: Create a straight line trajectory along the length of the detection area; Select collection points on the straight line of the movement trajectory. There are multiple collection points, and the collection points are set at intervals on the straight line of the movement trajectory. Signals are acquired at the acquisition points to obtain the acquired waveforms; Calculate the peak value of the acquired waveform and calculate the acquisition position of the signal acquisition terminal based on the peak value of the acquired waveform; The sampling depths at all sampling points were the same.
3. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 2, characterized in that, The acquisition location of the signal acquisition terminal is calculated based on the peak values of the acquired waveform, including: Construct discrete points in a coordinate system. The horizontal coordinate of the discrete points is the position coordinate of the acquisition point, and the vertical coordinate of the discrete points is the wave peak value at the discrete points. Construct numerical curves of wave crests using discrete points in a coordinate system; Calculate the peak of the peak numerical curve and use the corresponding position of the peak numerical curve as the acquisition position of the signal acquisition terminal.
4. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 1, characterized in that, The drive signal acquisition terminal determines the acquisition height at the acquisition position, including: The drive signal acquisition end moves linearly at the acquisition position, synchronously recording the moving distance and deformation. Create a movement curve using the movement distance as the horizontal axis and the deformation as the vertical axis. Analyze and identify abrupt change points on the moving curve and record the corresponding moving distances at these points; The sampling height range is determined by the movement distance corresponding to the abrupt change point, and there are multiple sampling height ranges. Select a value as the acquisition height within each acquisition height range.
5. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 4, characterized in that, The number of height ranges to be collected is three; The surface height is within the first acquisition height range, the elastic height is within the second acquisition height range, and the rigid height is within the third acquisition height range; The first data collection showed a positive correlation between the distance traveled and the deformation within the height range. The distance traveled and the deformation within the second acquisition height range are positively correlated, and the positive correlation coefficient between the distance traveled and the deformation within the second acquisition height range is greater than that between the distance traveled and the deformation within the first acquisition height range. The third finding shows a linear correlation between the distance traveled and the deformation within the height range.
6. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 1, characterized in that, Aligning the waveforms of the floating pulse, the middle pulse, and the deep pulse includes: The floating pulse waveform, middle pulse waveform, and deep pulse waveform are filtered and grouped according to the acquisition location. The floating pulse waveform, middle pulse waveform, and deep pulse waveform in the same group are generated at the same acquisition location. Calculate the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform respectively; Align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform according to their peak points. Among them, the peak points of the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform are located at the same time point.
7. The pulse sampling method for auxiliary diagnosis in traditional Chinese medicine according to claim 6, characterized in that, Filtering the floating pulse waveform includes: The floating pulse waveform is decomposed in the frequency domain to obtain a low-frequency decomposed waveform group and a high-frequency decomposed waveform group. Calculate the residual between the waveform composed of a low-frequency decomposition waveform group and the floating pulse waveform; Based on the residuals, the low-frequency decomposition waveform group and the high-frequency decomposition waveform group are re-divided, and the residuals of the waveforms composed of the low-frequency decomposition waveform group and the floating pulse waveform are calculated again until the residuals are less than or equal to the reference value. The low-frequency decomposition waveform group is processed by grouping and residual processing to obtain the second low-frequency decomposition waveform group and the second high-frequency decomposition waveform group, and so on, to obtain the Nth low-frequency decomposition waveform group and the Nth high-frequency decomposition waveform group. Filter the high-frequency decomposition waveform groups from the first to the Nth high-frequency decomposition waveform groups to remove high-frequency decomposition waveform groups containing noise. The floating pulse waveform is formed by combining the low-frequency decomposition waveform group from one to N times with the remaining high-frequency decomposition waveform. The filtering process for the middle pulse waveform and the deep pulse waveform is the same as that for the floating pulse waveform. N is a natural number greater than or equal to 2.
8. A pulse sampling device for auxiliary diagnosis in Traditional Chinese Medicine, characterized in that, include: The data acquisition unit is used to acquire hand samples of the sampling object and calculate the converted size ratio based on the hand samples; The first position adjustment unit is used to adjust the spacing of the signal acquisition terminals and determine the detection area in response to the received position adjustment command; The second position adjustment unit is used to adjust the position of the signal acquisition terminal in the detection area according to the signal feedback from the signal acquisition terminal in the detection area, and to determine the acquisition position of the signal acquisition terminal. The third position adjustment unit is used to drive the signal acquisition end to determine the acquisition height at the acquisition position. The acquisition height includes the surface height, the elastic height, and the rigid height. The signal acquisition unit is used to acquire signals at the acquisition height to obtain pulse signal waveforms, which include floating pulse waveforms, medium pulse waveforms, and deep pulse waveforms. The sample processing unit is used to align the floating pulse waveform, the middle pulse waveform, and the deep pulse waveform to obtain a pulse sample.
9. An AI pulse diagnosis system for auxiliary diagnosis in Traditional Chinese Medicine, characterized in that, The system includes: One or more memories for storing instructions; and One or more processors are configured to retrieve and execute the instructions from the memory to perform the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes: The program, when run by the processor, executes the method as described in any one of claims 1 to 7.