Depth-controllable piezoelectric shock wave treatment system based on automatic focusing

Abstract

The invention relates to the technical field of medical instrument control, in particular to a depth-controllable piezoelectric shock wave treatment system based on automatic focusing. The data acquisition module system collects sample data of patients with the same disease type, and the sample data comprises historical and to-be-tested samples and covers multi-dimensional information such as treatment feedback texts and treatment area feature parameters. As patients have individual differences, the sample data analysis module obtains sample tolerance sensitivity by analyzing pain test data, compares treatment feedback texts to determine the complication influence degree, and accurately grasps individual physiological and pathological characteristics of the patients. The linkage feature analysis module comprehensively considers similar features such as feature parameters of a treatment area, determines the expression degree of linkage features between samples, and mines internal relation and similarity rules. And the treatment control parameter determination module trains a neural network based on the linkage feature representation degree and the treatment control parameters, and accurately determines the treatment control parameters of the to-be-detected sample by using the learning and prediction capabilities of the neural network, thereby avoiding excessive or insufficient treatment.

Description

Technical Field

[0001] This invention relates to the field of medical device control technology, specifically to a depth-controllable piezoelectric shock wave therapy system based on automatic focusing. Background Technology

[0002] Piezoelectric shock wave therapy, especially depth-controlled piezoelectric shock waves, is a non-invasive treatment widely used in the medical field, such as orthopedic rehabilitation and chronic pain management. Piezoelectric shock wave therapy uses high-energy sound waves to act on human tissue, effectively stimulating blood circulation. Depth-controlled piezoelectric shock wave therapy systems combine precise focusing technology (the core of which is controlling the focusing position of the energy wave to ensure precise energy application to the target area) to adjust various treatment control parameters of the piezoelectric shock wave according to different symptoms, thereby improving treatment efficacy.

[0003] In existing technologies, when setting various treatment control parameters for piezoelectric shock waves during patient treatment, the treatment control parameters of historical samples with similar treatment area characteristics are often selected as a reference. However, since each patient has unique physiological and pathological characteristics, their pain tolerance and treatment effects vary from person to person. Therefore, if the similarity of treatment area characteristics is relied upon and the individual differences of patients are ignored, the accuracy of the final treatment control parameter settings for the test sample will be low, which may seriously affect the treatment effect. Summary of the Invention

[0004] To address the technical problem that each patient possesses unique physiological and pathological characteristics, resulting in varying pain tolerance and treatment outcomes, relying solely on similarities in treatment area characteristics while ignoring individual patient differences can lead to inaccurate settings of treatment control parameters for the test samples, potentially impacting treatment effectiveness, this invention aims to provide a depth-controllable piezoelectric shock wave therapy system based on automatic focusing. The specific technical solution adopted is as follows: This invention proposes a depth-controllable piezoelectric shock wave therapy system based on automatic focusing, the system comprising: The data acquisition module is used to acquire sample data from patients with the same disease type. The sample data includes historical samples and samples to be tested. Each sample data includes the patient's treatment feedback text, treatment area feature parameters, treatment control parameters, and the patient's pain test data for each treatment. The sample data analysis module is used to analyze the changing trends and numerical similarities of pain test data in the sample data to obtain the tolerance sensitivity of each sample data; within each sample data, the similarity between treatment feedback texts is compared to determine the complication impact of each sample data. The linkage feature analysis module is used to determine the linkage feature performance between any two sample data by comprehensively considering the similarity features of treatment area feature parameters, the similarity features of tolerance sensitivity, and the similarity features of complication impact between sample data. The treatment control parameter determination module is used to train a neural network based on the correlation feature performance between sample data and the treatment control parameters, thereby determining the treatment control parameters of the sample to be tested.

[0005] Furthermore, the method for obtaining the tolerance sensitivity includes: The pain test data includes subjective pain scores and electromyographic signal data at each test energy level, with the test energy gradually increasing. In each sample data, the subjective pain sensitivity of each sample data is determined based on the changing trend of subjective pain scores; By analyzing the similarity and numerical characteristics of electromyographic signals among the sample data, the objective pain sensitivity of each sample data was determined. The normalized value of the product of subjective pain sensitivity and objective pain sensitivity for each sample is taken as the tolerance sensitivity for each sample.

[0006] Furthermore, the method for obtaining the subjective pain sensitivity includes: In each sample data, the subjective pain scores are sorted in ascending order of test energy to obtain a sorting sequence; In the sorting sequence, the ratio of each subjective pain score to the adjacent previous pain score is calculated as the perceived pain enhancement for each subjective pain score. In each sample data, the subjective pain score is weighted and fused using the pain enhancement perception of each subjective pain score, and the normalized value of the weighted result is used as the subjective pain sensitivity of each sample data.

[0007] Furthermore, the method for obtaining the objective pain sensitivity includes: In the pain test data of each sample data, the absolute area of ​​the electromyographic signal data and the horizontal axis under each test energy is used as the electromyographic energy characterization value; For any two sample data, analyze the differences in electromyographic energy characterization values ​​of the two sample data under the same test energy, and combine the test energy numerical characteristics to obtain the objective stimulus performance difference between the two sample data. Cluster analysis was performed on all sample data based on the objective stimulus performance differences among the sample data to obtain all clusters; Within each cluster, the mean value of the peak values ​​of the electromyography signals of all sample data at all test energies is normalized and used as the objective pain sensitivity of each sample data in each cluster.

[0008] Furthermore, the method for obtaining the objective stimulus performance difference includes: For any two sample data, under the same test energy, the absolute value of the difference between the electromyographic energy characterization values ​​of the two sample data is taken as the electromyographic difference characteristic value of the two sample data under the same test energy. The ratio of each test energy to the sum of all test energies is used as the stimulus intensity weight for each test energy; For any two sample data, the electromyographic difference feature values ​​of the two sample data under the same test energy are weighted and fused using the stimulus intensity weight of the test energy, and the normalized value of the weighted result is used as the objective stimulus performance difference between the two sample data.

[0009] Furthermore, the method for obtaining the impact of the complications includes: For each sample data point, complication keywords were extracted from the treatment feedback text at each treatment session to obtain the complication keyword set for each treatment session. In each sample data, the Jaccard correlation coefficient between the complication keyword set at the last treatment and the complication keyword set at each previous treatment is calculated as the complication similarity factor; The sum of all complication similarity factors corresponding to each sample data is negatively correlated and mapped to the value, which is then used as the complication impact of each sample data.

[0010] Furthermore, the method for obtaining the linkage feature representation degree includes: Analyze the similarity features of the treatment area characteristic parameters between any two sample data to obtain the treatment area similarity between the two sample data; The absolute value of the difference in tolerance sensitivity between the two sample data is negatively correlated and normalized, and this value is taken as the tolerance sensitivity similarity between the two sample data. The absolute value of the difference in the impact of complications between the two sample data is negatively correlated and normalized, and this value is taken as the similarity of the impact of complications between the two sample data. The similarity of the treatment area between the two sample data is used to weight the similarity of tolerance sensitivity, and the difference between the treatment area similarity and the constant 1 is used to weight the similarity of complication impact. The weighted result is normalized and used as the linkage feature performance between the two sample data.

[0011] Furthermore, the method for obtaining the similarity of the treatment areas includes: The treatment area feature parameters include at least the distance from the skin to the target point and the thickness of the subcutaneous fat layer. The treatment area feature parameters of each sample data for each treatment are combined into a feature parameter vector. For any two sample data, calculate the Euclidean distance between any two feature parameter vectors between the two sample data, perform negative correlation mapping and normalization to obtain the treatment area similarity factor between the two sample data, and take the mean of all treatment area similarity factors between the two sample data as the treatment area similarity.

[0012] Furthermore, the neural network trained based on the linkage feature representation degree between sample data and treatment control parameters, thereby determining the treatment control parameters of the test sample, includes: Cluster analysis is performed on all historical samples based on the correlation characteristics among all historical samples to obtain clusters; Within each cluster, the mean of the linkage feature performance between each historical sample and other historical samples is used as the treatment performance feature value of each historical sample. In each cluster, the treatment performance feature value of each historical sample is used as the input of the neural network. The mean value of each historical sample under each treatment control parameter is calculated as the control parameter factor. The control parameter factor of each historical sample under all kinds of treatment control parameters is used as the output, thereby training the neural network and obtaining a trained neural network. The test sample is assigned to the cluster of the historical sample with the highest linkage feature performance. The mean of the linkage feature performance between the test sample and all historical samples in the cluster is used as the input of the trained neural network corresponding to the cluster, thereby obtaining the value of each treatment control parameter for the test sample.

[0013] Furthermore, the cluster analysis of all historical samples based on the correlation feature performance among all historical samples is used to obtain clusters, including: In all historical samples, cluster analysis is performed on all historical sample data based on the K-means clustering algorithm and a preset optimal K value to obtain all clusters. The distance metric is the value after negatively mapping the correlation characteristics between historical sample data.

[0014] The present invention has the following beneficial effects: The data acquisition module systematically collects sample data from patients with the same disease type. This data is categorized into historical samples and samples to be tested, encompassing multi-dimensional information such as treatment feedback text, treatment area characteristic parameters, treatment control parameters, and pain test data. Due to individual patient differences, each person's pain tolerance and post-treatment feedback vary. Therefore, the sample data analysis module accurately determines the tolerance sensitivity of each sample by analyzing the changing trends and numerical similarities of pain test data. Simultaneously, by comparing the similarity of treatment feedback texts among the sample data, the impact of complications is precisely determined. This module accurately grasps the individual physiological and pathological characteristics of patients, providing crucial evidence for personalized treatment and contributing to improved treatment targeting and effectiveness. Furthermore, the linkage feature analysis module comprehensively considers the similarities in treatment area characteristic parameters, tolerance sensitivity, and complication impact among sample data to determine the linkage feature performance between any two sample data. By uncovering the inherent connections and similar patterns between different sample data, it helps to more rationally reference historical samples in subsequent processes, providing a more scientific reference for setting treatment control parameters for the samples to be tested. Therefore, in the final treatment control parameter determination module, a neural network is trained based on the correlation characteristics between sample data and the treatment control parameters. Leveraging the powerful learning and prediction capabilities of the neural network, the treatment control parameters for the test sample can be accurately determined according to the individual characteristics of the patient and the correlation between sample data. In summary, this invention fully considers individual differences and similarities between sample data, enabling precise setting of treatment control parameters and effectively avoiding overtreatment or undertreatment. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0016] Figure 1 This is a system block diagram of a depth-controllable piezoelectric shock wave therapy system based on autofocus, provided in one embodiment of the present invention. Figure 2 This is a flowchart of a method for obtaining tolerance sensitivity according to an embodiment of the present invention; Figure 3 This is a flowchart of a method for obtaining the performance of linkage features according to an embodiment of the present invention; Figure 4This is a schematic diagram of the system structure of a depth-controllable piezoelectric shock wave therapy system based on automatic focusing, provided in one embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an autofocus-based depth-controllable piezoelectric shock wave therapy system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following description, in conjunction with the accompanying drawings, details a specific solution for a depth-controllable piezoelectric shock wave therapy system based on automatic focusing provided by the present invention.

[0020] Please see Figure 1 The diagram illustrates a system block diagram of a depth-controllable piezoelectric shock wave therapy system based on autofocus, according to an embodiment of the present invention. The system includes: a data acquisition module 101, a sample data analysis module 102, a linkage feature analysis module 103, and a treatment control parameter determination module 104.

[0021] The data acquisition module 101 is used to acquire sample data of patients with the same disease type. The sample data includes historical samples and samples to be tested. Each sample data includes the patient's treatment feedback text, treatment area feature parameters, treatment control parameters, and the patient's pain test data for each treatment.

[0022] During piezoelectric shock wave therapy, the system integrates a real-time imaging monitor (such as ultrasound or X-ray). By locking onto the target (such as a kidney stone), the computer system automatically calculates the precise depth and location of the target and automatically controls the piezoelectric ceramic array. Through complex electronic phase control technology, it dynamically changes the focal point of the shock wave, so that it always automatically aligns with and tracks the moving target (such as a tumor in the kidney or liver that moves slightly when breathing).

[0023] However, this highly efficient real-time treatment system also relies on a comprehensive, multi-dimensional analysis of the patient's condition before surgery. By comprehensively considering factors such as the patient's physical condition, the location of the lesion, and the size and shape of the target, doctors can better adjust the treatment control parameters during the treatment process, such as the frequency, number of shocks, energy intensity, and treatment time. Optimizing these treatment control parameters can improve treatment effectiveness, reduce side effects, and provide a guarantee for personalized treatment.

[0024] In this embodiment of the invention, the data acquisition module can extract sample data of patients with the same disease type from the hospital's database, such as kidney stones or liver tumors; wherein, the sample data of historical patients are recorded as historical samples, and the sample data of the patients to be tested are recorded as samples to be tested.

[0025] Because piezoelectric shock wave therapy is not a one-time procedure, but rather a multi-stage treatment process considering the patient's condition and the discomfort of a single treatment, treatment feedback text can be extracted from the patient's medical records after each treatment to assess the patient's experience during each treatment and any potential complications. Furthermore, the spatial geometry of the treatment area determines the shape, location, and size of the lesion, and the treatment control parameters reflect the treatment plan. These are also key indicators of individual patient differences; therefore, characteristic parameters of the treatment area and treatment control parameters are also extracted from the medical records after each treatment.

[0026] Meanwhile, since different patients have different pain tolerance levels, in order to ensure the comfort and effectiveness of the treatment, it is necessary to consider the patient's pain tolerance. Therefore, in this embodiment of the invention, a pain test was performed on each patient, and the pain test data of each patient was obtained.

[0027] At this point, multiple historical samples and samples to be tested under the same disease type can be obtained.

[0028] It should be noted that, in this embodiment of the present invention, the characteristic parameters of the treatment area include at least the distance from the skin to the target point and the thickness of the subcutaneous fat layer; the treatment control parameters include at least the frequency of the shock wave, the number of shocks, the energy intensity, and the treatment time.

[0029] Pain testing procedure: Test area: Normal tissue near the planned treatment target.

[0030] Test parameters: (1) Fixed benchmark test energy: Use a benchmark energy that is far below the treatment level and is absolutely safe.

[0031] (2) Incremental energy / frequency: emit a series of shock waves (e.g., 3-5 times), and the energy is increased slightly according to a preset gradient (e.g., each time increasing by 0.2 times the original base).

[0032] (3) Fixed focal depth: placed at a specific depth under the skin (e.g., 1 cm).

[0033] Thus, after each pulse, the patient scores their pain (0-10 points, with the pain level gradually increasing as the score increases; no scoring is required after the first pulse, and a default score of 1 is used to provide a reference for comparison after subsequent pulses). This score serves as the patient's subjective pain rating. Simultaneously, an attached surface electromyography (sEMG) sensor acquires electromyographic signal data after each pulse. Based on the aforementioned steps, the pain test data corresponding to each sample can be obtained, and the pain test data includes the subjective pain score and electromyographic signal data for each test energy level.

[0034] In this embodiment of the invention, the collection and acquisition of patients' personal information data are all authorized by the relevant users, and the process does not violate relevant laws and regulations, nor does it violate public order and good morals.

[0035] The sample data analysis module 102 is used to analyze the changing trends and numerical similarity characteristics of pain test data in the sample data to obtain the tolerance sensitivity of each sample data; in each sample data, the similarity between treatment feedback texts is compared to determine the complication impact of each sample data.

[0036] In actual treatment, individual differences among patients, particularly their pain threshold and tolerance to treatment, can vary significantly. This process depends not only on the patient's physiological condition but may also be influenced by psychological factors. Therefore, to ensure the comfort and effectiveness of subsequent treatment, a better understanding of patients' pain sensitivity differences is needed. This can be achieved by analyzing the changing trends and numerical similarities of pain test data in the sample data, and calculating the tolerance sensitivity for each sample data point to characterize the patient's pain tolerance for each sample data point.

[0037] Preferably, in one embodiment of the present invention, the method for obtaining tolerance sensitivity includes: Please see Figure 2 The diagram illustrates a method flowchart for obtaining tolerance sensitivity in one embodiment of the present invention, the method comprising the following steps: Step S201: In each sample data, determine the subjective pain sensitivity of each sample data based on the changing trend of the subjective pain score.

[0038] Test energy is a key factor affecting patients' pain perception. As test energy increases, patients' pain response usually increases as well. Therefore, in each sample of data, subjective pain scores are sorted in ascending order of test energy to obtain a sorting sequence.

[0039] Then, in the sorted sequence, the ratio of each subjective pain score to the adjacent previous pain score is calculated as the pain enhancement perception for each subjective pain score. The pain enhancement perception can measure the relative change in pain intensity between two adjacent test energies, and the larger the index, the greater the patient's pain sensitivity is compared to the previous test energy level as the test energy increases.

[0040] Finally, in each sample data, the pain score is weighted and fused using the pain enhancement perception of each subjective pain score: the pain enhancement perception is normalized and used as the pain perception weight; the product of the pain perception weight under each test energy and the subjective pain score is used as the weighted score; and the sum of all weighted scores for each sample data is normalized and used as the subjective pain sensitivity of each sample data. The higher the subjective pain sensitivity, the greater the patient's sensitivity to pain. Normalization is a technique well-known to those skilled in the art; the normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here; the sum of the pain perception weights in each sample data must be 1.

[0041] It should be noted that, for the first subjective pain score in the ranking sequence, since it does not have the previous subjective pain score, the pain enhancement perception of the first subjective pain score is set to be consistent with that of the second subjective pain score.

[0042] Step S202: Analyze the similarity and numerical characteristics of electromyographic signals among the sample data to determine the objective pain sensitivity of each sample data.

[0043] Electromyography (EMG) signals are the direct physiological manifestation of muscle activity. When stimulated by pain, muscles generate electrical activity. Therefore, in the pain test data of each sample, the absolute area of ​​the EMG signal data at each test energy level with respect to the horizontal axis is used as the EMG energy representation value. The absolute area is calculated by integrating the EMG signal data. The integral value of the EMG signal data is essentially the area between the signal data curve and the horizontal axis. The area above the horizontal axis (vertical coordinate greater than 0) is positive, and the area below the horizontal axis (vertical coordinate less than 0) is negative. Therefore, the absolute area is the area above the horizontal axis minus the area below the horizontal axis.

[0044] In pain testing, increasing the energy intensity of the test often helps to reveal differences in patients’ perception of stimuli. In particular, the greater the intensity, the more obvious the differences in individual perception of stimuli become, especially in the response of the nervous system. For patients with certain diseases or physiological conditions, stronger stimuli may elicit more significant physiological responses, thus allowing for a better assessment of the objective pain tolerance of different patients.

[0045] Therefore, for any two sample data points, under the same test energy, the absolute value of the difference in the electromyographic energy characterization values ​​of these two sample data points is taken as the electromyographic difference feature value of these two sample data points under the same test energy. The larger the electromyographic difference feature value, the greater the difference in pain perception tolerance between the two sample data points. Then, the ratio of each test energy to the sum of all test energies is taken as the stimulus intensity weight of each test energy. For any two sample data points, the electromyographic difference feature values ​​of these two sample data points under the same test energy are weighted and fused using the stimulus intensity weight of the test energy. That is, the stimulus intensity weight of each test energy is multiplied by the electromyographic difference feature value of these two sample data points under each test energy to obtain the weighted difference feature value. The larger this value, the lower the similarity of pain perception between the two sample data points under higher energy stimulation. Then, the sum of the weighted difference feature values ​​of these two sample data points under all test energies is normalized to obtain the objective stimulus performance difference between the two sample data points. The larger the objective stimulus performance difference, the greater the difference in pain perception between the two sample data points under the same energy test. Normalization is a technique well-known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0046] Cluster analysis is performed on all sample data based on the objective differences in stimulus performance among the sample data to obtain all clusters. The K-means algorithm can be used for this cluster analysis, with the distance metric being the objective differences in stimulus performance among the sample data. The default value for K is 5, but the specific value can be adjusted according to the implementation scenario and is not limited here.

[0047] At this point, the sample data in each cluster exhibit a relatively consistent level of pain stimulus perception. Since the peak value of the electromyographic (EMG) signal represents the maximum intensity of muscle contraction, and the more intense the pain, the stronger the muscle contraction tends to be, the mean of the peak values ​​of the EMG signals of all sample data under all test energies is normalized within each cluster. This normalized value is then used as the objective pain sensitivity of each sample data in each cluster. The higher the objective pain sensitivity, the more pronounced the pain response represented by the patient's EMG signal. Normalization is a technique well-known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0048] Step S203: Combine the subjective pain sensitivity and objective pain sensitivity of each sample data to obtain the tolerance sensitivity of each sample data.

[0049] Based on the analyses in steps S201 and S202, the subjective pain sensitivity and objective pain sensitivity of each sample data can be calculated. Both of these indicators are positively correlated with the patient's sensitivity to pain. Therefore, the normalized product of the subjective pain sensitivity and objective pain sensitivity of each sample data is used as the tolerance sensitivity of each sample data. The higher the tolerance sensitivity, the more obvious the pain perception during the treatment process. This indicator can serve as a key indicator reflecting individual differences in the sample data. Normalization is a technique well-known to those skilled in the art. The choice of normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0050] Furthermore, since each patient's physical condition and underlying diseases may differ, these factors can affect the occurrence and severity of complications. Therefore, we can continue to compare the similarity between treatment feedback texts in each sample data to determine the degree of complication impact of each sample data, which also serves as a key indicator reflecting individual differences in the sample data.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining the impact of complications includes: Based on medical data, keywords related to complications were extracted from the treatment feedback text of each sample data at each treatment session to obtain a set of keywords related to complications at each treatment session.

[0052] In each sample, the Jaccard correlation coefficient between the complication keyword set at the last treatment and the complication keyword set at each previous treatment was calculated as the complication similarity factor. The larger the complication similarity factor, the higher the similarity of the complication keyword sets between the two treatments, the more similar the complication manifestations, and the more likely the response is to conform to the patient's common response pattern and thus be within the expected range. The treatment plan did not cause new abnormalities and is more within the normal treatment response range. Conversely, if the complication similarity factor is smaller, it indicates that the complication manifestations after the two treatments are significantly different, and the last treatment may have had an inappropriate impact on the patient's tissue or physiological state.

[0053] Therefore, finally, the sum of all complication similarity factors corresponding to each sample data is subjected to negative correlation mapping to correct the logical relationship and obtain the complication impact degree of each sample data. The greater the complication impact degree, the greater the probability that the patient corresponding to the sample data may suffer from unexpected tissue damage or other complications during treatment, and the more severe the impact of the complications may be. The negative correlation mapping here can be performed using the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0054] The linkage feature analysis module 103 is used to determine the linkage feature performance between any two sample data by comprehensively considering the similarity features of treatment area feature parameters, the similarity features of tolerance sensitivity, and the similarity features of complication impact between sample data.

[0055] The treatment area characteristic parameters of the samples reflect key information such as the lesion site and treatment range. Therefore, incorporating the similarity of treatment area characteristic parameters into the analysis ensures that the impact of differences in treatment sites on treatment outcomes is fully considered when considering sample linkage. Individual differences exist in the tolerance sensitivity of the sample data; therefore, the similarity of tolerance sensitivity between sample data can reflect the differences in patient responses during treatment. Simultaneously, complications are an important factor that cannot be ignored during treatment, affecting the patient's recovery process and treatment effect. The similarity of the impact of complications between sample data can reflect the similarity of the risk and severity of complications among different patients during treatment. These three dimensions of similarity are interrelated and mutually influential. Therefore, in this embodiment of the invention, by comprehensively considering the similarity of these three dimensions, the degree of linkage between sample data can be determined more comprehensively and accurately, providing a more reliable reference for subsequent analysis.

[0056] Preferably, in one embodiment of the present invention, the method for obtaining the performance degree of linkage features includes: Please see Figure 3The diagram illustrates a method flowchart for obtaining the performance of linkage features in one embodiment of the present invention. The method includes the following steps: Step S301: Analyze the similarity features of the treatment area feature parameters between any two sample data to obtain the treatment area similarity between the two sample data.

[0057] The treatment area feature parameters include at least the distance from the skin to the target point and the thickness of the subcutaneous fat layer. Here, the treatment area feature parameters of each sample data for each treatment are combined into a feature parameter vector.

[0058] The similarity of treatment area feature parameters between sample data determines whether two sample data sets maintain similar structural features of their treatment areas. Therefore, for any two sample data sets, the Euclidean distance between any two feature parameter vectors is calculated. The smaller the Euclidean distance, the more similar the data. Therefore, the Euclidean distance is negatively correlated and normalized to correct the logical relationship, resulting in a treatment area similarity factor between the two sample data sets. At this point, there is a treatment area similarity factor between every two treatments, and the larger the value, the higher the structural feature similarity of the treatment areas between the two sample data sets. Therefore, the mean of all treatment area similarity factors between the two sample data sets is taken as the treatment area similarity. The negative correlation mapping and normalization here can be performed using the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0059] It should be noted that, for example, if sample data A has 3 treatment records, denoted as treatment 1, treatment 2, and treatment 3, and sample data B has 2 treatment records, denoted as treatment 4 and treatment 5, then there are a total of 6 treatment region similarity factors between sample data A and sample data B. That is, there is one treatment region similarity factor between treatment 1 and treatment 4, one between treatment 1 and treatment 5, one between treatment 2 and treatment 4, one between treatment 2 and treatment 5, one between treatment 4 and treatment 4, and one between treatment 3 and treatment 5.

[0060] Step S302: Analyze the similarity features of tolerance sensitivity between any two sample data to determine the tolerance sensitivity similarity between the two sample data.

[0061] The tolerance sensitivity of the sample data reflects the pain experience of the patients corresponding to the sample data during the treatment process. This indicator can serve as a key indicator reflecting the individual differences in the sample data. Different patients have different tolerance to treatment, and this difference will affect the formulation of treatment plans and treatment effects.

[0062] Therefore, for any two sample data, the absolute value of the difference in tolerance sensitivity between them is calculated. The smaller the absolute value of the difference, the higher the similarity in tolerance sensitivity between the two sample data. Thus, a negative correlation mapping and normalization are performed on the absolute value of the difference to obtain the tolerance sensitivity similarity between the two sample data. A higher similarity in tolerance sensitivity indicates that the patients corresponding to the two sample data have a relatively consistent physiological state and pain perception ability during piezoelectric shock wave treatment. This negative correlation mapping and normalization can be performed using the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0063] Step S303: Analyze the similarity characteristics of the impact of complications between any two sample data to determine the similarity of the impact of complications between the two sample data.

[0064] The complication impact of the sample data illustrates the probability of unexpected tissue damage and other complications that patients may suffer during treatment, as well as the degree of impact of these complications. As a key indicator reflecting individual differences in the sample data, the complication impact of different patients is affected by a variety of factors, such as treatment methods, underlying diseases, and physical conditions. Therefore, the differences in the risk and severity of complications that patients may experience during treatment will also affect the formulation of treatment plans and treatment outcomes.

[0065] Therefore, for any two sample data, the absolute value of the difference in the impact of complications between them is calculated. The smaller the absolute value of this difference, the higher the consistency of the impact of complications between the two sample data. Thus, this absolute value of the difference is negatively correlated and normalized to correct the logical relationship, obtaining the similarity of the impact of complications between the two sample data. The greater the similarity of the impact of complications, the higher the consistency of the abnormal physiological reactions and tissue damage reflected in the treatment process between the two sample data. The negative correlation mapping and normalization here can be performed using the formula... ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0066] Step S304: Use the treatment area similarity between any two sample data to perform weighted fusion of tolerance sensitivity similarity and complication impact similarity, thereby obtaining the linkage feature performance between the two sample data.

[0067] Based on the analysis in the preceding steps, the similarity of treatment intervals, tolerance sensitivity, and complication impact between any two sample data can be obtained. The similarity of treatment intervals reflects the similarity of the structural features of the treatment areas of the two sample data. The larger the value, the more attention needs to be paid to the similarity of the physiological state and pain perception of patients under piezoelectric shock wave therapy represented by the two sample data, so as to reasonably grasp the consistent direction of the overall treatment control parameters. Conversely, if the similarity of treatment intervals is smaller, it indicates that the similarity of the structural features of the treatment areas of the two sample data is weaker, and then it is necessary to more carefully evaluate the similarity of the patients' post-treatment complication risk represented by the sample data.

[0068] Therefore, here, the similarity of treatment areas between two sample data is used to weight the similarity of tolerance sensitivity (the similarity of treatment areas and the similarity of tolerance sensitivity are multiplied), and the difference between the constant 1 and the similarity of treatment areas is used to weight the similarity of complication impact (the difference between the constant 1 and the similarity of treatment areas is multiplied by the similarity of complication impact). This realizes the logic in the above analysis: when the similarity of treatment areas is greater, the proportion of tolerance sensitivity similarity is larger, and conversely, when the similarity of treatment areas is smaller, the proportion of complication impact similarity is larger. Finally, the weighted result (the sum of the two products mentioned above) is normalized and used as the linkage feature performance degree between the two sample data. The linkage feature performance degree uses the similarity characteristics of treatment areas to adjust the relative importance of tolerance sensitivity similarity and treatment area similarity, and integrates multiple indicators. Therefore, this value can more comprehensively reflect the overall similarity between sample data in the piezoelectric shock wave therapy process, and the larger the value, the higher the degree of similarity. Normalization is a technique well-known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0069] The treatment control parameter determination module 104 is used to train a neural network based on the linkage feature performance between sample data and the treatment control parameters, so as to determine the treatment control parameters of the sample to be tested.

[0070] The aforementioned module integrates various personalized differences in patients represented by the sample data, thereby analyzing the correlation feature performance between any two sample data. This can comprehensively and accurately reflect the overall similarity between sample data. Therefore, in this module, based on the correlation feature performance between sample data and the treatment control parameters of the sample data, a neural network can be trained to determine the treatment control parameters of the sample to be tested by leveraging the powerful learning and prediction capabilities of the neural network.

[0071] Preferably, in one embodiment of the present invention, a neural network is trained based on the correlation feature performance between sample data and treatment control parameters to determine the treatment control parameters of the sample to be tested, including: Cluster analysis is performed on all historical samples based on the correlation feature performance among them to obtain clusters. Specifically, K-means clustering algorithm with a preset optimal K value is used to cluster all historical sample data, resulting in clusters where historical samples within each cluster exhibit high similarity. The distance metric is the value obtained by negatively mapping the correlation feature performance among historical sample data; this negative correlation mapping can be expressed using the formula... or ,in, Let x represent an exponential function with the natural constant e as the base, and let x represent the independent variable.

[0072] Then, within each cluster, the mean of the linkage feature performance between each historical sample and other historical samples is used as the treatment performance feature value of each historical sample; and the treatment performance feature value of each historical sample is used as the input of the neural network. The mean of each historical sample under each treatment control parameter is calculated as the control parameter factor, and the control parameter factor of each historical sample under all kinds of treatment control parameters is used as the output, thereby training the neural network to obtain a trained neural network. At this time, each cluster obtained based on the clustering analysis of historical samples corresponds to a trained neural network.

[0073] Finally, among all the linkage feature performance values ​​corresponding to the test sample, the test sample is assigned to the cluster to which the historical sample with the highest linkage feature performance value belongs. The mean of the linkage feature performance values ​​between the test sample and all historical samples in the cluster is used as the input of the trained neural network corresponding to the cluster, thereby obtaining the specific value of each treatment control parameter corresponding to the test sample.

[0074] It should be noted that the preset optimal K value can be determined based on the silhouette coefficient method, and the silhouette coefficient method and K-means clustering algorithm are well-known techniques, and the specific process will not be elaborated here; the training process of the neural network is a well-known technique, and the specific process will not be elaborated here. The neural network can specifically use CNN or RNN, etc., and will not be limited or elaborated here.

[0075] In summary, the data acquisition module systematically collects sample data from patients with the same disease type. This sample data is divided into historical samples and samples to be tested, encompassing multi-dimensional information such as treatment feedback text, treatment area characteristic parameters, treatment control parameters, and pain test data. Due to individual patient differences, each person's pain tolerance and post-treatment feedback vary. Therefore, the sample data analysis module accurately determines the tolerance sensitivity of each sample by analyzing the changing trends and numerical similarities of pain test data within the sample data. Simultaneously, by comparing the similarity of treatment feedback text among the sample data, the impact of complications is precisely determined. This module accurately grasps the individual physiological and pathological characteristics of patients, providing crucial evidence for personalized treatment and contributing to improved treatment targeting and effectiveness. Furthermore, the linkage feature analysis module comprehensively considers the similarities in treatment area characteristic parameters, tolerance sensitivity, and complication impact among sample data to determine the linkage feature performance between any two sample data. Thus, by uncovering the inherent connections and similar patterns between different sample data, it helps to more rationally refer to historical samples in subsequent processes, providing a more scientific reference for setting treatment control parameters for the samples to be tested. Therefore, in the final treatment control parameter determination module, a neural network is trained based on the correlation characteristics between sample data and the treatment control parameters. Leveraging the powerful learning and prediction capabilities of the neural network, the treatment control parameters for the test sample can be accurately determined according to the individual characteristics of the patient and the correlation between sample data. Thus, this embodiment of the invention can fully consider individual differences and similarities between sample data, accurately set treatment control parameters, and effectively avoid overtreatment or undertreatment.

[0076] Please see Figure 4 This illustration shows a schematic diagram of the system structure of an autofocus-based depth-controllable piezoelectric shock wave therapy system according to an embodiment of the present invention. It includes a processor 400, a memory 401, a bus 402, and a communication interface 403. The processor 400, communication interface 403, and memory 401 are connected via the bus 402. The memory 401 may include a high-speed random access memory, and the bus 402 may be an ISA bus, PCI bus, or EISA bus, etc. The processor 400 may be an integrated circuit chip with signal processing capabilities. The memory 401 stores at least one instruction, at least one program, code set, or instruction set. When the processor loads and executes the at least one instruction, at least one program, code set, or instruction set, it implements the steps of each module in the autofocus-based depth-controllable piezoelectric shock wave therapy system.

[0077] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0078] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0079] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A depth-controllable piezoelectric shock wave therapy system based on automatic focusing, characterized in that, The system includes: The data acquisition module is used to acquire sample data from patients with the same disease type. The sample data includes historical samples and samples to be tested. Each sample data includes the patient's treatment feedback text, treatment area feature parameters, treatment control parameters, and the patient's pain test data for each treatment. The sample data analysis module is used to analyze the changing trends and numerical similarities of pain test data in the sample data to obtain the tolerance sensitivity of each sample data; within each sample data, the similarity between treatment feedback texts is compared to determine the complication impact of each sample data. The linkage feature analysis module is used to comprehensively analyze the similarity features of treatment area feature parameters, tolerance sensitivity, and complication impact between sample data to determine the linkage feature performance between any two sample data. The treatment control parameter determination module is used to train a neural network based on the correlation feature performance between sample data and the treatment control parameters, thereby determining the treatment control parameters of the sample to be tested.

2. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 1, characterized in that, The method for obtaining tolerance sensitivity includes: The pain test data includes subjective pain scores and electromyographic signal data at each test energy level, with the test energy gradually increasing. In each sample data, the subjective pain sensitivity of each sample data is determined based on the changing trend of subjective pain scores; By analyzing the similarity and numerical characteristics of electromyographic signals among the sample data, the objective pain sensitivity of each sample data was determined. The normalized value of the product of subjective pain sensitivity and objective pain sensitivity for each sample is taken as the tolerance sensitivity for each sample.

3. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 2, characterized in that, The methods for obtaining the subjective pain sensitivity include: In each sample data, the subjective pain scores are sorted in ascending order of test energy to obtain a sorting sequence; In the sorting sequence, the ratio of each subjective pain score to the adjacent previous pain score is calculated as the perceived pain enhancement for each subjective pain score. In each sample data, the subjective pain score is weighted and fused using the pain enhancement perception of each subjective pain score, and the normalized value of the weighted result is used as the subjective pain sensitivity of each sample data.

4. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 2, characterized in that, The methods for obtaining the objective pain sensitivity include: In the pain test data of each sample data, the absolute area of ​​the electromyographic signal data and the horizontal axis under each test energy is used as the electromyographic energy characterization value; For any two sample data, analyze the differences in electromyographic energy characterization values ​​of the two sample data under the same test energy, and combine the test energy numerical characteristics to obtain the objective stimulus performance difference between the two sample data. Cluster analysis was performed on all sample data based on the objective stimulus performance differences among the sample data to obtain all clusters; Within each cluster, the mean value of the peak values ​​of the electromyography signals of all sample data at all test energies is normalized and used as the objective pain sensitivity of each sample data in each cluster.

5. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 4, characterized in that, The methods for obtaining the objective stimulus performance differences include: For any two sample data, under the same test energy, the absolute value of the difference between the electromyographic energy characterization values ​​of the two sample data is taken as the electromyographic difference characteristic value of the two sample data under the same test energy. The ratio of each test energy to the sum of all test energies is used as the stimulus intensity weight for each test energy; For any two sample data, the electromyographic difference feature values ​​of the two sample data under the same test energy are weighted and fused using the stimulus intensity weight of the test energy, and the normalized value of the weighted result is used as the objective stimulus performance difference between the two sample data.

6. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 1, characterized in that, The methods for obtaining the impact of the complications include: For each sample data point, complication keywords were extracted from the treatment feedback text at each treatment session to obtain the complication keyword set for each treatment session. In each sample data, the Jaccard correlation coefficient between the complication keyword set at the last treatment and the complication keyword set at each previous treatment is calculated as the complication similarity factor; The sum of all complication similarity factors corresponding to each sample data is negatively correlated and mapped to the value, which is then used as the complication impact of each sample data.

7. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 1, characterized in that, The method for obtaining the expressiveness of the linkage features includes: Analyze the similarity features of the treatment area characteristic parameters between any two sample data to obtain the treatment area similarity between the two sample data; The absolute value of the difference in tolerance sensitivity between the two sample data is negatively correlated and normalized, and this value is taken as the tolerance sensitivity similarity between the two sample data. The absolute value of the difference in the impact of complications between the two sample data is negatively correlated and normalized, and this value is taken as the similarity of the impact of complications between the two sample data. The similarity of the treatment area between the two sample data is used to weight the similarity of tolerance sensitivity, and the difference between the treatment area similarity and the constant 1 is used to weight the similarity of complication impact. The weighted result is normalized and used as the linkage feature performance between the two sample data.

8. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 7, characterized in that, The method for obtaining the similarity of the treatment areas includes: The treatment area feature parameters include at least the distance from the skin to the target point and the thickness of the subcutaneous fat layer. The treatment area feature parameters of each sample data for each treatment are combined into a feature parameter vector. For any two sample data, calculate the Euclidean distance between any two feature parameter vectors between the two sample data, perform negative correlation mapping and normalization to obtain the treatment area similarity factor between the two sample data, and take the mean of all treatment area similarity factors between the two sample data as the treatment area similarity.

9. The depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 1, characterized in that, The neural network trained based on the linkage feature representation degree between sample data and treatment control parameters is used to determine the treatment control parameters of the sample to be tested, including: Cluster analysis is performed on all historical samples based on the correlation characteristics among all historical samples to obtain clusters; Within each cluster, the mean of the linkage feature performance between each historical sample and other historical samples is used as the treatment performance feature value of each historical sample. In each cluster, the treatment performance feature value of each historical sample is used as the input of the neural network. The mean value of each historical sample under each treatment control parameter is calculated as the control parameter factor. The control parameter factor of each historical sample under all kinds of treatment control parameters is used as the output, thereby training the neural network and obtaining a trained neural network. The test sample is assigned to the cluster of the historical sample with the highest linkage feature performance. The mean of the linkage feature performance between the test sample and all historical samples in the cluster is used as the input of the trained neural network corresponding to the cluster, thereby obtaining the value of each treatment control parameter for the test sample.

10. A depth-controllable piezoelectric shock wave therapy system based on automatic focusing according to claim 9, characterized in that, The cluster analysis is performed on all historical samples based on the correlation feature performance among all historical samples to obtain clusters, including: In all historical samples, cluster analysis is performed on all historical sample data based on the K-means clustering algorithm and a preset optimal K value to obtain all clusters. The distance metric is the value after negatively mapping the correlation characteristics between historical sample data.

Images