A human-machine collaborative high-precision acupuncture system based on digital twins

Abstract

The present invention discloses a human-machine collaborative high-precision acupuncture system based on digital twins, which relates to the field of acupuncture technology, including: obtaining user surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; generating a three-dimensional point cloud model based on the surface image data, mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; dynamically correcting the initial acupoint coordinates based on real-time electrophysiological data to obtain individualized acupoint coordinates; constructing a layered tissue model containing nerve distribution based on local subcutaneous tissue structure data; its beneficial effects are: realizing dynamic correction of acupoints by fusing three-dimensional point cloud modeling with real-time electrophysiological feedback, generating accurate acupuncture paths by combining nerve activation probability prediction, and performing real-time path adjustment based on physiological feedback, which has the advantages of improving acupoint positioning accuracy, realizing dynamic optimization of acupuncture paths and improving treatment safety.

Description

Technical Field

[0001] The present invention relates to the field of acupuncture technology, and in particular to a human-machine collaborative high-precision acupuncture system based on digital twins. Background Art

[0002] With the rapid development of digital medicine and human-computer collaboration technology, acupuncture, as an important treatment method in traditional Chinese medicine, is gradually integrating information and intelligent means to achieve refined and personalized upgrades. Traditional acupuncture methods mainly rely on experienced practitioners to complete acupoint positioning and acupuncture operations based on standard meridian maps and palpation techniques. Although it has certain therapeutic effects, due to the lack of objective quantification and precise control, there are still many technical bottlenecks in treatment consistency, safety and intelligent execution.

[0003] In traditional acupuncture practice, physicians usually need to rely on visual observation and palpation experience to determine the location of acupoints. This method is easily affected by individual differences in body surface morphology, resulting in insufficient accuracy in acupoint positioning. At the same time, due to the lack of precise modeling of subcutaneous tissue structure and nerve distribution, acupuncture operations often find it difficult to accurately activate the target nerve area, affecting the treatment effect. In addition, traditional methods cannot monitor the patient's physiological reactions in real time, making it difficult to adjust acupuncture parameters in a timely manner, and there are certain safety risks.

[0004] In this context, the introduction of human-machine collaboration technology in medical scenarios provides a new development path for achieving personalized, high-precision acupuncture. Human-machine collaboration can enable the system to have interactive perception, autonomous decision-making and dynamic adjustment capabilities with patients by integrating multimodal perception, human factors modeling and intelligent execution control. However, the current acupuncture technology based on human-machine collaboration still has significant shortcomings: First, most existing solutions are based on standard meridian models, lacking accurate modeling of the user's actual body surface morphology and nerve distribution, making it difficult to ensure the personalized accuracy of acupoint positioning; second, the generation of acupuncture paths has not yet effectively combined multi-dimensional data of tissue deformation, stress distribution and nerve activation probability, making it difficult to accurately activate the target nerve area; finally, during the acupuncture process, there is a lack of real-time collection of physiological feedback such as patient pain signals and facial expressions and adaptive adjustment of the path, affecting safety and comfort.

[0005] Therefore, a human-machine collaborative high-precision acupuncture system based on digital twin is proposed. Summary of the Invention

[0006] In view of the above-mentioned state of the art, this application is proposed. The embodiments of this application provide a human-machine collaborative high-precision acupuncture system based on digital twins, which can improve the accuracy of acupoint positioning, realize dynamic optimization of acupuncture paths, and improve treatment safety.

[0007] According to one aspect of the present application, a human-machine collaborative high-precision acupuncture method based on digital twins is provided, including: obtaining a user's body surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; generating a three-dimensional point cloud model based on the body surface image data, and mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; dynamically correcting the initial acupoint coordinates based on the real-time electrophysiological data to obtain corrected individualized acupoint coordinates; constructing a layered tissue model containing nerve distribution based on the local subcutaneous tissue structure data, and simulating the acupuncture angle and depth in combination with preset tissue elasticity parameters to generate a neural activation probability prediction value; generating three-dimensional acupuncture path data including starting point coordinates, direction vector and target depth based on the corrected individualized acupoint coordinates and the neural activation probability prediction value; converting the three-dimensional acupuncture path data into actuator control instructions, and obtaining the user's physiological feedback data in real time during the acupuncture process; and dynamically adjusting the three-dimensional acupuncture path data according to the physiological feedback data.

[0008] According to another aspect of the present application, a human-machine collaborative high-precision acupuncture system based on digital twin is provided, comprising: a data acquisition module for acquiring the user's body surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; a model construction module for generating a three-dimensional point cloud model based on the body surface image data, and mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; an acupoint correction module for dynamically correcting the initial acupoint coordinates based on the real-time electrophysiological data to obtain corrected individualized acupoint coordinates; an acupuncture simulation module for dynamically correcting the initial acupoint coordinates based on the local subcutaneous tissue structure data to obtain corrected individualized acupoint coordinates; and acupuncture simulation module for dynamically correcting the initial acupoint coordinates based on the local subcutaneous tissue structure data to obtain corrected individualized acupoint coordinates. A hierarchical tissue model including nerve distribution is constructed, and the acupuncture angle and depth are simulated in combination with preset tissue elasticity parameters to generate a neural activation probability prediction value; an acupuncture path generation module is used to generate three-dimensional acupuncture path data including starting point coordinates, direction vector and target depth based on the corrected individualized acupoint coordinates and the neural activation probability prediction value; an instruction generation module is used to convert the three-dimensional acupuncture path data into an actuator control instruction; an acupuncture path adjustment module is used to obtain the user's physiological feedback data during the acupuncture process when the actuator executes the control instruction, and dynamically adjust the three-dimensional acupuncture path data according to the physiological feedback data.

[0009] According to another aspect of the present application, an electronic device is provided, comprising a memory and a processor, wherein the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which implement the steps of the system described above when executed by the processor.

[0010] According to another aspect of the present application, a computer storage medium is provided, on which computer executable instructions are stored. When the computer executable instructions are executed by a processor, the steps of the system described above are implemented.

[0011] Compared with the existing technology, the human-machine collaborative high-precision acupuncture system based on digital twins according to the embodiment of the present application can realize dynamic correction of acupoints by integrating three-dimensional point cloud modeling and real-time electrophysiological feedback, generate precise acupuncture paths by combining nerve activation probability prediction, and perform real-time path adjustment based on physiological feedback. It has the advantages of improving acupoint positioning accuracy, realizing dynamic optimization of acupuncture paths and improving treatment safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0013] Figure 1 This is a flow chart of the present invention's human-machine collaborative high-precision acupuncture method based on digital twins.

[0014] Figure 2 This is a block diagram of a human-machine collaborative high-precision acupuncture system based on digital twins of the present invention.

[0015] Figure 3 The present invention is a block diagram of an electronic device. DETAILED DESCRIPTION

[0016] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0017] Application Overview

[0018] Traditional acupuncture methods rely primarily on experienced practitioners to locate acupoints and perform needle insertion according to standard meridian maps and palpation techniques. While these methods offer some efficacy, they lack objective quantification and precise control, resulting in numerous technical bottlenecks in treatment consistency, safety, and intelligent execution. For example, during acupuncture treatment, a patient experienced deviations from the standard meridian model due to individual body morphological differences, resulting in the needle path failing to effectively activate the target nerve area. Furthermore, the lack of real-time physiological feedback and adjustments resulted in discomfort during the treatment.

[0019] To address these issues, it is necessary to resolve the spatial matching errors between the standard model and the actual body surface morphology, establish a dynamic acupoint positioning mechanism, integrate tissue mechanical properties with nerve distribution data to optimize acupuncture paths, and build a real-time feedback system based on physiological signals. Analysis has shown that building a personalized digital twin model based on multimodal sensory data can improve spatial matching accuracy, and combining it with real-time electrophysiological data can achieve dynamic acupoint calibration. Finite element simulation technology can predict the probability of nerve activation under different acupuncture parameters, and closed-loop regulation through physiological feedback can improve treatment safety.

[0020] Exemplary Methods

[0021] Figure 1 The diagram illustrates a human-machine collaborative high-precision acupuncture method based on digital twins according to an embodiment of the present application, including: obtaining the user's body surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; generating a three-dimensional point cloud model based on the body surface image data, and mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; dynamically correcting the initial acupoint coordinates based on real-time electrophysiological data to obtain corrected individualized acupoint coordinates; constructing a layered tissue model containing nerve distribution based on local subcutaneous tissue structure data, and simulating the acupuncture angle and depth in combination with preset tissue elasticity parameters to generate a neural activation probability prediction value; generating three-dimensional acupuncture path data including starting point coordinates, direction vector and target depth based on the corrected individualized acupoint coordinates and the neural activation probability prediction value; converting the three-dimensional acupuncture path data into actuator control instructions; obtaining the user's physiological feedback data during the acupuncture process during the actuator's execution of the control instructions, and dynamically adjusting the three-dimensional acupuncture path data based on the physiological feedback data.

[0022] Among them, body surface image data refers to the geometric morphological information of the skin surface obtained by three-dimensional scanning equipment, which can be specifically achieved by laser scanning or structured light imaging technology to construct the spatial topological structure of the patient's body surface.

[0023] Among them, the three-dimensional point cloud model refers to a digital model that converts a two-dimensional image sequence into a set of spatial coordinates through a three-dimensional reconstruction algorithm. Specifically, it can be implemented using point cloud registration and surface reconstruction technology to carry out spatial mapping of anatomical structure features.

[0024] Among them, real-time electrophysiological data refers to the skin surface impedance or potential change signal collected by the bioelectrode array, which can be specifically achieved by using a multi-channel bioelectric detection device to reflect the bioelectric activity characteristics of the acupoint area.

[0025] Among them, the hierarchical tissue model refers to a three-dimensional structural model of subcutaneous tissue established through medical image segmentation technology, which can be specifically implemented using CT or MRI image hierarchical reconstruction technology to characterize the elastic modulus distribution of tissues at different depths.

[0026] Among them, the predicted value of nerve activation probability refers to the evaluation index of the possibility of nerve fibers being activated by mechanical stimulation obtained through mechanical simulation calculation. It can be specifically achieved by finite element analysis and stress field mapping algorithm to quantify the degree of impact of acupuncture operation on the target nerve area.

[0027] Specifically, a 3D scanning device is used to capture patient surface image data. A 3D digital model is generated through point cloud registration and surface reconstruction. Anatomical fiducials in the standard meridian model are matched to patient surface landmarks, and a preliminary mapping of acupoint coordinates is achieved using a spatial transformation matrix. A biopotential sensing electrode array is placed within the initial coordinate region. The peak location of the electrophysiological response intensity is identified based on real-time impedance change signals. A gradient descent algorithm is used to dynamically offset the initial coordinates. A hierarchical tissue model, including the distribution of nerve bundles, is constructed based on the medical imaging data. The stress field distribution for different acupuncture angle and depth combinations is calculated using the finite element method. The stress data is compared with preset neural activation thresholds, and activation probabilities corresponding to each parameter combination are output. The corrected acupuncture point coordinates are used to determine the acupuncture starting point. The direction vector and target depth are optimized based on the neural activation probability. A collision detection algorithm is used to eliminate path plans that pose a risk of tissue penetration. During the acupuncture procedure, real-time physiological feedback data is obtained through heart rate monitoring and facial expression recognition. When abnormal pain signals are detected, the preset acupuncture depth is automatically shortened and the path replanned.

[0028] Compared with existing technologies, existing solutions mostly use static standard meridian models for acupoint positioning, without considering individual body surface morphological differences and real-time bioelectric characteristic changes, resulting in systematic deviations in acupoint positioning. This solution combines a three-dimensional point cloud model with dynamic electrophysiological correction to construct an adaptive acupoint positioning mechanism, effectively eliminating the coordinate offset error caused by body surface morphological differences. Traditional acupuncture path planning relies only on empirical depth parameters. This solution achieves acupuncture parameter optimization based on biomechanical characteristics by establishing a hierarchical tissue mechanics model and a neural activation probability prediction model. Existing technologies lack real-time monitoring and feedback regulation of physiological status during operation. This solution constructs a closed-loop control system by fusing multi-source physiological signals, significantly improving the safety of the treatment process.

[0029] Through the above technical solutions, the present application can achieve precise acupoint positioning based on individual anatomical characteristics, eliminating the spatial mapping error between the standard model and the actual body surface; by combining tissue mechanical properties and nerve activation probability prediction, the acupuncture angle and depth parameter combination is optimized to improve the effective activation rate of the target nerve area; through real-time physiological feedback and path dynamic adjustment mechanism, the operating parameters can be corrected in time when abnormal pain reactions are detected, reducing the risk of tissue damage during treatment.

[0030] The present application further proposes a method for mapping standard meridian model data to a three-dimensional point cloud model to generate initial acupoint coordinates, including marking anatomical landmark points on the surface image data, and marking anatomical reference points and acupoint reference coordinates on the standard meridian model; aligning the anatomical landmark points in the surface image data with the anatomical reference points in the standard meridian model through a feature point matching algorithm to obtain an initial transformation matrix; mapping the acupoint reference coordinates in the standard meridian model to the three-dimensional point cloud model based on the initial transformation matrix; using an iterative closest point algorithm to optimize the spatial alignment error between the three-dimensional point cloud model and the standard meridian model to generate a corrected transformation matrix; adjusting the acupoint reference coordinates according to the corrected transformation matrix to obtain the initial acupoint coordinates.

[0031] Among them, anatomical landmark annotation refers to marking points with fixed anatomical significance in human body surface images. It can be achieved by combining image recognition algorithms with manual calibration to establish the correspondence between the body surface and the standard model.

[0032] Among them, the feature point matching algorithm refers to the technology of spatial alignment by extracting image feature points. Specifically, it can be implemented using a scale-invariant feature transformation algorithm or an accelerated robust feature algorithm to solve the spatial position deviation between the surface anatomical landmark points and the standard model reference points.

[0033] Among them, the iterative closest point algorithm refers to an optimization method that minimizes the distance between point clouds through iterative calculation. Specifically, it can be implemented using a rigid registration algorithm based on singular value decomposition to eliminate the nonlinear deformation error between the three-dimensional point cloud model and the standard meridian model.

[0034] Specifically, after the surface image data is three-dimensionally reconstructed to generate a point cloud model, it needs to be spatially aligned with the standard meridian model. First, the anatomical landmarks are marked in the surface image, and the corresponding reference points are marked in the standard model. The correspondence between the two point sets is established through the feature point matching algorithm, and the initial transformation matrix is ​​calculated to preliminarily map the acupoint coordinates in the standard model to the surface point cloud. Since individual surface morphological differences may cause errors in the initial mapping, the iterative nearest point algorithm is further used to perform non-rigid registration of the point cloud model and the standard model. Through multiple iterations to optimize the transformation matrix, the acupoint reference coordinates are finally made to accurately fit the three-dimensional structure of the individual's body surface.

[0035] Compared with existing technologies, traditional methods only rely on standard meridian models for acupoint positioning, without considering the impact of surface morphology differences on mapping accuracy. However, this scheme combines feature point matching with iterative optimization to achieve dynamic adaptation to individual surface characteristics while retaining the acupoint distribution law of the standard model, thus solving the problem of acupoint coordinate offset caused by differences in surface convexity and concavity.

[0036] Through the above technical solution, the present application can effectively eliminate the morphological differences between the standard meridian model and the individual body surface, making the acupoint positioning results more in line with the actual anatomical structure, avoiding the coordinate mapping error caused by changes in the body surface curvature, and providing a high-precision initial position reference for subsequent dynamic correction.

[0037] The present application further proposes a dynamic correction of the initial acupoint coordinates, including: calculating the electrophysiological response intensity around the initial acupoint coordinates, and comparing it with a preset conductivity threshold; if the electrophysiological response intensity exceeds the preset conductivity threshold, the initial acupoint coordinates are offset toward the peak direction of the response intensity until the electrophysiological response intensity reaches a preset confidence range of the maximum value.

[0038] Among them, the electrophysiological response intensity refers to the conductivity measurement value of the local area collected by the electrophysiological sensor, which can be specifically achieved by using a multi-channel bioelectric signal acquisition device to characterize the meridian activity level of the target area.

[0039] Among them, the preset conductivity threshold refers to the acupoint conductivity benchmark range established based on historical data statistics, which can be established through statistical analysis of clinical samples and used to screen candidate areas with significant conductivity abnormalities.

[0040] The peak direction of the response intensity refers to the vector direction of the conductivity gradient rising when radiating outward from the initial coordinates. Specifically, the potential distribution gradient can be calculated by measuring the annular electrode array to determine the spatial offset trend of the actual position of the acupoint.

[0041] The preset confidence range refers to the error range allowed for fluctuations in the conductivity measurement value, which can be specifically set through a probability distribution model and is used to determine the termination condition of the dynamic correction process.

[0042] Specifically, the dynamic correction process first measures the conductivity distribution of the area surrounding the initial acupoint coordinates using a ring electrode array to create a three-dimensional conductivity heat map. When the conductivity value in a certain direction is detected to continuously exceed a preset threshold, the system generates a potential gradient vector field for that area and gradually adjusts the coordinate position along the ascending gradient. During this adjustment process, the peak conductivity change is monitored in real time. The correction is considered complete when the measured value enters a predetermined confidence interval and remains stable. For example, in the localization of the Zusanli acupoint, if an abnormally high conductivity area appears outside the initial coordinates, the system will drive the coordinates to shift along the potential gradient until a stable and significant conductivity peak is obtained.

[0043] Compared to existing technologies, traditional acupoint localization methods rely solely on static anatomical landmark registration and fail to reflect real-time changes in an individual's physiological state. This solution, through dynamic electrophysiological monitoring and an adaptive coordinate offset mechanism, can capture acupoint position deviations caused by muscle contraction and changes in body fluid distribution, achieving truly personalized positioning. For example, in scenarios where body fluid distribution changes after exercise, this technology can effectively correct acupoint position deviations caused by tissue swelling.

[0044] Through the above-mentioned technical solution, this application solves the technical problems of static acupoint positioning and lack of physiological adaptability in existing technologies, significantly improving the physiological relevance and individualized accuracy of acupoint positioning. This solution can accurately identify acupoint position deviations caused by individual differences or changes in physiological status, ensuring that acupuncture procedures always act on the optimal bioelectrically active area, thereby improving treatment efficacy and reducing the risk of mis-puncture.

[0045] This application further proposes a method for simulating acupuncture angles and depths, specifically including: discretizing the layered tissue model into grid units based on the finite element analysis method; calculating the relationship between acupuncture force and tissue deformation according to preset tissue elastic parameters, and obtaining needle tip displacement and tissue stress distribution data under different acupuncture angle and depth combinations; mapping the tissue stress distribution data to the nerve area based on the preset nerve distribution area, and calculating the stress concentration intensity in the area; and generating a predicted value of the nerve activation probability under the corresponding acupuncture angle and depth combination based on the stress concentration intensity.

[0046] Among them, the finite element analysis method refers to a numerical calculation method that discretizes a continuous physical model into a finite number of grid units for mechanical analysis. Specifically, it can be implemented using three-dimensional tetrahedral or hexahedral grid division tools to simulate the interaction process between the needle and biological tissue.

[0047] Among them, the preset tissue elastic parameters refer to the mechanical parameters that characterize the elastic characteristics of different tissue layers such as skin, muscle and fascia. They can be obtained through tissue stretching experiments or existing biomechanical databases and used to establish a quantitative relationship between needle tip displacement and tissue stress distribution.

[0048] Among them, stress concentration intensity refers to the cumulative result of tissue stress values ​​per unit volume, which can be calculated using the stress integration algorithm and is used to characterize the mechanical stimulation intensity of the needle body on the nerve endings.

[0049] Among them, the predicted value of nerve activation probability refers to a quantitative indicator generated based on the comparison between stress concentration intensity and nerve excitation threshold. It can be implemented using logistic regression or support vector machine models to screen acupuncture parameter combinations that effectively activate nerves.

[0050] Specifically, the layered tissue model is first discretized into grid units with physical properties, such as using tetrahedral units to mesh the skin layer and muscle layer. During the acupuncture simulation process, the mechanical response relationship between the needle tip and each tissue layer is established based on the preset elastic parameters, and the displacement trajectory of the needle tip and the stress distribution of the surrounding tissue at different angles and depths are calculated by the finite element solver. Subsequently, the stress distribution data is mapped to the preset nerve distribution area, for example, the stress value is transferred to the neural model node through a spatial interpolation algorithm, and the stress accumulation in each neural unit is counted. When the stress concentration intensity in a certain area exceeds the preset threshold, it is determined that there is a possibility of activation of the nerve endings in the area, and the probability model outputs the neural activation probability value corresponding to the acupuncture parameters, for example, the stress intensity and the probability value are nonlinearly mapped to generate a prediction result in the range of 0-1.

[0051] Compared to existing technologies, traditional acupuncture simulation methods typically only consider the mechanical response of a single tissue layer and fail to integrate activation probability predictions with the nerve distribution region. This results in acupuncture parameter selection lacking bioelectrical signal correlation. For example, existing approaches may only determine needle insertion depth based on tissue deformation, without quantitatively analyzing the actual stimulating effect of stress distribution on nerve endings. This approach, by constructing a layered model that incorporates nerve distribution and combining elastic parameters with finite element analysis, achieves modeling of the correlation between acupuncture angle, depth, and nerve activation probability. This addresses the problem of traditional methods relying on empirical judgment and failing to predict neural responses.

[0052] Through the above technical solution, this application can accurately predict the activation probability of nerve endings under different acupuncture parameter combinations, and provide data support for the clinical selection of effective stimulation parameters. By combining tissue elasticity and nerve distribution characteristics, ineffective stimulation caused by acupuncture angle deviation or insufficient depth can be avoided, while reducing the risk of nerve damage caused by excessive insertion. For example, in the acupuncture operation of the Zusanli acupoint, the system can screen out the needle insertion path that can activate the target nerve and conform to the tissue mechanics safety range, thereby improving the accuracy and safety of treatment.

[0053] The present application further proposes a method for generating three-dimensional acupuncture path data including starting point coordinates, direction vectors and target depths, including: setting the corrected individualized acupoint coordinates as the starting point coordinates of the acupuncture path; calculating the direction vector based on the starting point coordinates pointing to the central area of ​​nerve distribution at the corresponding acupoint in the hierarchical tissue model; selecting the acupuncture depth that meets the preset activation threshold as the target depth based on the predicted value of the nerve activation probability; generating acupuncture trajectory data in three-dimensional space according to the direction vector and the target depth, and performing collision detection on the trajectory data with the three-dimensional point cloud model, eliminating trajectory paths with the risk of tissue penetration, and obtaining three-dimensional acupuncture path data.

[0054] Among them, the corrected individualized acupoint coordinates refer to the acupoint position coordinates that are dynamically adjusted through real-time electrophysiological data. Specifically, they can be achieved by combining the conductivity peak detection algorithm with the coordinate offset compensation algorithm to eliminate the acupoint positioning deviation caused by differences in the user's body surface morphology.

[0055] Among them, the direction vector refers to the unit vector in three-dimensional space pointing from the starting point to the central area of ​​nerve distribution. It can be achieved through a three-dimensional vector projection algorithm combined with nerve distribution density gradient calculation to ensure that the acupuncture direction accurately points to the target nerve area.

[0056] Among them, the neural activation probability prediction value refers to a quantitative indicator of neural response intensity generated based on the tissue stress distribution simulation results. Specifically, it can be implemented by combining the stress-neural response mapping model with the probability density function calculation to screen the numerical range of the effective stimulation depth.

[0057] Among them, collision detection refers to the interference analysis between three-dimensional space trajectories and anatomical structures. It can be implemented by combining a ray casting algorithm with a spatial grid collision detection algorithm to identify and eliminate abnormal paths that may penetrate blood vessels or bones.

[0058] Specifically, this method first uses the dynamically corrected acupuncture point coordinates as the starting point for acupuncture, and determines the optimal needle insertion direction by analyzing the spatial density gradient of nerve distribution in the layered tissue model. Based on the nerve activation probability curve obtained by finite element simulation, the maximum effective depth at which the activation probability exceeds the preset threshold is selected as the target value. When generating a three-dimensional trajectory, multi-dimensional collision detection is performed in conjunction with a high-precision point cloud model. Spatial geometric operations are used to identify potential intersections with blood vessels, bones, or other sensitive tissues in the trajectory path, automatically eliminating path plans with penetration risks and retaining three-dimensional acupuncture path data that meets safety standards.

[0059] Compared with existing technologies, traditional acupuncture methods typically use fixed acupoint coordinates and empirical angles for acupuncture. These methods lack a coordinate correction mechanism based on dynamic electrophysiological feedback and are unable to optimize needle insertion depth based on neural activation probability. Existing collision detection is often limited to two-dimensional planar analysis, making it difficult to effectively identify tissue penetration risks in three dimensions. This solution ensures positioning accuracy by dynamically correcting individualized acupoint coordinates, combines neural activation probability thresholds to screen for effective stimulation depth, and utilizes a three-dimensional collision detection mechanism to verify path safety, forming a closed-loop optimized path generation mechanism.

[0060] Through the above technical solution, this application can realize the automated generation of high-precision three-dimensional acupuncture paths, effectively eliminating positioning errors caused by individual anatomical differences, and ensuring that the acupuncture direction is accurately pointed to the target nerve area. By combining the neural activation probability threshold to screen the optimal needle insertion depth, ineffective stimulation caused by insufficient depth or tissue damage caused by excessive penetration can be avoided. The three-dimensional collision detection mechanism can actively identify and avoid blood vessels and bone structures, significantly reducing the risk of acupuncture operations and improving the safety of the treatment process.

[0061] This application further proposes to obtain the user's physiological feedback data in real time during the acupuncture process and dynamically adjust the three-dimensional acupuncture path data based on the data, where the physiological feedback data includes heart rate variability data and facial expression recognition data. The dynamic adjustment process is specifically manifested as follows: when the heart rate variability data is lower than the preset pain threshold or the facial expression recognition data matches the preset pain characteristics, the acupuncture path data with a shortened target depth is regenerated.

[0062] Among them, heart rate variability data refers to the characteristic data of heart rhythm changes collected in real time by wearable devices. Specifically, it can be achieved by using an electrocardiogram sensor combined with a time-frequency domain analysis method, and is used to reflect the degree of stress response of the autonomic nervous system to acupuncture stimulation. Facial expression recognition data refers to the characteristic data of facial muscle movement captured by a visual sensor. Specifically, it can be achieved by using a three-dimensional dynamic capture camera combined with a convolutional neural network model, and is used to extract micro-expression characteristic patterns related to pain. The preset pain threshold refers to the pain intensity grading standard established through clinical experiments. Specifically, it can be generated by training historical pain case data using a machine learning classifier. The acupuncture path data for shortening the target depth refers to reducing the original depth value by a preset step size through a path planning algorithm. Specifically, it can be achieved by using an iterative optimization algorithm based on gradient descent, and verifying whether the adjusted path meets the neural activation probability requirements in the simulation model.

[0063] Specifically, during the acupuncture phase, the heart rate variability monitoring module acquires ECG signals at a sampling rate of once per second and calculates the standard deviation of adjacent R-wave intervals to obtain an index of autonomic nerve activation intensity. Simultaneously, a binocular camera captures facial images at a rate of 30 frames per second, and the expression feature extraction network outputs a pain probability value. If the heart rate variability index monitors that it falls below a preset pain threshold for three consecutive sampling cycles, or if the pain feature match output by the expression recognition model exceeds a set threshold, the control algorithm immediately triggers the acupuncture path adjustment process. During this process, the path optimizer gradually shortens the target depth by a preset step size (e.g., 0.5 mm) based on the spatial relationship between the current acupuncture depth and the target nerve area, and recalculates the nerve activation probability through finite element simulation. If the adjusted depth still maintains a valid activation probability within the allowable range, a new control instruction is generated to drive the actuator.

[0064] Compared with existing technologies, traditional acupuncture pain feedback mainly relies on the patient's subjective description or single physiological indicator monitoring, which has the risk of response delay and misjudgment. The existing technology that uses electromyographic signal monitoring cannot distinguish between pain and normal acupuncture reactions, and contact detection based on pressure sensors will interfere with the operation process. This solution uses non-contact multimodal data fusion to achieve real-time capture and quantitative analysis of pain signals while maintaining the continuity of the acupuncture operation. In addition, existing path adjustment methods mostly use a fixed depth attenuation strategy, while this solution combines a neural activation probability prediction model to perform dynamic step size optimization, which not only ensures the treatment effect but also avoids excessive adjustment.

[0065] Through the above technical solution, this application effectively solves the problem of overstimulation caused by individual differences in pain sensitivity during acupuncture. Through cross-validation of multi-dimensional physiological feedback data, the accuracy and response speed of pain recognition are significantly improved. This solution can quickly generate a safe path adjustment plan when an abnormal pain response occurs, avoiding the risk of tissue damage caused by delayed manual judgment in traditional methods, while maintaining effective stimulation intensity to the target nerve area.

[0066] The present application further proposes to obtain in real time the needle resistance during the execution of the control instruction by the actuator, and to stop the execution of the control instruction when the needle resistance exceeds a preset resistance threshold.

[0067] Among them, needle body resistance refers to the mechanical feedback signal received when the needle contacts human tissue during the acupuncture operation. Specifically, it can be achieved by using a pressure sensor or a torque sensor to collect the needle's axial propulsion force and rotational resistance data in real time, which is used to dynamically determine whether abnormal tissue resistance is encountered during acupuncture. The preset resistance threshold refers to the mechanical safety boundary value set according to different tissue types and acupoint areas. Specifically, it can be achieved by establishing a multi-level threshold library based on statistical data of elastic modulus and nerve distribution density of different tissues in clinical experiments, which is used to identify abnormal puncture conditions that may cause tissue damage. The execution of the stop control instruction refers to immediately cutting off the power output of the actuator when it is detected that it exceeds the safety range. Specifically, it can be achieved by using an embedded controller to trigger an emergency stop signal and start a reverse backoff mechanism to avoid tissue penetration or nerve damage caused by abnormal resistance.

[0068] Specifically, during the acupuncture execution phase, the sensor integrated into the needle handle collects needle resistance data in real time and dynamically compares it with the preset resistance threshold. For example, during the acupuncture depth advancement process, when the resistance sensor detects that the needle tip contacts the bone or fascia layer, the resistance value will be significantly higher than the preset soft tissue safety threshold. At this time, the system will immediately terminate the actuator's advancement action and retract the needle to a safe position through the reverse drive module. This process, combined with the layered tissue elasticity parameters in the digital twin model, can correct the resistance threshold in real time to adapt to the tissue characteristics of different individuals, thereby building an active safety protection mechanism at the mechanical execution level.

[0069] Compared to existing technologies, existing acupuncture equipment typically relies on the operator to manually sense resistance changes, lacking quantitative monitoring and automatic response mechanisms. This can easily lead to tissue damage due to delayed human judgment. This solution, through real-time resistance monitoring and threshold control, can accurately identify abnormal puncture conditions and automatically trigger protective actions, avoiding the safety risks of traditional methods that rely on manual experience.

[0070] Through the above-mentioned technical solution, this application solves the technical problem of tissue damage caused by sudden changes in resistance during acupuncture, realizes real-time identification and automatic braking of abnormal resistance, and significantly improves the safety of automated acupuncture operations. At the same time, through the synergy of the preset dynamic threshold library and the reverse fallback mechanism, it can effectively prevent the needle from accidentally entering high-risk areas without relying on human intervention, providing reliable safety guarantees for high-precision human-machine collaborative acupuncture.

[0071] Exemplary Systems

[0072] Figure 2The figure shows a human-machine collaborative high-precision acupuncture system based on digital twin according to an embodiment of the present application, including a data acquisition module for acquiring the user's body surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; a model construction module for generating a three-dimensional point cloud model based on the body surface image data, and mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; an acupoint correction module for dynamically correcting the initial acupoint coordinates based on real-time electrophysiological data to obtain corrected individualized acupoint coordinates; an acupuncture simulation module for dynamically correcting the initial acupoint coordinates based on the local subcutaneous tissue structure data A hierarchical tissue model including nerve distribution is constructed, and the acupuncture angle and depth are simulated in combination with preset tissue elasticity parameters to generate a prediction value of nerve activation probability; an acupuncture path generation module is used to generate three-dimensional acupuncture path data including starting point coordinates, direction vector and target depth based on the corrected individualized acupoint coordinates and nerve activation probability prediction value; an instruction generation module is used to convert the three-dimensional acupuncture path data into actuator control instructions; an acupuncture path adjustment module is used to obtain the user's physiological feedback data during the acupuncture process when the actuator executes the control instructions, and dynamically adjust the three-dimensional acupuncture path data according to the physiological feedback data.

[0073] In one example, the model construction module generates initial acupoint coordinates including: marking anatomical landmark points on the surface image data, and marking anatomical reference points and acupoint reference coordinates on the standard meridian model; aligning the anatomical landmark points in the surface image data with the anatomical reference points in the standard meridian model through a feature point matching algorithm to obtain an initial transformation matrix; mapping the acupoint reference coordinates in the standard meridian model to a three-dimensional point cloud model based on the initial transformation matrix; using an iterative closest point algorithm to optimize the spatial alignment error between the three-dimensional point cloud model and the standard meridian model to generate a corrected transformation matrix; adjusting the acupoint reference coordinates according to the corrected transformation matrix to obtain the initial acupoint coordinates.

[0074] In one example, the acupoint correction module dynamically corrects the initial acupoint coordinates, including: calculating the electrophysiological response intensity around the initial acupoint coordinates, and comparing it with a preset conductivity threshold; if the electrophysiological response intensity exceeds the preset conductivity threshold, the initial acupoint coordinates are shifted toward the peak direction of the response intensity until the electrophysiological response intensity reaches a preset confidence range of the maximum value.

[0075] In one example, the acupuncture simulation module simulates the acupuncture angle and depth, including: discretizing the layered tissue model into grid units based on the finite element analysis method; calculating the relationship between the acupuncture force and tissue deformation according to the preset tissue elastic parameters, and obtaining the needle tip displacement and tissue stress distribution data under different acupuncture angle and depth combinations; mapping the tissue stress distribution data to the nerve area based on the preset nerve distribution area, and calculating the stress concentration intensity in the area; and generating a predicted value of the nerve activation probability under the corresponding acupuncture angle and depth combination based on the stress concentration intensity.

[0076] In one example, the acupuncture path generation module generates three-dimensional acupuncture path data including starting point coordinates, direction vectors and target depths, including: setting the corrected individualized acupoint coordinates as the starting point coordinates of the acupuncture path; calculating the direction vector based on the starting point coordinates pointing to the central area of ​​nerve distribution at the corresponding acupoint in the hierarchical tissue model; selecting the acupuncture depth that meets the preset activation threshold as the target depth based on the predicted value of the nerve activation probability; generating acupuncture trajectory data in three-dimensional space according to the direction vector and the target depth, and performing collision detection on the trajectory data with the three-dimensional point cloud model to eliminate trajectory paths with the risk of tissue penetration, thereby obtaining three-dimensional acupuncture path data.

[0077] In one example, the physiological feedback data acquired by the acupuncture path adjustment module includes heart rate variability data and facial expression recognition data.

[0078] In one example, the acupuncture path adjustment module dynamically adjusts the three-dimensional acupuncture path data including: regenerating acupuncture path data with a shortened target depth when the heart rate variability data is lower than a preset pain threshold or the facial expression recognition data matches a preset pain feature.

[0079] Exemplary electronic devices

[0080] Figure 3 The figure shows a block diagram of an electronic device according to an embodiment of the present application.

[0081] like Figure 3 As shown, the electronic device includes one or more processors and memory.

[0082] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0083] The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may execute the program instructions to implement the functions expected by the various embodiments of the present application described above.

[0084] In one example, the electronic device may further include an input device and an output device, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0085] Of course, to simplify, Figure 3 Only some of the components in the electronic device related to the present application are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application scenarios.

[0086] Exemplary computer-readable media

[0087] The embodiment of the present application may also be a computer-readable storage medium having stored thereon a computer program indicating

[0088] When the computer program instructions are executed by the processor, the processor executes the steps according to various embodiments of the present application described in the above “Exemplary Method” section of this specification.

[0089] Computer readable storage media can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0090] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this application are merely illustrative and not restrictive, and it should not be assumed that these advantages, strengths, and effects are required of each embodiment of this application. In addition, the specific details disclosed above are merely illustrative and facilitating understanding, and are not restrictive. The above details do not limit this application to necessarily being implemented using the above specific details.

[0091] The block diagrams of the devices, devices, equipment, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0092] It should also be noted that in the apparatus, device, and method of the present application, each component or each step can be decomposed and / or recombined, and such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0093] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0094] The above description has been provided for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A human-machine collaborative high-precision acupuncture system based on digital twins, characterized by: include: Data acquisition module, used to obtain the user's body surface image data, local subcutaneous tissue structure data, standard meridian model data and real-time electrophysiological data; a model building module, configured to generate a three-dimensional point cloud model based on the body surface image data, and map the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates; an acupoint correction module, configured to dynamically correct the initial acupoint coordinates based on the real-time electrophysiological data to obtain corrected individualized acupoint coordinates; an acupuncture simulation module for constructing a hierarchical tissue model including nerve distribution based on the local subcutaneous tissue structure data, and simulating acupuncture angle and depth in combination with preset tissue elasticity parameters to generate a predicted value of nerve activation probability; an acupuncture path generation module, configured to generate three-dimensional acupuncture path data including starting point coordinates, direction vectors, and target depths based on the corrected individualized acupoint coordinates and the predicted value of nerve activation probability; An instruction generation module, configured to convert the three-dimensional acupuncture path data into actuator control instructions; an acupuncture path adjustment module, configured to obtain physiological feedback data from the user during acupuncture performed by the actuator executing the control instruction, and dynamically adjust the three-dimensional acupuncture path data according to the physiological feedback data; The dynamic correction of the initial acupoint coordinates includes: Calculating the electrophysiological response intensity around the initial acupoint coordinates and comparing it with a preset conductivity threshold; If the electrophysiological response intensity exceeds the preset conductivity threshold, the initial acupoint coordinates are shifted toward the peak direction of the response intensity until the electrophysiological response intensity reaches a preset confidence range of the maximum value; The simulation of acupuncture angle and depth includes: Based on the finite element analysis method, the hierarchical tissue model is discretized into grid units; Calculating the relationship between acupuncture force and tissue deformation based on the preset tissue elasticity parameters to obtain needle tip displacement and tissue stress distribution data under different combinations of acupuncture angles and depths; Based on a preset nerve distribution area, mapping the tissue stress distribution data to the nerve area, and calculating the stress concentration intensity in the area; generating a predicted value of the probability of neural activation corresponding to the combination of acupuncture angle and depth according to the stress concentration intensity; The generating of three-dimensional acupuncture path data including starting point coordinates, direction vectors and target depths includes: Setting the corrected individualized acupoint coordinates as the starting point coordinates of the acupuncture path; Calculating a direction vector based on the starting point coordinate pointing to the nerve distribution center area at the corresponding acupuncture point in the layered tissue model; Based on the predicted value of the nerve activation probability, selecting an acupuncture depth that meets a preset activation threshold as a target depth; Acupuncture trajectory data in three-dimensional space is generated according to the direction vector and the target depth, and collision detection is performed between the trajectory data and the three-dimensional point cloud model to eliminate trajectory paths with tissue penetration risks and obtain three-dimensional acupuncture path data.

2. A digital twin-based human-machine collaborative high-precision acupuncture system according to claim 1, characterized in that: Mapping the standard meridian model data to the three-dimensional point cloud model to generate initial acupoint coordinates includes: Marking the body surface image data with anatomical landmarks, and marking the standard meridian model with anatomical reference points and acupoint reference coordinates; Registering the anatomical landmarks in the body surface image data with the anatomical reference points in the standard meridian model using a feature point matching algorithm to obtain an initial transformation matrix; Mapping the acupoint reference coordinates in the standard meridian model to the three-dimensional point cloud model based on the initial transformation matrix; An iterative closest point algorithm is used to optimize the spatial alignment error between the three-dimensional point cloud model and the standard meridian model to generate a corrected transformation matrix; The acupoint reference coordinates are adjusted according to the corrected transformation matrix to obtain initial acupoint coordinates.

3. The human-machine collaborative high-precision acupuncture system based on digital twin according to claim 1, characterized in that: The physiological feedback data includes heart rate variability data and facial expression recognition data, and the dynamic adjustment includes: When the heart rate variability data is lower than a preset pain threshold or the facial expression recognition data matches a preset pain feature, acupuncture path data with a shortened target depth is regenerated.

4. The human-machine collaborative high-precision acupuncture system based on digital twin according to claim 1, characterized in that: Also includes: Real-time acquisition of the needle resistance during the execution of the control instruction by the actuator; When the needle resistance exceeds a preset resistance threshold, execution of the control instruction is stopped.

5. An electronic device comprising a memory and a processor, characterized in that : The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the system as described in any one of claims 1 to 4 are implemented.

6. A computer storage medium having computer-executable instructions stored thereon, characterized in that: When the computer executable instructions are executed by a processor, the steps of the system according to any one of claims 1 to 4 are implemented.

Images