Topological mapping monitoring and correcting method for abnormal scapular humeral rhythm after breast surgery

Abstract

The invention discloses a topological mapping monitoring and correcting method for abnormal scapular brachial rhythm after breast surgery, and belongs to the field of medical rehabilitation and computer vision. The method comprises the following steps: acquiring a trunk dynamic three-dimensional point cloud and infrared thermal imaging data; constructing a multi-layer collaborative dynamic topology network fusing a skeleton structure, soft tissue deformation and physiological activeness; extracting an intrinsic motion field and comparing the intrinsic motion field with the ideal model to generate deviation measurement; and calculating a dynamic instability index, comparing the dynamic instability index with a threshold value, and triggering optimization energy correction path calculation and instruction output. According to the method, the multi-dimensional topological mapping and dynamics prediction technology is adopted, accurate recognition and feed-forward early warning of motion abnormity under the influence of deep tissue adhesion after the breast surgery can be achieved, rehabilitation of a patient is guided through an energy consumption minimization path, and the accuracy of shoulder joint function remodeling and the individualized treatment effect are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical rehabilitation and computer vision, and in particular to a topological mapping monitoring and correction method for abnormal scapulohumeral rhythm after breast surgery. BACKGROUND

[0002] Medical rehabilitation and computer vision technology is widely used in clinical medicine, especially in the process of motion function evaluation and rehabilitation guidance of patients after breast cancer surgery. Precise monitoring and real-time correction are of great significance for preventing tissue adhesion and dysfunction.

[0003] In related technologies, a breast puncture surgery robot navigation system and device based on multi-modal image fusion are disclosed in Chinese patent CN120168115A. The system includes a multi-modal image data registration and fusion module, which realizes high-precision three-dimensional reconstruction of breast tissue. It provides surgical navigation information by real-time tracking of the position of the puncture needle and spatial registration with image data. At the same time, this technology evaluates the effectiveness of the navigation path by real-time acquisition of puncture force feedback information of the puncture needle during needle insertion. When risks are found in the puncture path, dynamic adjustments are made to re-plan new paths or the system automatically stops needle insertion and gives an audible and visual warning.

[0004] For the related technologies in the above, although the robot navigation system realizes high-precision three-dimensional reconstruction and static path planning at the image level, its core focuses on local path control of the puncture needle, and cannot effectively model the large-scale and dynamic scapulohumeral rhythm movement after breast surgery. It also lacks a comprehensive instability risk warning mechanism under the coupling state of bone movement, soft tissue deformation and physiological activity, making it difficult to be directly applied to complex postoperative shoulder joint function remodeling rehabilitation scenarios. SUMMARY

[0005] To solve the above problems, the present application provides a topological mapping monitoring and correction method for abnormal scapulohumeral rhythm after breast surgery, which uses the technical means of fusing dynamic three-dimensional point cloud and infrared thermal imaging data to construct a multi-layer collaborative dynamic topological net, which can realize multidimensional deep monitoring of abnormal movement patterns, prospective instability risk warning and individualized optimal energy path correction guidance.

[0006] To achieve the above purpose, the present application adopts the following technical solutions:

[0007] In a first aspect, a method for monitoring and correcting scapulohumeral rhythm abnormalities after breast surgery is provided, comprising: acquiring dynamic three-dimensional point cloud data and synchronous infrared thermal imaging data of a patient's torso; based on the dynamic three-dimensional point cloud data and infrared thermal imaging data, constructing a multi-layer collaborative dynamic topology network that integrates bone structure, soft tissue deformation, and physiological activity information represented by infrared thermal imaging; based on the multi-layer collaborative dynamic topology network, extracting an intrinsic motion field representing motion coordination patterns during continuous motion; acquiring an ideal intrinsic motion field matched to the individual patient, comparing the real-time extracted intrinsic motion field with the ideal intrinsic motion field in terms of pattern similarity, and generating a motion coordination deviation metric; calculating a dynamic instability degree index representing the risk of motion system instability based on the motion coordination deviation metric and its time sequence variation characteristics; comparing the dynamic instability degree index with a preset risk threshold, and determining whether to trigger a correction program based on the comparison result; when the risk threshold is exceeded, calculating an optimized energy correction path based on the current state of the multi-layer collaborative dynamic topology network and the ideal intrinsic motion field, and outputting guidance instructions corresponding to the optimized energy correction path.

[0008] Based on the above technical solution, in the method for monitoring and correcting scapulohumeral rhythm abnormalities after breast surgery provided in the present application, the use of the technique of constructing a multi-layer collaborative dynamic topology network by integrating dynamic three-dimensional point cloud and infrared thermal imaging data enables multi-dimensional and deep monitoring of abnormal motion patterns, early warning of instability risks, and individualized optimal energy path correction guidance.

[0009] In combination with the above first aspect, in a possible implementation, the construction of the multi-layer collaborative dynamic topology network that integrates bone structure, soft tissue deformation, and physiological activity information includes: identifying and extracting the motion trajectories of rigid bone key points from the dynamic three-dimensional point cloud data to generate rigid skeleton layer data; based on the geometric features of the dynamic three-dimensional point cloud data, constructing an elastic grid simulating the mechanical conduction path of subcutaneous soft tissue to generate soft tissue mechanical conduction layer data; calculating the deformation vector of the elastic grid during motion to generate dynamic deformation layer data; spatially registering and mapping the infrared thermal imaging data with the elastic grid to generate functional image layer data reflecting local physiological activity; establishing an associated mapping relationship between the dynamic deformation layer data and the functional image layer data to represent the matching characteristics of tissue metabolic level and motion amplitude in the deformation area, and integrating the rigid skeleton layer data, soft tissue mechanical conduction layer data, dynamic deformation layer data, and functional image layer data to generate a multi-layer collaborative dynamic topology network.

[0010] In a possible implementation manner of the first aspect, the extracting the intrinsic motion field representing the motion coordination mode in the continuous motion process comprises: collecting time sequence state series of the rigid skeleton layer data, the soft tissue mechanics conduction layer data, the dynamic deformation layer data and the functional mapping layer data in the multi-layer coordinated dynamic topology network in a complete motion cycle; performing high-dimensional feature coding on the time sequence state series, and performing spatial dimension compression through manifold learning or principal component analysis to generate a multi-dimensional feature vector set; and defining the multi-dimensional feature vector set as the intrinsic motion field representing the internal coordination logic of the motion.

[0011] In a possible implementation manner of the first aspect, the acquiring the ideal intrinsic motion field matched with the patient individual comprises: acquiring reference data of a healthy shoulder joint of the patient in a standard action, or acquiring statistical model data of a preset healthy sample matched with an anatomical feature of the patient; and constructing a multi-layer coordinated dynamic topology network in an ideal state and extracting an ideal intrinsic motion field based on the reference data or the statistical model data, wherein when the healthy reference data is used, mirror processing for reversing spatial directionality characteristics is performed on the extracted intrinsic motion field to generate the ideal intrinsic motion field.

[0012] In a possible implementation manner of the first aspect, the calculating the dynamic instability degree index representing the instability risk of the motion system comprises: reconstructing a phase space of the motion system based on the time sequence of the motion coordination deviation degree, and calculating a dynamic chaos parameter representing a motion trajectory prediction difficulty; analyzing an inconsistency degree of the dynamic deformation layer data and the functional mapping layer data in the multi-layer coordinated dynamic topology network in a correlation mapping relationship; and fusing the dynamic chaos parameter and the inconsistency degree to calculate the dynamic instability degree index.

[0013] In a possible implementation manner of the first aspect, the method further comprises: reversely adjusting a soft tissue mechanics parameter used when the multi-layer coordinated dynamic topology network is constructed according to the dynamic instability degree index and the inconsistency degree, to realize adaptive optimization of the multi-layer coordinated dynamic topology network.

[0014] In a possible implementation manner of the first aspect, the solving the optimized energy correction path comprises: identifying a current abnormal mechanics constraint region according to the soft tissue mechanics conduction layer data and the dynamic deformation layer data in the multi-layer coordinated dynamic topology network; simulating a plurality of candidate correction trajectories in a motion space to relieve abnormal stress of the mechanics constraint region and return to the ideal intrinsic motion field; and selecting a trajectory with the minimum estimated overall muscle group energy consumption from the candidate correction trajectories based on a biomechanical model to calculate joint torque and muscle power consumption, to determine the optimized energy correction path.

[0015] In conjunction with the first aspect above, in one possible implementation, the output guidance instruction corresponding to the optimized energy correction path includes: converting the optimized energy correction path into visual motion trajectory data; superimposing the visual motion trajectory data onto the patient's torso image in a real-time video stream using an augmented reality device to form a visual guidance signal; and simultaneously generating voice or tactile prompt signals corresponding to key nodes of the optimized energy correction path.

[0016] In conjunction with the first aspect above, in one possible implementation, the method further includes: continuously recording the motion coordination deviation metric, dynamic instability index, and corrective instruction execution data to generate a rehabilitation progress assessment report; and dynamically adjusting the risk threshold and ideal intrinsic motion field used to trigger correction based on the rehabilitation progress assessment report.

[0017] Secondly, a topological mapping monitoring and correction system for scapuhumeral rhythm abnormalities after breast surgery is provided, including: a multimodal data acquisition module for acquiring dynamic three-dimensional point cloud data of the patient's trunk and synchronous infrared thermal imaging data; a multi-layer topological network construction module for constructing a multi-layer collaborative dynamic topological network that integrates skeletal structure, soft tissue deformation, and physiological activity information characterized by infrared thermal imaging based on the dynamic three-dimensional point cloud data and infrared thermal imaging data; an intrinsic feature extraction module for extracting intrinsic motion fields that characterize motion coordination patterns during continuous movement based on the multi-layer collaborative dynamic topological network; and a comparison and measurement module for obtaining an ideal intrinsic feature matching the individual patient. The system employs a motion field analysis module, which compares the real-time extracted intrinsic motion field with the ideal intrinsic motion field to generate a motion coordination deviation metric. An instability analysis module calculates a dynamic instability index characterizing the instability risk of the motion system based on the motion coordination deviation metric and its temporal variation characteristics. A strategy execution and solution module compares the dynamic instability index with a preset risk threshold and determines whether to trigger a correction procedure based on the comparison result. When the risk threshold is exceeded, an optimized energy correction path is calculated based on the current state of the multi-layer collaborative dynamic topology and the ideal intrinsic motion field, and a guidance instruction corresponding to the optimized energy correction path is output.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] This invention constructs a multi-layered, collaborative dynamic topological network that integrates dynamic three-dimensional point cloud data with infrared thermal imaging data, encompassing skeletal structure, soft tissue deformation, and physiological activity. This enables a comprehensive and in-depth assessment of the scapulothorax motor system. This method transcends the limitations of traditional approaches that focus solely on skeletal kinematics, capturing latent biomechanical characteristics and physiological responses such as muscle compensation and soft tissue adhesions. This allows for a more precise qualitative and quantitative analysis of the root causes of motor dysfunction.

[0020] This invention proposes a prospective risk assessment mechanism based on intrinsic motion fields and dynamic instability indices. By reducing complex motion patterns to feature expressions of their inherent coordination logic and analyzing their temporal dynamic chaotic characteristics, this method can shift from passively identifying motion errors to actively predicting the risk of motion system instability. This early warning capability allows for earlier intervention and effectively prevents secondary injuries or rehabilitation bottlenecks caused by the accumulation of uncoordinated movements.

[0021] This invention provides a personalized correction path calculation and guidance scheme based on biomechanical energy consumption optimization. When a risk of motion instability is detected, the system not only provides a simple correction direction, but also calculates and outputs a motion trajectory that effectively alleviates abnormal mechanical constraints while ensuring minimal overall energy consumption of the muscle groups. Combined with multimodal guidance commands such as augmented reality, voice, and haptic feedback, it provides patients with intuitive, easy-to-follow, and physiologically minimal rehabilitation training, improving correction effectiveness and patient compliance.

[0022] This invention establishes a dynamic, adaptive, closed-loop rehabilitation management system. By continuously recording and analyzing rehabilitation progress data, it can dynamically adjust its core parameters, including the risk threshold for triggering correction and the ideal intrinsic motor field as the target. This allows the rehabilitation program to intelligently evolve as the patient's functional state improves, transforming from a standardized target model into a personalized program that truly matches the patient's individual characteristics and recovery potential, achieving a higher level of precision rehabilitation.

[0023] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0024] To more clearly illustrate the technical solutions 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A structural architecture diagram of a topological mapping monitoring and correction method for scapular-humeral rhythm abnormalities after breast surgery provided in an embodiment of this application;

[0026] Figure 2 A flowchart illustrating the topological mapping monitoring and correction method for postoperative scapular-humeral rhythm abnormalities provided in this application embodiment;

[0027] Figure 3 This is a comparison diagram of intrinsic motion field trajectories provided in the embodiments of this application.

[0028] Figure 4 This is a local physiological and mechanical inconsistency distribution map provided in the embodiments of this application. Detailed Implementation

[0029] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0030] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0031] The topological mapping monitoring and correction method for scapular-humeral rhythm abnormalities after breast surgery provided in this application can be applied to, for example... Figure 1 The topological mapping monitoring and correction system 100 shown here is for postoperative scapular-humeral rhythm abnormalities, such as... Figure 1 As shown, the system includes:

[0032] The multimodal data acquisition module is used to acquire dynamic three-dimensional point cloud data of the patient's torso and synchronous infrared thermal imaging data.

[0033] A multi-layer topology network construction module is used to construct a multi-layer collaborative dynamic topology network that integrates skeletal structure, soft tissue deformation, and physiological activity information characterized by infrared thermal imaging based on the dynamic three-dimensional point cloud data and infrared thermal imaging data.

[0034] The intrinsic feature extraction module is used to extract intrinsic motion fields that characterize motion coordination patterns during continuous motion based on the multi-layer cooperative dynamic topology network.

[0035] The comparison and measurement module is used to obtain the ideal intrinsic motion field that matches the individual patient, and compare the pattern similarity between the real-time extracted intrinsic motion field and the ideal intrinsic motion field to generate a motion coordination deviation measurement.

[0036] The instability analysis module is used to calculate a dynamic instability index that characterizes the instability risk of the motion system based on the motion coordination deviation metric and its temporal variation characteristics.

[0037] The strategy execution and calculation module is used to compare the dynamic instability index with a preset risk threshold, and determine whether to trigger the correction procedure based on the comparison result. When the risk threshold is exceeded, the module calculates the optimized energy correction path based on the current state of the multi-layer cooperative dynamic topology network and the ideal intrinsic motion field, and outputs the guidance instruction corresponding to the optimized energy correction path.

[0038] like Figure 2 As shown, this application provides a method for topological mapping monitoring and correction of scapuhumeral rhythm abnormalities after breast surgery, including:

[0039] Acquire dynamic three-dimensional point cloud data of the patient's torso and synchronized infrared thermal imaging data;

[0040] Based on the dynamic three-dimensional point cloud data and infrared thermal imaging data, a multi-layer collaborative dynamic topology network is constructed that integrates information on skeletal structure, soft tissue deformation, and physiological activity characterized by infrared thermal imaging.

[0041] Based on the multi-layered collaborative dynamic topology network, the intrinsic motion field that characterizes the motion coordination mode during continuous motion is extracted.

[0042] Obtain an ideal intrinsic motion field that matches the individual patient, compare the real-time extracted intrinsic motion field with the ideal intrinsic motion field to generate a motion coordination deviation metric;

[0043] Based on the aforementioned motion coordination deviation metric and its temporal variation characteristics, a dynamic instability index characterizing the instability risk of the motion system is calculated.

[0044] The dynamic instability index is compared with a preset risk threshold, and a correction procedure is triggered based on the comparison result. When the risk threshold is exceeded, an optimized energy correction path is calculated based on the current state of the multi-layer collaborative dynamic topology and the ideal intrinsic motion field, and a guidance instruction corresponding to the optimized energy correction path is output.

[0045] It should be noted that by simultaneously acquiring dynamic three-dimensional geometric information and infrared physiological thermal information, a multi-layered collaborative dynamic topological network integrating skeletal kinematics, soft tissue biomechanics, and physiological metabolic activities is constructed, thereby establishing a digital twin model that can comprehensively reflect the intrinsic state of the scapulothorax motor system. Based on this model, data dimensionality reduction and feature extraction techniques are used to condense high-dimensional, complex multimodal time-series data into a low-dimensional intrinsic motion field that can characterize the core pattern of individual motor coordination. By quantitatively comparing the real-time extracted intrinsic motion field with the individualized ideal pattern, not only is an immediate measurement of motor coordination deviation generated, but also, by analyzing the dynamic evolution trend of this deviation measurement, a dynamic instability index that can predict the stability risk of the motor control system is calculated. Finally, this method forms a closed-loop monitoring and intervention system. Once the instability risk exceeds a preset threshold, a correction procedure is immediately initiated, with the dual objectives of returning to the ideal movement pattern and minimizing bioenergy consumption. An optimized energy correction path is calculated and output, guiding the patient in real time through multimodal commands.

[0046] In one possible implementation of the embodiments of this application, combined with Figure 2 The construction of a multi-layered collaborative dynamic topology network integrating information on skeletal structure, soft tissue deformation, and physiological activity includes:

[0047] From the dynamic 3D point cloud data, the motion trajectory of key points of the rigid skeleton is identified and extracted to generate rigid skeleton layer data;

[0048] Based on the geometric features of the dynamic three-dimensional point cloud data, an elastic mesh simulating the mechanical conduction path of subcutaneous soft tissue is constructed to generate soft tissue mechanical conduction layer data.

[0049] Calculate the deformation vector of the elastic mesh during the motion process to generate dynamic deformation layer data;

[0050] The infrared thermal imaging data is spatially registered and mapped with the elastic grid to generate functional image layer data that reflects local physiological activity.

[0051] A correlation mapping relationship is established between the dynamic deformation layer data and the functional image layer data to characterize the matching characteristics of tissue metabolic level and movement amplitude in the deformation region. The rigid skeleton layer data, soft tissue mechanical conduction layer data, dynamic deformation layer data and functional image layer data are fused to generate a multi-layer collaborative dynamic topology network.

[0052] In some implementations, the acquired dynamic 3D point cloud data is processed at a rate of 30 to 60 frames per second, maintaining a point cloud density of 500 to 1000 data points per square centimeter. A pre-trained convolutional neural network model identifies and locates anatomical landmarks on the torso surface, such as the acromion, sternoclavicular joint, and inferior scapular angle, thereby extracting the motion trajectories of key points in rigid bones and forming a rigid skeletal layer data composed of a temporal 3D coordinate sequence. Based on the static 3D point cloud data of the initial motion frame, a Poisson surface reconstruction algorithm is used to generate an initial elastic mesh. This mesh covers the entire shoulder and back region and contains approximately 3000 to 5000 nodes, used to construct a soft tissue biomechanical conduction layer data simulating the biomechanical conduction path of subcutaneous soft tissue. The topological connectivity of this mesh remains unchanged throughout the motion cycle, representing the mechanical connections within the soft tissue. A non-rigid registration algorithm is used to align the 3D point cloud data of each subsequent frame to this elastic mesh, and the spatial displacement of each mesh node relative to its initial position is calculated. These displacement vector sets constitute the dynamic deformation layer data, quantifying the stretching and compression of soft tissue during motion. Infrared thermal imaging data acquired at the same frame rate with a thermal sensitivity better than 0.05 degrees Celsius is processed. Using pre-calibrated intrinsic and extrinsic parameters, the 2D infrared thermal image is spatially registered and mapped to a 3D elastic grid. The temperature value of each pixel is assigned to the corresponding node on the elastic grid, generating functional image layer data. This layer data intuitively reflects the surface temperature distribution caused by changes in muscle activity and blood circulation, i.e., local physiological activity. Multi-layer data fusion and correlation are then performed. The core is to establish a correlation mapping relationship between dynamic deformation layer data and functional image layer data to characterize the matching characteristics of tissue metabolic level and motion amplitude in the deformation area. This correlation is achieved by calculating the synergy index of deformation intensity and temperature change within a specific anatomical region. For example, the correlation degree is calculated within a region of interest composed of multiple grid nodes. :

[0053] ;

[0054] in, It represents the deformation vector magnitude of each node in the region per unit time, reflecting the mechanical strain intensity. The temperature change at the corresponding node reflects the level of physiological metabolic activity. This represents the total number of nodes in the region. This is a normalized function. This correlation degree The dynamic attributes of this region are added. By integrating data from the rigid skeletal layer, soft tissue biomechanical conduction layer, dynamic deformation layer, and functional imaging layer, a multi-layered collaborative dynamic topology network under a unified spatiotemporal reference is finally generated. Through the above steps, the final multi-layered collaborative dynamic topology network is a structured, multi-attribute dynamic graph model. The nodes of this model not only carry geometric position, deformation vector, and temperature information, but their connecting edges also contain preset biomechanical parameters. As a high-dimensional digital twin, this topology network is no longer limited to single skeletal kinematic analysis, but incorporates the dynamic response of soft tissue and metabolic activities at the physiological level, providing a comprehensive, multimodal data foundation for the subsequent accurate extraction of the intrinsic kinetic field representing deep motor coordination patterns, thereby enabling a deeper understanding of the intrinsic mechanisms of scapuhumeral rhythm abnormalities after breast cancer surgery.

[0055] For example, suppose a patient who has undergone modified radical mastectomy for breast cancer 5 months ago is undergoing rehabilitation training to abduct the affected shoulder joint from 60 degrees to 120 degrees, using a depth camera to... Real-time acquisition of 3D point clouds of the back was performed, and displacement sequences were identified at the acromion, inferior angle of the scapula, and sternoclavicular joint to generate rigid skeletal layer data. Simultaneously, an elastic mesh containing 4500 nodes was generated based on Poisson reconstruction of the initial frame. At a specific phase with the arm raised to 90 degrees, the functional area of ​​the serratus anterior muscle was selected. Deep analysis is performed on specific grid nodes to calculate the average deformation vector magnitude of each node in the region in real time. for The average temperature rise of nodes in the corresponding synchronous infrared thermal imaging data was measured after spatial registration. for Define the normalization function The synergy index of the region is calculated by substituting it into the correlation formula. The resulting multi-layered collaborative dynamic topology network accurately characterizes the biomechanical-physiological mismatch caused by local fascial adhesions; subsequently, the extracted real-time intrinsic motion field shows that its scapuhumeral rhythm ratio is... Compared to the standard scale of an ideal mirror model The comparison was performed, and the generated value was... The coordination deviation metric was further calculated using phase space reconstruction to obtain the maximum Lyapunov exponent of dynamics. Because it exceeds the preset risk threshold The correction procedure was triggered; abnormally high stress was detected in the upper trapezius muscle. From 100 simulated candidate trajectories, a biomechanical model was used to select the trajectory that would reduce the total power consumption of the muscle group from the original posture. Down to The optimized energy correction path is used, and a dynamic "green flow trajectory" is overlaid on the patient's back in real time via an AR terminal, with the trigger intensity at the maximum lifting point being [value missing]. The tactile vibration alerts enable precise closed-loop guidance.

[0056] In one possible implementation, combining Figure 2 The extraction of the intrinsic motion field characterizing the motion coordination pattern during continuous motion includes:

[0057] Within a complete motion cycle, the temporal state sequence of data from the rigid skeleton layer, soft tissue mechanical transmission layer, dynamic deformation layer, and functional image layer in the multi-layer collaborative dynamic topology network is collected;

[0058] The time-series state sequence is encoded with high-dimensional features, and spatial dimension compression is performed through manifold learning or principal component analysis to generate a multi-dimensional feature vector set.

[0059] The set of multidimensional feature vectors is defined as the intrinsic motion field that characterizes the inherent coordination logic of the motion.

[0060] In some implementations, the input to this process is a sequence of temporal states of a multi-layered cooperative dynamic topology network over a complete motion cycle, such as a single shoulder elevation and descent movement, typically collecting data from 100 to 200 time frames. Within each time frame... The three-dimensional coordinates of key points in the rigid skeleton layer data, the node topology in the soft tissue mechanical conduction layer data, the node deformation vectors in the dynamic deformation layer data, and the node temperature values ​​in the functional image layer data are concatenated into a high-dimensional state vector. The dimension of this vector is... Typically, the number of vectors is between 20,000 and 30,000, forming the original feature space. All state vectors throughout the entire motion cycle are combined into a single... dimensional temporal state matrix ,in This represents the total number of frames. To achieve spatial dimensionality compression, Principal Component Analysis (PCA) is used to analyze the temporal state matrix. Processing is required. First, the matrix... Centering is performed column-wise, i.e., subtracting the mean of each feature. Then, the covariance matrix is ​​calculated and eigenvalue decomposition is performed, resulting in a series of eigenvectors sorted by their eigenvalues. (Selecting the first...) The eigenvectors with the largest contribution rates collectively constitute the optimal linear projection from the high-dimensional original space to the low-dimensional eigenspace. The criterion for selecting the value is to ensure that its cumulative variance contribution rate reaches a preset threshold, typically set between 95% and 99%, to ensure that while significantly reducing dimensionality, Typically ranging from 10 to 50, this preserves the dynamic characteristics of the original motion information to the greatest extent possible. This is achieved by using the original temporal state matrix... Projected onto this On a basis composed of principal eigenvectors, a... A multidimensional set of feature vectors. The generation process is represented by the following formula:

[0061] ;

[0062] in, It is a set of generated multidimensional feature vectors, where each row represents the low-dimensional state of the motion system at a given time. yes The original time-series state matrix of dimension , where each element represents the quantized value of a feature at a certain time. yes A projection matrix of dimension, whose column vectors are from The covariance matrix calculated from One principal feature vector. This multidimensional feature vector set. This is defined as the intrinsic motion field, which characterizes the inherent coordination logic of this motion. This intrinsic motion field is based on... The trajectory exists in 3D space, and its shape, smoothness, and topology imply the cooperative patterns and coupling relationships of the scapula, humerus, and related soft tissues during movement. Transforming complex, multimodal physiological and mechanical information into mathematical objects that can be quantitatively compared and pattern recognized provides a solid data foundation for subsequent deviation measurement and instability analysis. For example... Figure 3 As shown, this illustrates the deviation of the patient's real-time intrinsic motion trajectory from the ideal intrinsic motion trajectory in a low-dimensional feature space. By comparing the geometrical differences between the two curves, the system can quantitatively identify abnormal patterns in the scapuhumeral rhythm.

[0063] For example, a post-masturbation patient undergoing a 120-degree shoulder abduction test had her shoulder measured over a complete movement cycle. Data from each time frame is used to form a dimension by stitching together features such as bone displacement, mesh deformation, and physiological temperature. High-dimensional vectors are used to construct... Temporal state matrix First, for the matrix Column centering preprocessing is performed, and the covariance matrix is ​​calculated and eigenvalue decomposition is performed. Under the set requirement of 98% cumulative variance contribution rate, the top performers are automatically selected. The projection matrix is ​​composed of the principal eigenvectors. , dimension Then apply the projection formula. Perform a linear mapping to obtain a dimension of low-dimensional feature vector set This set is the intrinsic motion field that characterizes the coordination pattern of this movement. The motion trajectory it outlines in 20-dimensional space not only achieves a data dimensionality reduction of about 1250 times, but also eliminates anatomical redundancy and condenses the complex subcutaneous soft tissue stress changes and skeletal kinematic parameters into a quantitative mathematical trajectory that reflects the core logic of the scapulohumeral rhythm.

[0064] In one possible implementation, combining Figure 2 The process of obtaining an ideal intrinsic kinetic field that matches the individual patient includes:

[0065] Obtain reference data of the patient's healthy shoulder joint under standard movements, or obtain statistical model data of a pre-set healthy sample that matches the patient's anatomical features;

[0066] Based on the reference data or statistical model data, a multi-layer cooperative dynamic topology network under ideal conditions is constructed and an ideal intrinsic motion field is extracted. When based on the healthy side reference data, the extracted intrinsic motion field is subjected to mirror processing to reverse the spatial directional characteristics, thereby generating an ideal intrinsic motion field.

[0067] In some implementations, this process involves two technical paths. The first path involves acquiring dynamic 3D point cloud data of the patient's healthy shoulder joint during standard movements, such as shoulder abduction from 0 to 120 degrees, along with simultaneous infrared thermal imaging data. Subsequently, based on these reference data, a multi-layered collaborative dynamic topology network of the healthy side under ideal conditions is constructed, and its intrinsic motion field is further extracted. Since this intrinsic motion field represents the movement pattern of the healthy side, a mirroring process is performed to reverse the spatial directional features, generating an ideal intrinsic motion field suitable for the affected side. This mirroring process manipulates the 3D point cloud coordinates during the data preprocessing stage, achieved through a transformation matrix:

[0068] ;

[0069] in, This represents the homogeneous coordinate vector of any three-dimensional point acquired by the healthy side. It is a 4x4 mirror transformation matrix, which is used to reflect three-dimensional coordinates relative to the median sagittal plane of the human body. This is the transformed mirror image point coordinate vector. A multi-layered collaborative dynamic topology network is generated using the transformed point cloud data, and the ideal intrinsic kinetic field, serving as the gold standard, is ultimately extracted. The second approach, when the patient's healthy side also has functional impairment or valid data cannot be obtained, utilizes a pre-defined healthy sample statistical model. This model database stores the intrinsic kinetic fields of hundreds of healthy individuals, indexed by key anatomical features such as height, weight, arm length, and shoulder width. First, these anatomical feature parameters of the current patient are obtained, and then... The nearest neighbor algorithm retrieves 5 to 10 healthy samples with the best anatomical features. Then, a weighted average of the intrinsic motion fields of these samples is calculated, with weights based on similarity to patient features, to generate a statistically ideal intrinsic motion field that highly matches the patient's anatomical features. Regardless of the path used, the final ideal intrinsic motion field is a low-dimensional feature trajectory that mathematically encodes the coordination, efficiency, and stability expected of the scapuhumeral rhythm in a healthy state. This provides a clear and quantifiable comparison target for real-time monitoring of patient movement patterns and is a core prerequisite for realizing a closed-loop monitoring and correction system.

[0070] For example, in a patient who has undergone left breast surgery, reference point cloud data of the healthy (right) shoulder joint performing a standard abduction movement is first collected. In the mirror processing stage of constructing the ideal intrinsic motion field, homogeneous coordinates from the healthy side point cloud are selected as... Anatomical feature points, unit: Assuming the mid-sagittal plane coordinates Transformation matrix via mirror Using formulas The coordinates of the mirror image point on the affected side were calculated as follows: Then, the ideal intrinsic motion field is generated by mirroring the point cloud after full transformation; for patients who cannot obtain data from the healthy side after bilateral surgery, their height is extracted. arm length Shoulder width Characteristic parameters, through The nearest neighbor algorithm retrieves six healthy samples from the database that best match the anatomical features, and calculates a weighted average based on the similarity of each sample. For example, the weight coefficient sequence is as follows: This generates a statistically ideal intrinsic motion field that highly matches its anatomical features.

[0071] In one possible implementation, combining Figure 2 The dynamic instability index used to calculate the instability risk of the motion system includes:

[0072] Based on the time series of the motion coordination deviation metric, the phase space of the motion system is reconstructed, and dynamic chaotic parameters characterizing the difficulty of motion trajectory prediction are calculated.

[0073] Analyze the degree of inconsistency between the dynamic deformation layer data and the functional image layer data in the multi-layer collaborative dynamic topology network under the association mapping relationship;

[0074] By combining the dynamic chaos parameters and the degree of inconsistency, a dynamic instability index is calculated.

[0075] In some implementations, processing is based on a time series of motion coordination deviation metrics. This time series is a sequence of deviation metric values ​​for each frame within a continuous monitoring period. A time-delay embedding method is used to reconstruct the phase space of the motion system, mapping the one-dimensional time series to a high-dimensional space to reveal its inherent dynamic structure. This process requires determining two key parameters: the embedding dimension. With time delay ,generally The value range is from 3 to 7. The time frame range is 5 to 10 time frames, and the specific values ​​are determined through optimization using the minimum mutual information method and the pseudo-nearest neighbor method. In the reconstructed phase space, the Rosenstein algorithm is used to calculate the dynamic chaos parameter characterizing the difficulty of motion trajectory prediction, namely the maximum Lyapunov exponent. This exponent reflects the rate at which adjacent trajectories separate exponentially over time in the phase space, and its positive value is directly related to the inherent instability and unpredictability of the motion pattern. The data correlation within the multi-layered cooperative dynamic topology network is analyzed in depth. Focusing on the dynamic deformation layer data and the functional image layer data, the degree of inconsistency between the two under the correlation mapping relationship in specific muscle regions is analyzed. Several key anatomical regions of interest are delineated around the scapular girdle, such as the upper trapezius and serratus anterior. Within each region, the Pearson correlation coefficient is calculated between the time series of the average deformation vector magnitude of the nodes within a time window, such as 0.5 to 1 second, and the time series of the average temperature change of the corresponding nodes. The degree of inconsistency is defined as the difference between the absolute value of the correlation coefficient and 1, with a value ranging from 0 to 1. A higher value indicates a decoupling or mismatch between high-intensity mechanical activity and a high-metabolic physiological response, potentially predicting muscle fatigue, compensatory dysfunction, or neural control disorders. The results of the above two levels of analysis are fused to calculate a comprehensive dynamic instability index. The fusion process is achieved through a weighted model, with the formula:

[0076] ;

[0077] in, This represents the final dynamic instability index. It is the maximum Lyapunov exponent calculated from the phase space reconstruction. It is the degree of inconsistency calculated for a specific anatomical region. and These are two independent normalization functions that map their respective input values ​​to a unified, dimensionless interval, such as 0 to 1, thereby ensuring that quantities with different physical meanings can be effectively mathematically integrated. and These are preset weighting coefficients, summing to 1, used to adjust the relative importance of dynamic instability and biomechanical inconsistency in the final risk assessment. These weights can be calibrated based on clinical experience data, typically initially set between 0.4 and 0.6. The calculated dynamic instability index is a multidimensional risk measure that transcends simple geometric deviation measurements. It integrates the macroscopic dynamic stability of movement patterns with the microscopic biomechanical coupling state, enabling more sensitive and earlier detection of the critical state of the movement system transitioning from order to disorder, providing a more reliable and forward-looking decision-making basis for triggering corrective procedures. Figure 4 As shown, this illustrates the degree of local biomechanical inconsistency in the scapular girdle region. The spatial distribution of the shadows and contour values ​​in the figure reflects the coupling quality between soft tissue deformation and physiological thermo-metabolic response in different anatomical regions. The areas with deeper shadows and higher values ​​directly indicate the risk areas where mechanical load and physiological compensation in the musculoskeletal system are significantly disconnected, providing a spatial distribution basis for the accurate calculation of dynamic instability and the reverse adjustment of soft tissue parameters.

[0078] For example, in a post-masturbation patient performing shoulder abduction, a continuous time series of their motor coordination deviation measurement is extracted. The phase space of the motor system is reconstructed using a time-delay embedding method, and the embedding dimension is determined through optimization. Time delay The frame was used to calculate the dynamic maximum Lyapunov exponent using the Rosenstein algorithm. After normalization, the following was obtained Meanwhile, for the upper trapezius muscle region, time series of the changes in the average deformation vector magnitude and average temperature of the nodes were extracted within a 0.8-second time window, and the degree of inconsistency between the two was obtained by calculating the Pearson correlation coefficient. The corresponding normalized value is Finally, the dynamic weights are set. With physiological and mechanical weights Substitute into the formula By performing fusion calculations, dynamic instability indices are obtained. This result comprehensively assesses the macroscopic stability of the movement pattern and the coupling state of its microscopic physiological and mechanical properties.

[0079] In one possible implementation, combining Figure 2 The method further includes: adjusting the soft tissue mechanical parameters used in constructing the multi-layer collaborative dynamic topology network in reverse according to the dynamic instability index and the degree of inconsistency, so as to realize the adaptive optimization of the multi-layer collaborative dynamic topology network.

[0080] In some implementations, this adaptive optimization process is triggered by a dynamic instability index or a local inconsistency exceeding a preset adjustment threshold. Once triggered, a reverse parameter adjustment procedure is initiated. The core of this procedure is locating the anatomical regions with the highest inconsistencies, where the model deviates most from the actual physiological response. When constructing soft tissue biomechanical conduction layer data, each elastic grid's connecting edges are assigned initial soft tissue biomechanical parameters, such as Young's modulus representing tissue stiffness and damping coefficient representing energy dissipation. These initial values ​​are typically based on a standard human model database. An iterative optimization algorithm fine-tunes the grid biomechanical parameters of the located regions. This adjustment follows a negative feedback principle, aiming to reduce the observed inconsistencies. For example, if a region exhibits high deformation and low temperature response, it is inferred that the actual tissue in that region may be stiffer or in a fatigue-protected state than the model presets. In this case, the Young's modulus parameter of the corresponding grid connecting edge for that region will be moderately increased. The adjustment amount is determined by an update rule, as follows:

[0081] ;

[0082] in, These are the updated mechanical parameter values, such as Young's modulus, that will be used in the next calculation cycle. This is the current parameter value. It is a learning rate coefficient, usually ranging from 0.01 to 0.1, used to control the adjustment step size and prevent system oscillation. It represents the degree of inconsistency currently measured in this region. It is an ideal reference inconsistency level, with a value close to zero, representing a perfect coupling between mechanical deformation and physiological response. It is a dynamic instability index The relevant gain function increases with overall system instability, making parameter adjustments more sensitive and drastic. This adjustment process is performed once after each rehabilitation training cycle or initiated in real time when persistent abnormalities in instability indicators are detected. Through repeated iterations, the soft tissue biomechanical parameters in the multi-layered collaborative dynamic topology network gradually converge, maximizing the consistency between the model-predicted biomechanical behavior and the actual physiological feedback observed through infrared thermography. Ultimately, this adaptive optimization process ensures that the multi-layered collaborative dynamic topology network is not only a static anatomical model but also a dynamic digital twin capable of learning and adapting to individual patient physiological changes. This continuous self-calibration significantly improves the accuracy and effectiveness of subsequent intrinsic motion field extraction, deviation measurement, and correction path calculation, giving the entire monitoring and correction system personalized and high-precision characteristics.

[0083] For example, in a patient who experienced a mismatch between biomechanical deformation and thermometabolic response in the serratus anterior muscle region during postoperative rehabilitation training, the degree of inconsistency in this specific muscle region was monitored in real time. And the overall dynamic instability index reached This triggered the adaptive optimization process, at which point the current Young's modulus parameters of the elastic mesh connection edges in that region were... By setting the learning rate coefficient Degree of inconsistency between ideal reference and the gain function value determined based on the instability index. Substitute the parameters into the update rule formula Calculations are performed to determine the Young's modulus value to be used in the next calculation cycle. Through this iterative fine-tuning based on the principle of negative feedback, dynamic calibration of the soft tissue mechanical parameters inside the multi-layer collaborative dynamic topology network is achieved, effectively reducing the deviation between the mechanical behavior predicted by the digital twin model and the real physiological feedback, and ensuring that the monitoring system can accurately capture and adapt to the evolution of biomechanical characteristics of individual patients due to tissue repair or fatigue.

[0084] In one possible implementation, combining Figure 2 The optimized energy correction path includes:

[0085] Based on the soft tissue mechanical transmission layer data and dynamic deformation layer data in the multi-layer collaborative dynamic topology network, identify the current abnormal mechanical constraint region;

[0086] With the goal of alleviating the abnormal stress in the mechanically constrained region and restoring the ideal intrinsic motion field, multiple candidate correction trajectories are simulated in the motion space;

[0087] From the candidate correction trajectories, the joint torque and muscle power consumption are calculated based on the biomechanical model, and the trajectory that minimizes the estimated overall energy consumption of the muscle group is selected as the optimized energy correction path.

[0088] In some implementations, this process is initiated after the correction procedure is triggered. Anomaly diagnosis is performed based on the multi-layered collaborative dynamic topology network at the current moment. By analyzing soft tissue mechanical transmission layer data and dynamic deformation layer data, the stress and strain distribution on the elastic mesh is calculated using finite element analysis. When the strain value of a certain region exceeds 1.5 to 2 times its mean strain under the corresponding state in the ideal intrinsic motion field, that region is identified as the current abnormal mechanical constraint region, which is the main physical source causing the deviation of the motion pattern. The dual optimization objectives are to alleviate the abnormal stress in the identified mechanical constraint regions and guide the motion trajectory back to the ideal intrinsic motion field. In the joint angle space of motion control, starting from the current abnormal state and ending at the nearest target point in the ideal intrinsic motion field, a set of 50 to 200 candidate correction trajectories is generated using path planning algorithms, such as fast exploratory random trees or spline interpolation. Each candidate correction trajectory represents a potential way to transition from the current incorrect posture to the target correct posture. A biomechanical energy consumption assessment is performed on each candidate correction trajectory to screen for the optimal path. This assessment is based on a simplified musculoskeletal biomechanical model. First, inverse dynamics calculations are performed on each candidate trajectory to determine the sequence of joint torques required to achieve that trajectory. Then, these joint torques are allocated to the corresponding muscle groups, and a muscle energy metabolism model is used to estimate the total energy consumption required to complete the trajectory. The core of the optimization lies in minimizing a comprehensive energy consumption index. The selection process for this indicator is expressed as follows:

[0089] ;

[0090] in, This is the final optimized energy correction path. This indicates the selection of the candidate trajectory that minimizes the integral value. The joint angle time-series vector representing the candidate correction trajectory. It is based on the trajectory The joint torque vector calculated using the inverse dynamics model. It is the rate of change of joint torque, reflecting the smoothness of muscle strength control. This is an estimate of the level of muscle activation required to generate that torque. It is a comprehensive cost function that integrates factors such as joint torque, rate of change of torque, and muscle activation level, quantifying them into instantaneous physiological energy consumption. The integral is the time of the entire corrective movement. The process is performed internally. This is achieved by traversing all candidate correction trajectories and calculating their... Value, the final choice makes The trajectory with the minimum value is identified as the optimal energy correction path. This path not only points to the correct target geometrically, but also ensures a smooth, efficient, and low-burden correction process dynamically, thereby minimizing patient fatigue and the risk of secondary injury.

[0091] For example, a patient four months post-right breast augmentation triggered a correction procedure when abducting the shoulder joint to 75 degrees. Finite element analysis based on the current multi-layered collaborative dynamic topology revealed a local strain value of 0.32 in the axillary fascia region, which is 1.78 times the average phase value of 0.18 corresponding to the ideal intrinsic motion field, thus identifying it as the current abnormal mechanical constraint region. Subsequently, within the joint angle space, starting from the current 75-degree abnormal pose and ending at the nearest 80-degree target point in the ideal field, a spline interpolation algorithm was used to simulate and generate 150 candidate correction trajectories in the motion space. Inverse dynamics was then performed on each trajectory using a musculoskeletal biomechanical model to obtain the joint torque vector required to achieve the trajectory. and its rate of change By substituting the formula of the comprehensive energy consumption cost function Integral calculations were performed, and candidate trajectories were identified through comparison. Total power consumption for The power consumption of other trajectories is all within In conclusion, the trajectory that minimizes the integral value is ultimately selected. As an optimized energy correction pathway, this pathway guides patients to correct their compensatory "shoulder-shrugging" movements and relieve abnormal stress in the armpit while ensuring muscle activation levels. Co-optimization with torque change rate.

[0092] In one possible implementation, combining Figure 2 The output guidance instructions corresponding to the optimized energy correction path include:

[0093] The optimized energy correction path is then converted into visualized motion trajectory data.

[0094] The visualized motion trajectory data is superimposed onto the patient's torso image in a real-time video stream using augmented reality devices to form a visual guidance signal.

[0095] Simultaneously generate voice or tactile prompts corresponding to key nodes of the optimized energy correction path.

[0096] In some implementations, this process first processes the optimized energy correction path as input, which is essentially a time series containing continuous posture information. This time-series data is then transformed into visual motion trajectory data that can be rendered by a graphics engine. This is typically represented as a semi-transparent 3D target model, a light band extending in space, or a series of continuous "ghost" posture silhouettes, updated synchronously with the real-time video stream, usually at 30 to 60 frames per second to ensure visual smoothness. Subsequently, this visual motion trajectory data is overlaid onto the patient's torso image in the real-time video stream using augmented reality devices, such as smart glasses or a display with a camera. This process requires real-time spatial positioning and registration techniques, such as simultaneous localization and mapping algorithms, to ensure that the virtual guidance trajectory is precisely aligned with the patient's real body in 3D space, with a registration error controlled within 2 cm and an end-to-end latency of less than 20 milliseconds to avoid user dizziness and operational delays. The resulting visual guidance signal provides the patient with a clear, WYSIWYG motion imitation target. Key nodes in the optimized energy correction path are processed simultaneously while outputting the visual guidance. These key nodes are pre-marked kinematic feature points along the path, such as the starting point of the movement, the position of the maximum abduction angle, and the point of change in velocity direction. These nodes are associated with a pre-defined instruction library. When the patient's real-time movement approaches or reaches a key node, for example, when the difference between the actual shoulder joint angle and the target angle of the key node is less than 5 degrees, a corresponding voice or tactile cue signal is triggered. Voice cues can be short instructions such as "start lifting," "keep stable," or "slowly lower." Tactile cues, through wearable vibration units, generate vibrations of different intensities or rhythms at specific body parts, such as short vibrations at the start of movement and continuous weak vibrations when stability is required. By integrating these three synchronously output guidance instructions, an immersive, multi-sensory corrective environment is created. Visual guidance provides the global path of movement, while voice and tactile cues reinforce the rhythm and key points of the movement at critical moments, reducing the cognitive load on the patient to perform the correct movement and improving the compliance and effectiveness of corrective training.

[0097] For example, a post-masturbation patient undergoing shoulder abduction correction training converts the calculated optimized energy correction path into high-frequency visual motion trajectory data at 60 frames per second, appearing as a semi-transparent green light band extending in three-dimensional space. This data is overlaid on the patient's torso image in real time using augmented reality glasses. Simultaneous localization and mapping algorithms control the spatial registration error to within 1.5 cm and the end-to-end latency to within 18 milliseconds, providing the patient with an intuitive "what you see is what you get" imitation target. During the movement, when the patient's real-time shoulder joint angle reaches 116 degrees, and the difference between this and the preset maximum abduction angle key node of 120 degrees is 4 degrees (meeting the trigger threshold of less than 5 degrees), the command library is immediately linked, and a voice prompt of "Reached the top, please hold" is triggered synchronously. Additionally, a wearable vibration unit generates a frequency of [missing information - likely a specific frequency] at the patient's acromion position. The continuous weak tactile signal; through the synchronous integration of visual path, key point voice and local tactile feedback, the guidance instruction significantly reduces the cognitive load of patients in complex rehabilitation movements and ensures that their movement trajectory is highly aligned with the optimized biomechanical energy-minimizing path.

[0098] In one possible implementation, combining Figure 2 The method further includes:

[0099] The system continuously records the measured motion coordination deviation, dynamic instability index, and corrective instruction execution data to generate a rehabilitation progress assessment report.

[0100] Based on the rehabilitation progress assessment report, the risk threshold and ideal intrinsic sports field used to trigger correction are dynamically adjusted.

[0101] In some implementations, the process continuously runs a data recording unit in the background. This unit records the time series of motion coordination deviation and dynamic instability indices at high frequency during each training session, and archives the guidance instructions triggered for each correction and the patient's actual execution data of those instructions. Execution data is generated by comparing and optimizing the energy correction path with the patient's actual motion trajectory, quantifying it as following error. At the end of each training cycle, this accumulated data is automatically processed. Key performance indicators, such as average deviation, frequency of instability events, and average following error, are calculated, and linear regression analysis is performed on the trends of these indicators over the past 5 to 10 training cycles. This ultimately generates a visualized rehabilitation progress assessment report. This report is not only for clinicians' reference but also serves as input for the system's self-adjustment. Based on the rehabilitation trends revealed in the report, two core parameters are dynamically adjusted. The first is the risk threshold used to trigger correction. When the report shows that the patient's mean dynamic instability index has shown a stable downward trend in the most recent assessment cycle, for example, a decrease exceeding 15%, the risk threshold will be automatically and appropriately increased. This adjustment is achieved through the following rules:

[0102] ;

[0103] in, This is the new risk threshold for the next stage. This is the current threshold. It is the improvement rate calculated based on the rehabilitation progress assessment report, and is a dimensionless positive value. This is a small adjustment coefficient, typically between 0.05 and 0.1, to ensure smooth adjustment. This aims to gradually reduce the system's intervention sensitivity as the patient's abilities improve, encouraging greater active exploration. Secondly, there is the ideal intrinsic motor field. When a patient's movement pattern has stabilized, but a stable systematic deviation remains between their best performance and the initially set ideal intrinsic motor field, this may indicate that the patient has formed a new, functional compensatory pattern. At this point, the target field is fine-tuned to moderately align the ideal intrinsic motor field with the patient's best performance. This process is achieved through weighted fusion of the fields:

[0104] ;

[0105] in, It is the updated Ideal Original Sports Field. It is the old ideal place. It is the intrinsic kinetic field extracted and averaged from the patient's best athletic performances in the most recent evaluation period. It is a fusion weighting factor with a very small value, typically between 0.01 and 0.05, ensuring that the ideal goal is adjusted only slightly and gradually. This adaptive adjustment allows rehabilitation goals to evolve from a universal "perfect" standard to achievable personalized goals that better match the individual patient's anatomy and functional potential, thereby achieving truly individualized and precise rehabilitation.

[0106] For example, a post-masturbation patient who has entered the third stage of rehabilitation, in 10 training sessions over the past two weeks, The frequency of recordings of motor coordination deviation and dynamic instability indices, along with the rehabilitation progress assessment report generated through linear regression analysis, showed that the mean dynamic instability index of patients decreased from [previous value]. Steadily declining to The improvement rate was calculated. for Based on this trend, adaptive threshold adjustment is initiated, adjusting the current risk threshold for triggering correction from... Substitute into the formula Set the smoothing coefficient The new risk threshold for the next stage is calculated. This approach encourages patients to actively explore a wider range of areas by reducing intervention sensitivity; simultaneously, monitoring revealed that although the patients' movement patterns had stabilized, their optimal performance trajectory still deviated from their initial ideal intrinsic movement field. A fixed systematic bias still exists, indicating that it has formed a functional compensation. Therefore, the average intrinsic field of its three most recent best performances is extracted. Set the fusion weight factor Through formula By performing weighted fusion and fine-tuning of the target field, the rehabilitation goals are brought closer to the functional potential of individual patients, realizing a dynamic adaptive evolution from universal standards to individualized and precise rehabilitation goals.

[0107] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0108] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A topological mapping monitoring and correction method for scapulohumeral rhythm abnormality after breast surgery, characterized in that, The method comprises: acquiring dynamic three-dimensional point cloud data and synchronous infrared thermal imaging data of a patient's torso; based on the dynamic three-dimensional point cloud data and infrared thermal imaging data, constructing a multi-layer collaborative dynamic topological network that integrates bone structure, soft tissue deformation, and physiological activity information represented by infrared thermal imaging; based on the multi-layer collaborative dynamic topological network, extracting an intrinsic motion field representing the motion coordination pattern during continuous motion; acquiring an ideal intrinsic motion field matched with the patient's individual, comparing the real-time extracted intrinsic motion field with the ideal intrinsic motion field in terms of pattern similarity, and generating a motion coordination deviation metric; according to the motion coordination deviation metric and its time sequence variation characteristics, calculating a dynamic instability degree index representing the risk of motion system instability; comparing the dynamic instability degree index with a preset risk threshold, and determining whether to trigger a correction program according to the comparison result; when the risk threshold is exceeded, based on the current state of the multi-layer collaborative dynamic topological network and the ideal intrinsic motion field, solving an optimized energy correction path, and outputting guidance instructions corresponding to the optimized energy correction path.

2. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 1, wherein, The construction of the multi-layer collaborative dynamic topological network that integrates bone structure, soft tissue deformation, and physiological activity information comprises: from the dynamic three-dimensional point cloud data, identifying and extracting the motion trajectory of rigid bone key points to generate rigid skeleton layer data; based on the geometric features of the dynamic three-dimensional point cloud data, constructing an elastic grid simulating the mechanical conduction path of subcutaneous soft tissue to generate soft tissue mechanical conduction layer data; calculating the deformation vector of the elastic grid during motion to generate dynamic deformation layer data; spatially registering and mapping the infrared thermal imaging data and the elastic grid to generate functional image layer data reflecting local physiological activity; establishing an associated mapping relationship between the dynamic deformation layer data and the functional image layer data to represent the matching characteristics of tissue metabolic level and motion amplitude in the deformation area, and integrating the rigid skeleton layer data, soft tissue mechanical conduction layer data, dynamic deformation layer data, and functional image layer data to generate a multi-layer collaborative dynamic topological network.

3. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 2, wherein, The extraction of the intrinsic motion field representing the motion coordination pattern during continuous motion comprises: within a complete motion cycle, collecting the time sequence state sequence of the rigid skeleton layer data, soft tissue mechanical conduction layer data, dynamic deformation layer data, and functional image layer data in the multi-layer collaborative dynamic topological network; high-dimensional feature coding is performed on the time sequence state sequence, and spatial dimension compression is performed through manifold learning or principal component analysis to generate a multi-dimensional feature vector set; the multi-dimensional feature vector set is defined as an intrinsic motion field representing the internal coordination logic of this motion.

4. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 1, wherein, The acquisition of the ideal intrinsic motion field matched with the patient's individual comprises: acquiring reference data of the patient's healthy side shoulder joint under standard action, or acquiring preset healthy sample statistical model data matched with the patient's anatomical features; Based on the reference data or statistical model data, a multi-layer collaborative dynamic topology network in an ideal state is constructed, and an ideal intrinsic motion field is extracted, wherein when the reference data of the healthy side is used, mirror processing for reversing the spatial directivity characteristics is performed on the extracted intrinsic motion field to generate the ideal intrinsic motion field.

5. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 2, wherein, The calculation of the dynamic instability degree index representing the instability risk of the motion system includes: Based on the time series of the motion coordination deviation degree, the phase space of the motion system is reconstructed, and a dynamic chaos parameter representing the prediction difficulty of the motion trajectory is calculated; The inconsistency degree of the dynamic deformation layer data and the functional mapping layer data in the multi-layer collaborative dynamic topology network is analyzed under the correlation mapping relationship; The dynamic instability degree index is calculated by fusing the dynamic chaos parameter and the inconsistency degree.

6. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 5, wherein, The method further includes: according to the dynamic instability degree index and the inconsistency degree, reversely adjusting the soft tissue mechanics parameters used when constructing the multi-layer collaborative dynamic topology network, to realize adaptive optimization of the multi-layer collaborative dynamic topology network.

7. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 2, wherein, The solving and optimization of the energy correction path include: According to the soft tissue mechanics conduction layer data and the dynamic deformation layer data in the multi-layer collaborative dynamic topology network, the current abnormal mechanics constraint region is identified; A plurality of candidate correction trajectories are simulated in the motion space to relieve the abnormal stress of the mechanics constraint region and return to the ideal intrinsic motion field; From the candidate correction trajectories, joint torque and muscle power consumption are calculated based on the biomechanical model, the trajectory with the minimum estimated overall muscle group energy consumption is selected as the optimization energy correction path.

8. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 1, wherein, The output of the guidance instruction corresponding to the optimization energy correction path includes: The optimization energy correction path is converted into visual motion trajectory data; The visual motion trajectory data is superimposed on the patient's trunk image in the real-time video stream through an augmented reality device to form a visual guidance signal; Synchronous generation of voice or tactile prompt signals corresponding to the key nodes of the optimization energy correction path.

9. The method for topographic mapping monitoring and correction of scapulohumeral rhythm abnormalities following breast surgery of claim 1, wherein, The method further includes: Continuously recording the motion coordination deviation degree, the dynamic instability degree index, and the correction instruction execution data to generate a rehabilitation progress evaluation report; Based on the rehabilitation progress evaluation report, the risk threshold for triggering correction and the ideal intrinsic motion field are dynamically adjusted.

10. A topographic mapping monitoring and correction system for scapulohumeral rhythm abnormalities post breast surgery, characterized by, The system is used for the topological mapping monitoring and correction method for breast postoperative scapulohumeral rhythm abnormalities as claimed in any one of claims 1-9, and the system includes: A multi-modal data acquisition module for acquiring dynamic three-dimensional point cloud data and synchronous infrared thermal imaging data of a patient's trunk; A multi-layer topology network construction module for constructing a multi-layer collaborative dynamic topology network that integrates bone structure, soft tissue deformation, and physiological activity information represented by infrared thermal imaging based on the dynamic three-dimensional point cloud data and infrared thermal imaging data; An intrinsic feature extraction module for extracting an intrinsic motion field representing a motion coordination mode in a continuous motion process based on the multi-layer collaborative dynamic topology network; a comparison measure module configured to obtain an ideal eigen-motion field matched with the patient individual, perform a pattern similarity comparison between the real-time extracted eigen-motion field and the ideal eigen-motion field, and generate a motion coordination deviation measure; an instability analysis module configured to calculate a dynamic instability degree index representing a risk of instability of the motion system according to the motion coordination deviation measure and a time sequence variation characteristic thereof; a strategy execution and solution module configured to compare the dynamic instability degree index with a preset risk threshold, determine whether to trigger a correction program according to a comparison result, and when the risk threshold is exceeded, solve an optimized energy correction path based on a current multi-layer collaborative dynamic topology net state and the ideal eigen-motion field, and output a guidance instruction corresponding to the optimized energy correction path.

Images