A Method for Analyzing Muscle Movement Abnormalities in the Rehabilitation Process

Abstract

This invention proposes a method for analyzing muscle movement abnormalities during rehabilitation, relating to the field of muscle movement abnormality detection. First, it extracts asymmetric collaborative offset rates between channels based on a muscle collaborative offset tensor. A normalized collaborative offset matrix is ​​constructed by combining a time weighting factor and a channel pair weighting factor, and a temporal curvature-enhanced offset trend perception is derived. Then, an abnormal response is dynamically integrated using an energy perturbation matrix and a cumulative deviation embedding structure. An asymmetric collaborative modeling path is constructed based on an adaptive adjacency matrix and a perturbation operator matrix. A tensor interaction structure and channel offset mapping are fused to generate an offset field coupling matrix. Finally, a deviation sensitivity response vector and a deviation intensity tensor are generated using a standard template. An abnormality scoring index is calculated by introducing a Gaussian perturbation kernel function, a sign function, and channel direction weights. Model training and optimization are completed using a multilayer perceptron and a cross-entropy loss function.

Description

Technical Field

[0001] This invention relates to the field of muscle movement abnormality detection, and more particularly to a method for analyzing muscle movement abnormalities in the rehabilitation process. Background Technology

[0002] With the development of rehabilitation medicine and intelligent detection technology, the demand for muscle movement function assessment in rehabilitation populations suffering from stroke, brain injury, and muscle atrophy is increasing. Current rehabilitation assessments largely rely on physician experience and patient subjective feelings, lacking objective and quantitative parameter support. Electromyography (EMG), as an important electrophysiological indicator reflecting muscle activity, can provide a basis for analyzing movement abnormalities during rehabilitation. However, existing methods suffer from poor anti-interference capabilities, limited analytical dimensions, and difficulty in adapting to individualized rehabilitation needs. Therefore, it is necessary to propose a muscle movement abnormality analysis method based on electrical measurements to support subsequent rehabilitation assessment and training plan adjustments.

[0003] Publication No. CN119564236A discloses a method, device, equipment, and storage medium for muscle movement analysis based on electromyography (EMG) data. This method utilizes EMG data for rehabilitation detection, first employing adaptive filtering for noise reduction, then using a sliding window to extract energy and zero-crossing rate, and combining wavelet thresholding for noise reduction to obtain time-frequency features. Finally, these features are fused into a vector and input into a trained random forest model for analysis, thereby identifying muscle movement states in real time and enabling the detection of the rehabilitation process. Publication No. CN112329640A discloses a rehabilitation detection system for facial nerve palsy based on eye muscle movement analysis. This system utilizes multidimensional data collected during rehabilitation, including eyelid area, eye movement velocity, eyebrow and forehead wrinkles, and facial symmetry. Parameters are extracted through semantic segmentation and 3D face reconstruction to establish a disease severity prediction and rehabilitation assessment model. The system integrates features and quantifies the rehabilitation level based on an exponential function, achieving objective detection of rehabilitation progress.

[0004] In existing technologies, CN119564236A mainly relies on the energy characteristics and zero-crossing rate characteristics of electromyography for motion state analysis. Its feature extraction method has a certain degree of manual dependence and dimensional fragmentation, making it difficult to accurately characterize the subtle perturbation differences and dynamic coupling characteristics between multiple channels in muscle synergistic movements during rehabilitation. At the same time, it uses random forest as a classification model, which has a static structure and is difficult to dynamically adapt to the non-stationarity of abnormal distributions during training. CN112329640A focuses more on the fusion modeling of facial geometric feature parameters and explicit posture information, lacking the ability to perceive the temporal dynamics of deviations at the level of muscle motor neural control. In addition, its method of scoring the degree of rehabilitation using an exponential function has problems such as insufficient global sensitivity and insufficient dynamic normalization, making it difficult to reflect the abnormal change trend under instantaneous perturbation, resulting in limitations in the accuracy and timeliness of rehabilitation abnormality identification. Summary of the Invention

[0005] This invention proposes a method for analyzing muscle movement abnormalities during rehabilitation. First, it extracts asymmetric collaborative offset rates between channels based on a muscle collaborative offset tensor. A normalized collaborative offset matrix is ​​constructed by combining a time weighting factor and a channel pair weighting factor, and the temporal curvature is further derived to enhance the perception of offset change trends. Then, an energy perturbation matrix characterizes the perturbation behavior, and a cumulative deviation embedding structure is used to achieve dynamic integration of abnormal responses. An asymmetric collaborative modeling path between multi-dimensional channels is established using an adaptive adjacency matrix and a perturbation operator matrix. A tensor interaction structure and channel offset mapping are fused to generate an offset field coupling matrix, constructing abnormal performance characteristics in a multi-channel space. Finally, a deviation sensitivity response vector is constructed based on the offset field coupling matrix and a standard template, and a multi-dimensional deviation intensity tensor is generated. An abnormality scoring index is calculated by combining a Gaussian perturbation kernel function with bandwidth adjustment capability, a sign function, and channel direction weights. The abnormality scoring index is used to train and optimize the model through a multilayer perceptron and a cross-entropy-based probabilistic discriminant loss function, enabling multi-dimensional modeling and accurate abnormality analysis of muscle movement abnormalities during rehabilitation.

[0006] A method for analyzing muscle movement abnormalities during the rehabilitation process, the specific method is as follows:

[0007] S1. Collect muscle movement datasets during rehabilitation testing, and preprocess the collected datasets to construct a muscle movement dataset.

[0008] S2. Divide the muscle motion dataset into muscle motion vectors, calculate the synergistic relationship between muscle groups to obtain a dynamic adjacency matrix sequence, and obtain the muscle synergistic offset tensor based on the structural change rate matrix.

[0009] S3. Based on the muscle co-shift tensor, the asymmetric co-shift rate is weighted and normalized to construct a co-shift matrix, and the continuous difference is extracted to obtain the temporal curvature. The co-shift matrix and the temporal curvature are weighted and superimposed, and channel coupling weights and perturbation adjustment factors are introduced to obtain the energy perturbation matrix and accumulate calculation to construct the co-embedding feature.

[0010] S4. Based on the collaborative embedding feature, a perturbation weight vector is introduced to obtain an adaptive adjacency matrix and a perturbation operator matrix is ​​constructed with an asymmetric adjustment factor. The matrix information of historical moments is fused to generate tensor interaction superposition results. A channel offset factor is introduced for differential coupling to form an offset field coupling matrix.

[0011] S5. By exponentially mapping the channel offset intensity to the offset field coupling matrix and the Gaussian perturbation kernel function, and by fusing the channel differences, a deviation sensitivity response vector is generated. Combined with the channel direction weight coefficient and the residual direction projection component, a deviation intensity tensor is generated and jointly modeled to obtain the anomaly scoring index.

[0012] S6. Input the muscle movement dataset, combine it with the loss function, construct a muscle movement anomaly analysis model, and iteratively train it until convergence to complete the training of the muscle movement anomaly analysis model.

[0013] Preferably, the construction process of the muscle movement dataset includes data acquisition and preprocessing. In the data acquisition stage, patients at different rehabilitation stages are selected as the detection subjects, and four types of standard rehabilitation movements are designed: downward gaze, upward gaze, flexion and extension, and fist clenching. Using a multi-channel surface electromyography sensor array and an inertial measurement unit, the raw electromyographic signals, triaxial acceleration, triaxial angular velocity, and joint angle change information of the target muscle group are collected respectively. Synchronous annotation is achieved through a unified timestamp to construct an initial dataset containing multi-source time-series data.

[0014] In the preprocessing stage, the raw electromyography (EMG) signals are first baseline-corrected by adjusting the initial values ​​to zero using a mean shift method. Then, a moving average method is used to average each sampling point with the preceding and following sampling points to suppress abrupt noise and smooth the waveform. Acceleration and angular velocity signals are simultaneously smoothed using the same method, and static calibration is completed through zero-point relocation. The entire signal is divided into segments with fixed time slices, and adjacent time slices are overlapped to enhance continuity. Within each time slice, EMG energy, root mean square value, zero-crossing rate, average frequency, and median frequency are extracted. The mean, maximum, peak value, and standard deviation of triaxial acceleration and triaxial angular velocity are statistically analyzed, and the change in joint angle and its rate of change between adjacent time slices are calculated. All feature data are uniformly subjected to min-max normalization to form a muscle motion dataset, which serves as input for subsequent anomaly analysis models and rehabilitation assessment modules.

[0015] Preferably, in step S2, the data collected at each time point is divided into muscle motion vectors, and the vectors are combined according to a fixed time step to form a muscle motion signal matrix.

[0016] Within each time step, the multi-channel muscle motion data features within the muscle motion vector are paired, the co-activation intensity between the channel pairs is calculated, and a stability constant is introduced during the calculation process to avoid numerical calculation errors. Finally, the co-activation intensity of each channel pair is concatenated according to the index position to obtain the muscle group adjacency matrix of the current time step, and a dynamic adjacency matrix sequence is formed through iterative calculation within the entire sampling window.

[0017] Based on the dynamic adjacency matrix sequence, the difference in co-activation intensity at different times is calculated to obtain the structural change rate. The difference component between adjacent times is calculated through the structural change rate to obtain the co-offset change difference of each channel pair. Based on the co-offset change difference and the consistency sign function, the asymmetric co-offset rate used to distinguish the direction of co-offset change is obtained. The asymmetric co-offset rates at each time are concatenated in sequence to form a complete muscle co-offset tensor.

[0018] Furthermore, by establishing a dynamic muscle group adjacency matrix, the co-activation intensity of all channel pairs can be recorded in matrix form at each time step, forming a dynamic adjacency graph sequence within the entire sampling window. The structural change rate matrix is ​​used to calculate the difference in co-activation intensity between adjacent time steps, and the resulting structural change rate can refine the instantaneous fluctuations in the muscle group synergy. Subsequently, based on the construction process of the co-offset rate, the structural change rate at different time steps is further decomposed to extract the offset information of channel pairs during evolution. On this basis, a consistency sign function is introduced to determine the direction of the co-offset rate, thereby constructing an asymmetric co-offset rate, which is then accumulated and generated as a muscle co-offset tensor through time iteration. This step ensures the organic connection between the muscle group adjacency matrix, the structural change rate matrix, the asymmetric co-offset rate, and the muscle co-offset tensor, enabling a complete modeling of muscle motion signals in terms of inter-channel co-activation intensity, time-series difference, directional offset, and asymmetric evolution trend.

[0019] Preferably, in step S3, based on the muscle co-shift tensor, the asymmetric co-shift rate between channels is extracted, and a channel pair weighting factor is introduced in combination with the influence of different muscle groups during rehabilitation training. Combined with a time weighting factor set with time decay, the asymmetric co-shift rate is jointly weighted. The weighting result is normalized according to the global maximum and minimum values ​​of the muscle co-shift tensor to construct the co-shift matrix.

[0020] Based on the cooperative offset matrix, second-order calculations are performed on the continuous differences of adjacent time steps, and then normalized to construct a time-series curvature sequence.

[0021] Based on the cooperative offset matrix and temporal curvature, the offset difference terms of adjacent time steps are extracted. The temporal curvature is used as a calculation component and is weighted and superimposed with the offset difference terms according to a preset weight relationship. During the weighting process, a channel coupling weight factor is introduced to distinguish the sensitivity of different channels in the calculation. At the same time, a perturbation adjustment factor is combined to balance the relative weight of curvature components and difference components to obtain the energy perturbation matrix.

[0022] A cumulative bias embedding method is constructed by using the energy perturbation matrix. Based on the historical average energy perturbation level, the energy perturbation of each channel pair is weighted and accumulated within a fixed time window W to obtain the cumulative bias result, which is then aggregated into a collaborative embedding feature.

[0023] Furthermore, this method dynamically represents the muscle synergy relationships during rehabilitation training at different levels. On the one hand, by combining the synergy offset matrix with temporal curvature, it captures the changes in muscle synergy patterns over time, sensitively reflecting the acceleration characteristics of abnormal offsets. On the other hand, by introducing channel weighting factors, time weighting factors, and perturbation adjustment factors, it enhances the ability to distinguish the differences in characteristics of different muscle groups and different time periods, effectively avoiding distortion caused by a single indicator. At the same time, the joint construction of the energy perturbation matrix and the cumulative bias embedding allows for the consideration of both instantaneous fluctuations and long-term cumulative effects, thereby characterizing the stability and offset trends of muscle synergy during rehabilitation at multiple scales. This method improves the robustness and comprehensiveness of feature extraction, providing discriminative embedded feature support for subsequent anomaly detection and rehabilitation assessment.

[0024] Preferably, in step S4, by jointly modeling the collaborative feature relationship between any channel pairs and introducing a perturbation weight vector to dynamically adjust the response degree between channels, combined with a set of adjustment mechanisms to enhance numerical stability, a weight value is formed to characterize the strength of the relationship between channels, and an adaptive adjacency matrix is ​​constructed.

[0025] Based on the adaptive adjacency matrix, the perturbation operator matrix is ​​formed to characterize the bidirectional cooperative strength between channels by differentiating the strength of the cooperative strength and introducing an asymmetric adjustment factor to regulate the relative weights between different computational components.

[0026] By perturbating the operator matrix and combining multi-level matrix information from historical moments, tensor outer product operations and element-wise multiplication operations are introduced at the matrix level to establish cross-time interaction structures and historical association paths. At the same time, historical memory weights are set to regulate the degree of jointness between current and past operators, thus constructing tensor interaction superposition results.

[0027] Based on the channel dimension of the tensor interaction superposition result, a channel offset factor is introduced to the corresponding elements at consecutive time steps to perform differential coupling operation; considering the change trend of the tensor in each channel dimension, combined with the tensor mapping result of the current time step and the previous time step, the cross-dimensional responsivity mapping is completed by relying on multiple independent offset factors to generate the offset field coupling matrix.

[0028] Furthermore, an adaptive adjacency matrix, perturbation operator matrix, tensor interaction superposition result, and offset field coupling matrix are constructed to achieve dynamic modeling of the asymmetric synergistic relationship between multi-channel electromyographic signals during rehabilitation training. This process effectively captures the differences in synergistic response and perturbation propagation characteristics between channels, integrates historical state information, and enhances the ability to characterize the nonlinear interaction path between channels. By introducing perturbation weight vectors, asymmetric adjustment factors, historical memory weights, and channel offset factors, multi-scale modeling and temporal sensitivity control of the evolution of relationships between channels are achieved, thereby improving the muscle movement abnormality analysis model's ability to identify potential abnormal synergistic behaviors in complex muscle group synergistic scenarios and providing a more discriminative synergistic embedding structure representation for subsequent abnormality detection and rehabilitation process assessment.

[0029] Preferably, in step S5, the perturbation performance of each channel dimension is extracted based on the offset field coupling matrix, a Gaussian perturbation kernel function with bandwidth adjustment capability is introduced to perform exponential mapping on the channel offset intensity, and the deviation sensitivity response vector is generated dimension by dimension by combining the activation difference between the standard template activation mode and the patient's actual activation mode in the corresponding channel dimension.

[0030] Channel dimension fusion is performed by combining the bias sensitivity response vector and the channel direction weight coefficient. Then, the difference direction between the reference template activation value and the patient activation value is biased and encoded by combining the sign function. Multi-channel direction projection components are further superimposed to complete the weighted normalization operation of the residual direction tensor, and finally the bias intensity tensor of multi-dimensional channel bias response intensity is generated.

[0031] Based on the deviation intensity value at the current moment, a normalization construction strategy with local logarithmic amplitude characteristics is introduced, which is combined with the deviation intensity value at the previous moment to form an anomaly scoring index.

[0032] Furthermore, a joint mapping mechanism of Gaussian perturbation kernel function and channel activation difference is introduced on the basis of the offset field coupling matrix, which can realize fine modeling of perturbation response in the channel dimension and improve the resolution of deviation sensitivity response vector. In addition, a multi-dimensional directional projection relationship is established by combining channel direction weight coefficient and sign function, so that the residual direction tensor can accurately reflect the intensity difference of deviation between channels. At the same time, by introducing a normalization construction strategy with local logarithmic amplitude characteristics, the deviation intensity value of the current time step and the previous time step are dynamically correlated to form a time-continuous anomaly scoring index, thereby realizing the dynamic characterization of channel offset behavior and accurate quantification of anomaly response, providing higher temporal sensitivity and discrimination accuracy for multi-channel collaborative deviation identification.

[0033] Preferably, in step S6, the abnormality scoring index is used to analyze abnormal muscle movement through a two-layer multilayer perceptron.

[0034] The muscle movement anomaly analysis model is trained using a cross-entropy loss function based on probability discrimination, thereby updating the hyperparameters of the muscle movement anomaly analysis model.

[0035] Furthermore, by inputting the final anomaly score into a two-layer multilayer perceptron to achieve nonlinear mapping analysis of muscle movement anomalies, the representational fit between the score and the actual anomaly pattern can be effectively enhanced, improving the model's ability to identify and analyze complex anomaly patterns. Supervised optimization of the anomaly analysis model using a probability-based cross-entropy loss function can achieve discrete alignment between the anomaly probability prediction results and the true anomaly labels. While ensuring the model's convergence stability, this helps improve its generalization performance and discrimination robustness in muscle coordination anomaly identification tasks, thereby enabling automatic tuning and adaptive updating of the key hyperparameters of the muscle movement anomaly analysis model.

[0036] In the above technical solution, the present invention has the following technical effects and advantages:

[0037] 1. This invention proposes to extract asymmetric collaborative offset rate based on muscle collaborative offset tensor and construct a normalized collaborative offset matrix by combining time weight factor and channel weight factor. Furthermore, it derives temporal curvature to enhance the dynamic modeling ability of collaborative relationship, which can realize continuous tracking and accurate perception of the collaborative relationship between muscle groups during rehabilitation training, thereby enhancing the accuracy and time sensitivity of abnormal activation pattern identification.

[0038] 2. This invention introduces an energy perturbation matrix to express perturbation behavior and constructs a cumulative deviation embedding structure to achieve dynamic aggregation of abnormal response information. It can capture the cumulative abnormal performance of muscle group activation over a long period of time while maintaining the response speed of local perturbation changes, thereby enabling effective modeling and analysis of latent abnormal states in the multi-period rehabilitation process.

[0039] 3. This invention establishes an asymmetric collaborative modeling path by constructing an adaptive adjacency matrix and a perturbation operator matrix, and introduces a tensor interaction structure and a channel offset mapping to construct an offset field coupling matrix. This can enhance the expression of spatial interaction features between cross-channel muscle groups, realize the identification and modeling of collaborative abnormal paths in the channel dimension, and ensure the accurate recovery of dynamic transmission relationships between channels and the complete analysis of perturbation transmission mechanisms.

[0040] 4. This invention introduces a Gaussian perturbation kernel function with bandwidth adjustment capability on the basis of the offset field coupling matrix to perform exponential mapping on the channel offset intensity, and constructs a deviation sensitivity response vector and a deviation intensity tensor by combining the sign function and the channel direction weight, so as to realize the multi-dimensional modeling process of the abnormal scoring index with high numerical resolution and direction discrimination capability, effectively distinguishing the edge abnormal state and normal fluctuation in muscle activation.

[0041] 5. This invention inputs the abnormality scoring index into a multilayer perceptron and introduces a probabilistic discriminant loss function based on cross-entropy for model training, thereby enhancing the convergence stability and generalization ability of the muscle movement abnormality analysis model in a high-dimensional nonlinear space, thus realizing the analysis of abnormal muscle group activation throughout the rehabilitation training process. Attached Figure Description

[0042] Figure 1 This is a flowchart of a muscle movement abnormality analysis method for the rehabilitation process provided by the present invention.

[0043] Figure 2 This is a structural diagram of the muscle co-shift tensor constructed according to the present invention.

[0044] Figure 3 This is a structural diagram of the collaborative embedding feature constructed according to the present invention.

[0045] Figure 4 This is a structural diagram of the offset field coupling matrix constructed according to the present invention.

[0046] Figure 5 This is a structural diagram of the anomaly scoring index constructed by the present invention.

[0047] Figure 6 This is a loss convergence curve during the training process of the muscle movement abnormality analysis model provided by this invention.

[0048] Figure 7 This invention provides a visualization of the high-dimensional bias intensity tensor using a dimensionality reduction graph.

[0049] Figure 8 This is a comparison curve of the actual value and the predicted value provided by the present invention. Detailed Implementation

[0050] This invention proposes a method for analyzing muscle movement abnormalities during rehabilitation. First, it extracts asymmetric collaborative offset rates between channels based on a muscle collaborative offset tensor. A normalized collaborative offset matrix is ​​constructed by combining a time weighting factor and a channel pair weighting factor, and the temporal curvature is further derived to enhance the perception of offset change trends. Then, an energy perturbation matrix characterizes the perturbation behavior, and a cumulative deviation embedding structure is used to achieve dynamic integration of abnormal responses. An asymmetric collaborative modeling path between multi-dimensional channels is established using an adaptive adjacency matrix and a perturbation operator matrix. A tensor interaction structure and channel offset mapping are fused to generate an offset field coupling matrix, constructing abnormal performance characteristics in a multi-channel space. Finally, a deviation sensitivity response vector is constructed based on the offset field coupling matrix and a standard template, and a multi-dimensional deviation intensity tensor is generated. An abnormality scoring index is calculated by combining a Gaussian perturbation kernel function with bandwidth adjustment capability, a sign function, and channel direction weights. The abnormality scoring index is used to train and optimize the model through a multilayer perceptron and a cross-entropy-based probabilistic discriminant loss function, enabling multi-dimensional modeling and accurate abnormality analysis of muscle movement abnormalities during rehabilitation.

[0051] Please see Figure 1 As shown in the figure, a method for analyzing abnormal muscle movement during the rehabilitation process is described in the following embodiments of this application.

[0052] S1. Collect muscle movement datasets during rehabilitation testing, and preprocess the collected datasets to construct a muscle movement dataset.

[0053] In this embodiment, the construction process of the muscle movement dataset includes data acquisition and preprocessing. During the data acquisition stage, patients at different rehabilitation stages are selected as the detection subjects. Four standard rehabilitation movements are designed: downward gaze, upward gaze, flexion and extension, and fist clenching, covering a variety of typical movement types. The acquisition device uses a multi-channel surface electromyography (EMG) sensor array. The sensors are placed at key locations on the surface of the target muscle group. The sampling rate per channel is set to 2000Hz, and the sampling accuracy is set to 16bit, to completely record the time-domain waveforms of muscle contraction and relaxation. Simultaneously, muscle kinematic parameters, including triaxial acceleration, triaxial angular velocity, and joint angle changes, are acquired. The sampling frequency of the inertial measurement unit is set to... 200Hz; The start and end times of the movement are marked by writing the wired trigger signal into a unified time base signal channel. All acquisition channels are synchronized based on a unified timestamp. The muscle movement dataset collected during the rehabilitation detection process includes: raw electromyography (EMG) signal, EMG energy, EMG root mean square value, EMG zero-crossing rate, EMG average frequency, EMG median frequency, acceleration component X, acceleration component Y, acceleration component Z, composite acceleration value, angular velocity component X, angular velocity component Y, angular velocity component Z, average angular velocity, peak-to-peak angular velocity, standard deviation of angular velocity, joint angle change, and joint angular velocity, which are used to characterize muscle contraction intensity, contraction rhythm, movement amplitude, and movement speed.

[0054] The preprocessing stage includes baseline correction, outlier removal, signal segmentation, statistical calculation, and numerical standardization of the collected data. First, the average value of each acquired EMG signal segment is calculated, and this average value is subtracted point by point from the signal to bring the initial signal value close to zero. Then, each sampling point in the entire signal segment is averaged with its two adjacent points, and this average is used to replace the original value point by point to reduce abrupt spikes and smooth the waveform. Acceleration and angular velocity data are processed in the same way, with each sampling point replaced by its average with its two adjacent points to eliminate jitter and stabilize fluctuations. All channel signals are divided into equal-length short segments in chronological order, each segment lasting approximately 0.2 seconds, with adjacent segments overlapping by half the time length. Within each time segment, the energy of the EMG signal is calculated by summing the squares of each sample value, calculating the root mean square (RMS) value, averaging the squares, and then taking the square root. The zero-crossing rate is calculated by dividing the number of times the signal crosses zero by the number of sampling points. The average frequency and median frequency are calculated using Fast Fourier Transform (FFT) for each segment's spectrum. The acceleration signal is obtained after calculation. Simultaneously, the average and maximum values ​​of the sampled values ​​in the three directions are taken, and the square root of the sum of the squares in these three directions is calculated to obtain the composite acceleration. Similarly, the average and maximum values ​​of the angular velocity data across the three axes are calculated, and the difference between the maximum and minimum values ​​for each sample segment is calculated to represent the peak-to-peak value. The standard deviation of the fluctuation amplitude is then calculated as a stability index. Joint angle changes are obtained by comparing the composite acceleration and angular velocity of consecutive time slices, approximating the positional difference between two time points as the angle change. The rate of change of angular velocity is calculated by dividing the angle difference between two adjacent time slices by the time interval. All calculation results are arranged in time slice order, and each data point is linearly scaled to the [-1,1] interval based on its minimum and maximum values ​​to ensure a uniform scale for data from different sources, ultimately constructing a muscle movement dataset. The processed muscle movement dataset is arranged in chronological order, with row indices corresponding to the chronological order and column indices corresponding to the feature dimensions. It also retains the category information and start and end times of the movements for input into subsequent rehabilitation muscle abnormality analysis models.

[0055] S2. Divide the muscle motion dataset into muscle motion vectors, calculate the synergistic relationship between muscle groups to obtain a dynamic adjacency matrix sequence, and obtain the muscle synergistic offset tensor based on the structural change rate matrix.

[0056] Furthermore, in step S2, the muscle cooperative offset tensor is constructed, and the process is as follows: Figure 2 As shown, the specific steps for constructing the muscle cooperative offset tensor are as follows.

[0057] S21. The data collected at each time point is divided into muscle motion vectors, and then combined according to a fixed time step to form a muscle motion signal matrix.

[0058] In this embodiment, the mathematical model for the sample data at time t is:

[0059] ;

[0060] Where C represents the number of channels, and its value is 18. Let C be the data value of the Cth channel at time t. Let be the sample at time t; the data is divided into segments with a fixed time step T, where T is 50. Each sample is combined according to the fixed time step to form a muscle movement signal matrix, the mathematical model of which is:

[0061] ;

[0062] in, This is a muscle movement signal matrix, which is a single muscle movement data sample.

[0063] S22. In each time step, the multi-channel muscle motion data features in the muscle motion vector are paired, the co-activation intensity between the channel pairs is calculated, and a stability constant is introduced in the calculation process to avoid numerical calculation errors. Finally, the co-activation intensity of each channel pair is concatenated according to the index position to obtain the muscle group adjacency matrix of the current time step, and a dynamic adjacency matrix sequence is formed by iterative calculation within the entire sampling window.

[0064] In this embodiment, to extract the co-activation intensity of multi-channel electromyography signals during the execution of rehabilitation movements, a muscle group adjacency matrix is ​​constructed based on each time t. This adjacency matrix characterizes the degree of co-activation between any two muscle groups at that time, and its mathematical model is as follows:

[0065] ;

[0066] in, , These are the data values ​​of channel i and channel j at time t; This is a numerical stability constant used to avoid division-by-zero errors; it is preferably set to 10e-8. for norm operations, for norm operations, To determine the co-activation intensity of channels i and j at time t, the co-activation intensity values ​​calculated for each channel are concatenated into a muscle group adjacency matrix. The complete dynamic adjacency graph sequence is obtained through iterative calculation using sliding time steps. Its mathematical model is as follows:

[0067] ;

[0068] in, It is a dynamic adjacency graph sequence. The co-activation intensity values ​​obtained from each channel at time t are concatenated into a muscle group adjacency matrix.

[0069] S23. Based on the dynamic adjacency matrix sequence, calculate the difference in co-activation intensity at different times to obtain the structural change rate. Calculate the difference between adjacent times using the structural change rate to obtain the co-offset change difference of each channel pair. Based on the co-offset change difference combined with the consistency sign function, obtain the asymmetric co-offset rate used to distinguish the direction of co-offset change. Concatenate the asymmetric co-offset rates at each time point to form a complete muscle co-offset tensor.

[0070] In this embodiment, the structural change rate matrix is ​​defined based on the difference in the adjacency matrix of each muscle group in the dynamic adjacency graph sequence, and its mathematical model is as follows:

[0071] ;

[0072] in, The structural change rate matrix consists of the structural change rate at each time step. The internal calculation of the structural change rate represents the cooperative activation intensity of channels a and b at different times. Its mathematical model is as follows:

[0073] ;

[0074] in, Let be the cooperative activation intensity of channels a and b at time t. Let the cooperative activation strength of channels a and b at time t-1 be denoted as . Let a and b be the rates of structural change of channels at time t;

[0075] To describe the asymmetric trend in the evolution of muscle group synergistic structures, the difference in synergistic offset is defined, and its mathematical model is as follows:

[0076] ;

[0077] in, Let be the difference in coordinated offset changes between muscle group pairs a and b at time t. Then, to determine the consistency of the perturbation direction, a direction consistency sign function is introduced, and an asymmetric coordinated offset rate is defined. Its mathematical model is as follows:

[0078] ;

[0079] in, This represents the asymmetric cooperative offset rate of muscle group pairs a and b at time t. A value less than 0 indicates a sudden change in the cooperative direction. The `sign` function is used to determine whether the cooperative directions are consistent. This structure, while maintaining the perturbation amplitude, highlights the asymmetric offset characteristics caused by inconsistent directions. After calculating the asymmetric cooperative offset rate for each channel, the muscle cooperative offset tensor is finally constructed, and its mathematical model is as follows:

[0080] ;

[0081] in, For muscle coordinating offset tensor, Let be the muscle coordination offset rate at time T in the sample.

[0082] S3. Based on the muscle cooperative offset tensor, the asymmetric cooperative offset rate is weighted and normalized to construct a cooperative offset matrix, and the continuous difference is extracted to obtain the temporal curvature. The cooperative offset matrix and the temporal curvature are weighted and superimposed, and channel coupling weights and perturbation adjustment factors are introduced to obtain the energy perturbation matrix and accumulate calculation to construct cooperative embedding features.

[0083] Furthermore, in step S3, collaborative embedding features are constructed, the process of which is as follows: Figure 3 As shown, the specific steps for constructing collaborative embedding features are as follows.

[0084] S31. Based on the muscle co-location tensor, extract the asymmetric co-location rate between channels, introduce channel pair weighting factors in combination with the influence of different muscle groups during rehabilitation training, and combine the time weighting factor set with time decay to perform joint weighting processing on the asymmetric co-location rate; perform normalization operation on the weighting result based on the global maximum and minimum values ​​of the muscle co-location tensor to construct the co-location matrix.

[0085] In this embodiment, the muscle co-shift tensor After initial construction, significant differences in amplitude and uneven distribution over time were observed between different muscle group channels. Therefore, a channel-weighting factor was introduced. With time weighting factor The asymmetric cooperative offset rate is weighted and normalized to obtain the cooperative offset matrix. The channel is weighted by factors The time weighting factor is set based on the importance of the dominant muscle groups in rehabilitation training; An exponential decay function is used to highlight recent abnormal contributions during rehabilitation training. The mathematical model for the synergistic offset matrix is ​​as follows:

[0086] ;

[0087] in, The channel-to-weighting factor is used to differentiate the influence of different muscle groups in anomaly analysis. During training, the channel-to-weighting factor with a larger influence value is set to 1, while the influence value for auxiliary muscle groups is set to 0.6. To obtain the maximum value in the muscle co-shift tensor, To obtain the minimum value in the muscle co-shift tensor, Let c and d be the asymmetric cooperative offset rates of muscle groups at time t. The time-weighted factor has the following mathematical model: ,in This is the attenuation parameter, with a value of 0.01. Let be the normalized cooperative offset matrix for the c-th and d-th pairs of muscle groups at time t.

[0088] S32. Based on the cooperative offset matrix, perform second-order calculations on the continuous differences of adjacent time steps and normalize them to construct a time-series curvature sequence.

[0089] In this embodiment, based on the normalized cooperative offset matrix The second-order difference between adjacent time steps is calculated to obtain the temporal curvature of the channel pair, highlighting the acceleration of the shift value, thereby capturing the inflection point characteristics of muscle group synergy patterns during rehabilitation movement execution. The mathematical model of the temporal curvature of the channel pair is as follows:

[0090] ;

[0091] in, Let be the normalized cooperative offset matrix for the c and d pairs of muscle groups at time t+1; Let be the normalized cooperative offset matrix for the c and d pairs of muscle groups at time t-1; Let be the temporal curvature of channel c and channel d at time t. By introducing the temporal curvature operator, not only can the direction of the shift change be reflected, but the acceleration effect of abnormal inflection points can also be amplified. When a patient experiences movement imbalance or a sudden excessive contraction of a muscle group during rehabilitation training, the curvature value will change significantly, thus providing an early signal for abnormal detection.

[0092] S33. Based on the cooperative offset matrix and temporal curvature, firstly, the offset difference terms of adjacent time steps are extracted. The temporal curvature is used as a calculation component and is weighted and superimposed with the offset difference terms according to a preset weight relationship. During the weighting process, a channel coupling weight factor is introduced to distinguish the sensitivity of different channels in the calculation. At the same time, a perturbation adjustment factor is combined to balance the relative weight of the curvature component and the difference component, thus obtaining the energy perturbation matrix.

[0093] In this embodiment, to characterize the abnormality of muscle group coordinated migration from an energy perspective, a coordinated energy perturbation field modeling method is introduced. Specifically, the temporal curvature and the migration difference are weighted and superimposed to construct an energy perturbation matrix. Its mathematical model is:

[0094] ;

[0095] in, This is the normalized cooperative offset difference term, whose function is to preserve the instantaneous change information of the offset amplitude. This is a perturbation adjustment factor, adaptively learned during training, with an initial value of 0.5. Its role is to balance the relative values ​​of the curvature term and the difference term in energy modeling. These are channel coupling weights, whose values ​​are updated during training. These weights are used to adjust the sensitivity of different channels to temporal curvature. Let be the energy perturbation matrix of channel c and channel d at time t.

[0096] S34. Construct a cumulative deviation embedding method based on the energy perturbation matrix. Using the historical average energy perturbation level as a benchmark, perform weighted accumulation of the energy perturbation of each channel pair within a fixed time window W to obtain the cumulative deviation result, and converge it into a collaborative embedding feature.

[0097] In this embodiment, within the time window W, the cumulative deviation of channel pairs c and d is calculated. The energy disturbance at each moment is compared with its historical average value, and a time decay factor is introduced. The results at different time points are weighted and summed to obtain the cumulative deviation. The mathematical model of the cumulative deviation is as follows:

[0098] ;

[0099] in, Let be the accumulated deviation of channel pairs c and d within window W at time t. This reflects the cumulative disturbance of the channel pairs within window W, where W is the length of the time window, with a value of 40. The historical average energy perturbation level of the channel with respect to c and d. For a moment The energy disturbance value, This is the time decay factor, used to assign weights to the perturbation values ​​within window W. The initial value is set to 0.5, and its value during training is adjusted accordingly. and descent parameters The related exponential decay, The initial value is set to 0.05;

[0100] After obtaining the cumulative deviation of all channel pairs, the cooperative embedding feature at time t is formed by summing their absolute values. The mathematical model is as follows:

[0101] ;

[0102] in, Let be the collaborative embedding feature at time t.

[0103] S4. Based on the collaborative embedding feature, a perturbation weight vector is introduced to obtain an adaptive adjacency matrix and a perturbation operator matrix is ​​constructed with an asymmetric adjustment factor. The matrix information of historical moments is fused to generate a tensor interactive superposition result. A channel offset factor is introduced for differential coupling to form an offset field coupling matrix.

[0104] Furthermore, in step S4, the offset field coupling matrix is ​​constructed, and the process is as follows: Figure 4 As shown, the specific steps for constructing the offset field coupling matrix are as follows.

[0105] S41. By jointly modeling the collaborative feature relationship between any channel pairs and introducing a perturbation weight vector to dynamically adjust the response degree between channels, combined with a set of adjustment mechanisms to enhance numerical stability, a weight value is formed to characterize the strength of the relationship between channels, and an adaptive adjacency matrix is ​​constructed.

[0106] In this embodiment, to characterize the dynamically changing muscle group synergy relationships during rehabilitation training, an adaptive adjacency matrix is ​​constructed based on synergistic embedding features. Each element in the adaptive adjacency matrix is ​​an adaptive adjacency weight. During the calculation process, a perturbation weight vector is introduced to progressively weight and correct the embedding dimensions, and a stability constant is added to the denominator to ensure numerical stability. At time t, the adaptive adjacency weights of channel c and channel d are obtained by multiplying their embedding features and perturbation weights dimension by dimension and then taking the inner product. The mathematical model of the adaptive adjacency weights is as follows:

[0107] ;

[0108] in, Let be the collaborative embedding feature of channel c at time t. Let be the collaborative embedding features of channel d at time t. This is a stability constant, taking the value 10e−6, used to avoid the denominator being zero. This is a perturbation weight vector, initially set to all 1s. Its function is to adjust the importance of feature dimensions and it is updated during training based on the energy perturbation signal. The adaptive adjacency weight represents the dynamic cooperative strength between channels c and d. for The second normal form operation, for The second normal form operation.

[0109] S42. Based on the adaptive adjacency matrix, the perturbation operator matrix is ​​formed to characterize the bidirectional cooperative strength between channels by differentiating the strength of the cooperative strength and introducing an asymmetric adjustment factor to regulate the relative weights between different computational components.

[0110] In this embodiment, after obtaining the adaptive adjacency weight Subsequently, to further characterize the asymmetric activation relationship between different muscle groups during rehabilitation training, an asymmetric perturbation operator was constructed. This asymmetric perturbation operator, by maintaining directional differences and introducing a product term to enhance bidirectional interaction, can characterize the imbalance between channels during the co-activation process. The mathematical model of the asymmetric perturbation operator is as follows:

[0111] ;

[0112] in, The strength of the cooperative relationship from channel c to channel d at time t is derived from the adaptive adjacency matrix of S41, reflecting the degree of cooperative activation of the two channels at that time. Let be the reverse cooperative strength from channel d to channel c at time t, and . They are independent of each other, reflecting the asymmetry inherent in bidirectional coupling; This is an asymmetric adjustment factor used to balance the contributions of the difference and product terms. The initial value is set to 0.3, and it is automatically updated during training through gradient descent. The range is [0,1]. The larger the value, the stronger the influence of the product term. The result of the asymmetric perturbation operator represents the asymmetric cooperative offset of the channel pair (c,d) at time t. The asymmetric perturbation operator results for each channel pair constitute the perturbation operator matrix. .

[0113] S43. By perturbating the operator matrix and combining multi-level matrix information at historical moments, tensor outer product operation and element-wise multiplication operation are introduced at the matrix level to establish cross-time interaction structure and historical association path respectively. At the same time, historical memory weight is set to regulate the degree of joint between current and past operators, and tensor interaction superposition result is constructed.

[0114] In this embodiment, to characterize the temporal continuity and high-order dynamic evolution of muscle group synergy during rehabilitation, a tensor interaction superposition mechanism is proposed based on the aforementioned asymmetric perturbation operator. The electromyographic signals of rehabilitation training typically exhibit a multi-layered progressive relationship in the time series: changes in adjacent time steps reflect short-term muscle activation responses, while changes across time steps reveal long-term adjustment trends between training phases. Traditional first-order difference methods can only capture transient changes and cannot simultaneously consider multi-temporal dependencies. Therefore, this step introduces a combined operation of tensor outer product and element-wise product at the matrix level to form a multi-order structure with historical memory characteristics. The mathematical model of the tensor interaction superposition mechanism is as follows:

[0115] ;

[0116] in, Let be the asymmetric perturbation operator matrix at time t. , These are the asymmetric perturbation operator matrices for the previous time step and the two time steps prior, respectively. The tensor outer product operation maps two matrices to a higher-order tensor structure, and its purpose is to characterize nonlinear interactions across time steps. For element-wise multiplication, it emphasizes the coupling relationship between the current matrix and the historical matrix at the item-by-item level. The historical memory weights are used to control the influence of element-wise product terms. The initial value is set to 0.5, and it is updated during training, ranging from [0,1]. This is the result of tensor interaction overlay, containing interaction information from multiple time points.

[0117] S44. Based on the channel dimension of the tensor interaction superposition result, a channel offset factor is introduced to the corresponding elements at consecutive time steps to perform differential coupling operation; for the change trend of the tensor in each channel dimension, combined with the tensor mapping result of the current time step and the previous time step, the cross-dimensional responsivity mapping is completed by relying on multiple independent offset factors to generate the offset field coupling matrix.

[0118] In this embodiment, to convert the dynamic features in the high-order tensor structure into an energy representation that can be used for anomaly detection, this step, based on the difference operation of the tensor channel dimension of the tensor cross-superposition result and the introduction strategy of the channel offset factor, obtains the offset field coupling matrix through offset field coupling mapping. The muscle group synergy pattern in rehabilitation training has temporal evolution, representing changes in energy coupling between channels. Different channels exhibit complex responses such as synchronous enhancement, phase lag, and signal attenuation at specific movement stages. To capture multi-channel offset patterns, this step generates the offset field coupling matrix by performing layer-by-layer analysis of the channel dimension of the tensor cross-superposition result. The offset field coupling matrix is ​​composed of offset field coupling elements; the mathematical model of the offset field coupling elements is:

[0119] ;

[0120] Where k is the tensor channel dimension index, ranging from [1, 18], and each k corresponds to an independent channel offset factor. ; This is the channel offset factor, used to measure the sensitivity of different dimensions to anomalies. The initial value is randomly distributed in the range of [0.1, 0.9] and adaptively adjusted. The element values ​​of the tensor interaction superposition result in the channel pair (c,d) and dimension k are derived from S43 and reflect the interaction strength of the channel pair in dimension k. Given the tensor interaction values ​​in the same dimension at the previous time step, extract the rate of change of interaction by differing from the current time step; The element values ​​of the offset field coupling matrix represent the energized offset response of the channel pair (c,d) at time t. The element values ​​of the offset field coupling matrix for each channel pair constitute the offset field coupling matrix. .

[0121] S5. By exponentially mapping the channel offset intensity to the offset field coupling matrix and the Gaussian perturbation kernel function, and by fusing the channel differences, a deviation sensitivity response vector is generated. Combined with the channel direction weight coefficient and the residual direction projection component, a deviation intensity tensor is generated and jointly modeled to obtain the anomaly scoring index.

[0122] Furthermore, in step S5, anomaly scoring metrics are constructed, the process of which is as follows: Figure 5 As shown, the specific steps for constructing anomaly scoring indicators are as follows.

[0123] S51. Based on the offset field coupling matrix, the perturbation performance of each channel dimension is extracted. A Gaussian perturbation kernel function with bandwidth adjustment capability is introduced to perform exponential mapping on the channel offset intensity. Combined with the activation difference between the standard template activation mode and the patient's actual activation mode in the corresponding channel dimension, the deviation sensitivity response vector is generated dimension by dimension.

[0124] In this embodiment, when quantitatively measuring the response difference between the standard template activation mode and the patient's actual activation mode, a Gaussian perturbation kernel function is introduced to exponentially modulate the channel response in order to accurately express the perturbation performance of different channel dimensions in the offset field coupling matrix. Specifically, a deviation sensitivity response vector is constructed by combining the offset field coupling matrix and the channel activation difference. The mathematical model of the deviation sensitivity response vector is as follows:

[0125] ;

[0126] in, To control the response sensitivity to channel perturbations, the core width is adjusted, preferably to 0.8. The activation value of the standard template activation mode in the k-th dimension. The activation value of the patient's actual activation mode in the k-th dimension. Let be the bias sensitivity response vector in the k-th channel dimension, representing the bias response in the k-th channel dimension. It is an exponential function.

[0127] S52. Channel dimension fusion is performed by combining the bias sensitivity response vector and the channel direction weight coefficient. Then, the difference direction between the reference template activation value and the patient activation value is biased and encoded by combining the sign function. The multi-channel direction projection components are further superimposed to complete the weighted normalization operation of the residual direction tensor, and finally the bias intensity tensor of the multi-dimensional channel bias response intensity is generated.

[0128] In this embodiment, to further describe the direction of patient deviation from multiple dimensions, a residual weighting mechanism is used to weight the deviation sensitivity response vector generated in the previous step. By fusing the channel direction weights, a multidimensional deviation direction tensor is constructed, and the deviation direction is specified using a sign function, resulting in a deviation intensity tensor. The mathematical model of the deviation intensity tensor is as follows:

[0129] ;

[0130] in, The value of the deviation intensity after projection of the residual direction at time t. These are the channel direction weight coefficients, initially set to 1, and iteratively updated during training. Let k be the bias sensitivity response vector in the k-th channel dimension. For a sign function, when When less than 0, the sign function value is -1; when... When equal to 0, the sign function value is 0; when... When the value is greater than 0, the sign function value is 1, and K is the total number of feature channels with a value of 18.

[0131] S53. Based on the deviation intensity value at the current moment, a normalization construction strategy with local logarithmic amplitude characteristics is introduced, which is combined with the deviation intensity value at the previous moment to form an anomaly scoring index.

[0132] In this embodiment, to comprehensively assess the degree of abnormal deviation of the patient's activation behavior at the temporal level, this step introduces a time-sensitive distribution structure entropy comparison mechanism based on the deviation intensity scalar to achieve dynamic detection of deviation across time periods. This distribution structure entropy comparison mechanism constructs an exponential deviation scoring function based on local logarithmic increase, fusing the current deviation intensity value with the deviation intensity value at the previous time point to form an abnormality scoring index. This index measures the magnitude of the temporal deviation abrupt change between the patient's activation pattern and the standard template. Its mathematical model is as follows:

[0133] ;

[0134] in, The constant term is introduced and has a value of 10e−6 to prevent the denominator from being zero and to ensure calculation stability; the logarithmic term is used to measure the rate of change of the current deviation value relative to the historical deviation, thereby amplifying and normalizing the abnormal amplitude. This is the final anomaly scoring indicator; the larger the value, the more significant the deviation. This represents the deviation intensity value after projection of the residual direction at time t-1.

[0135] S6. Input the muscle movement dataset, combine it with the loss function, construct a muscle movement anomaly analysis model, and iteratively train it until convergence to complete the training of the muscle movement anomaly analysis model.

[0136] S61. The abnormal scoring index is used to analyze abnormal muscle movement through a two-layer multilayer perceptron.

[0137] In this embodiment, a muscle movement anomaly analysis model is constructed. The model is input into a muscle movement dataset and sequentially progresses through multiple stages: muscle co-external offset tensor, co-embedding features, offset field coupling matrix, and anomaly scoring index. Based on this, channel coupling weights, perturbation adjustment factors, perturbation weight vectors, asymmetric adjustment factors, channel offset factors, and channel direction weight coefficients are introduced. To achieve a nonlinear discriminative mapping from the channel scoring vector to the anomaly probability, this step constructs a multilayer perceptron with two-layer mapping capabilities. Embedding space projection is achieved by combining ReLU linear transformation with a Sigmoid nonlinear activation function, ultimately yielding the results of the muscle movement anomaly analysis. .

[0138] S62. The muscle movement abnormality analysis model is trained based on the cross-entropy loss function of probability discrimination to update the hyperparameters of the muscle movement abnormality analysis model.

[0139] In this embodiment, during the training process after the abnormality scoring index is mapped to the deep layer, to ensure the consistency between the predicted output and the actual abnormal state, a cross-entropy loss function based on probability discrimination is adopted. Specifically, the probability value output by the muscle movement abnormality analysis model is compared with the actual label sample by sample, and the mathematical model of the loss function is constructed as follows:

[0140] ;

[0141] in, The output of the muscle movement abnormality analysis model. This is the label of the actual abnormal state corresponding to the current sample. The loss function value is used as the target for backpropagation to update the hyperparameters.

[0142] Furthermore, the muscle movement abnormality analysis model proposed in this invention is implemented using the Python programming language and modeled and trained based on the PyTorch deep learning framework. During model training, the Adam optimizer is used to iteratively update the model parameters, with the initial learning rate set to 0.001. A fixed step size decay strategy is used to dynamically adjust the learning rate to adapt to the model's convergence trend. The training batch size is set to 64, and the total number of training epochs is 2000. The loss function is a probabilistic discriminant loss structure based on cross-entropy, and a residual modulation term is embedded on the abnormality scoring index to enhance the training sensitivity to edge samples. At the same time, the abnormality scoring index is mapped at the output end through two layers of multilayer perceptron, combined with the Sigmoid activation function to achieve a stable output of the abnormality probability, effectively improving the model's training robustness and classification ability for multi-channel deviation responses during rehabilitation.

[0143] Furthermore, the collected muscle movement dataset is input into the constructed muscle movement anomaly analysis model for training and validation. The convergence trend of the joint loss during model training is as follows: Figure 6 As shown, with the increase of training rounds, the loss value of the muscle movement anomaly analysis model continuously decreases from an initial high level and gradually stabilizes, indicating that the proposed muscle movement anomaly analysis model structure has good convergence characteristics during training and can effectively complete the modeling of perturbation features and accurate estimation of anomaly scores. To further verify the model's discriminative ability in anomaly identification, the t-SNE algorithm is used to perform visualization dimensionality reduction on the high-dimensional bias intensity tensor, and the results are shown below. Figure 7 As shown in the figure, different categories of samples form clear boundaries in the two-dimensional embedding space, demonstrating the significant advantage of the constructed offset field coupling mechanism and Gaussian perturbation response joint modeling capability in sample discriminability; the prediction results are as follows. Figure 8 As shown in the figure, the horizontal axis represents the time step, and the vertical axis represents the anomaly score. In the figure, circles represent the anomaly level corresponding to the real label, and triangles represent the model prediction results. It can be observed from the figure that the overall consistency between the output of the muscle movement anomaly analysis model and the real label is high, and the constructed anomaly score index can accurately locate edge samples. Compared with existing methods, the proposed analysis model shows superior performance in terms of anomaly detection accuracy, response speed, and collaborative discrimination ability between cross-channel muscle groups. It can adapt to the diverse activation pattern perturbations in different rehabilitation stages and ensure the reliability and consistency of anomaly monitoring and evaluation results in the rehabilitation process.

[0144] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A muscle movement abnormality analysis method for a rehabilitation process, characterized by, The method comprises the following steps: Collecting a muscle movement data set in a rehabilitation detection process, and preprocessing the collected data set to construct a muscle movement data set; Dividing the muscle movement data set into muscle movement vectors, calculating the inter-muscle group coordination relationship to obtain a dynamic adjacency matrix sequence, and obtaining a muscle coordination shift tensor according to a structure change rate matrix; According to the muscle coordination shift tensor, the asymmetric coordination shift rate is weighted and normalized to construct a coordination shift matrix, and the continuous difference is extracted to obtain a time sequence curvature, the coordination shift matrix and the time sequence curvature are weighted and superimposed, and a channel coupling weight and a disturbance adjustment factor are introduced to obtain an energy disturbance matrix and cumulative calculation, and a coordination embedding feature is constructed; Based on the coordination embedding feature, a disturbance weight vector is introduced to obtain an adaptive adjacency matrix and construct a disturbance operator matrix with an asymmetric adjustment factor, fuse the matrix information at the historical time, generate a tensor interaction superposition result, introduce a channel shift factor for difference coupling to form a shift field coupling matrix; Through the shift field coupling matrix and the Gaussian disturbance kernel function, the exponential mapping of the channel shift strength is performed, and the channel difference is fused to generate a bias sensitivity response vector, the channel direction weight coefficient and the residual direction projection component are combined to generate a bias intensity tensor and jointly model to obtain an abnormal score index; Input the muscle movement data set, and construct a muscle movement anomaly analysis model combined with a loss function, and iteratively train until convergence to complete the training of the muscle movement anomaly analysis model.

2. The muscle movement abnormality analysis method for a rehabilitation process according to claim 1, characterized by, In the data collection stage, patients at different rehabilitation stages are selected as detection objects to form a feature set containing original electromyographic signals, electromyographic energy, root mean square values, zero-crossing rates, average frequencies, median frequencies, three-axis accelerations, three-axis angular velocities, and joint angle characteristics; In the data preprocessing stage, the collected data is baseline corrected; the electromyographic signal is denoised; the acceleration and angular velocity signals are smoothed and zero-point calibrated; the collected data is time-synchronized and the synchronized signals are cut into time slices; the statistical features of electromyography, acceleration and angular velocity are extracted in each time slice; The processed features are organized into a muscle movement data set in chronological order and feature dimension.

3. The rehabilitation process-oriented muscle movement abnormality analysis method according to claim 2, characterized by, The data collected at each time point is divided into a muscle movement vector, and combined according to a fixed time step to form a muscle movement signal matrix; In each time step, the multi-channel muscle movement data features in the muscle movement vector are paired, the coordination activation strength between the channel pairs is calculated, and a stability constant is introduced in the calculation process to avoid numerical calculation errors, and finally the coordination activation strength of each channel pair is spliced according to the index position to obtain the muscle group adjacency matrix of the current time step, and the dynamic adjacency matrix sequence is formed through iterative calculation in the entire sampling window; Based on the dynamic adjacency matrix sequence, the structural change rate is calculated by calculating the difference of the synergistic activation strength at different time points, the difference between adjacent time points is calculated by the structural change rate, the synergistic offset change difference of each channel pair is obtained, based on the synergistic offset change difference combined with the consistency symbol function, the asymmetric synergistic offset rate for distinguishing the synergistic change direction is obtained, and the asymmetric synergistic offset rate at each time point is spliced in turn to form a complete muscle synergistic offset tensor.

4. The abnormal muscle movement analysis method for a rehabilitation process according to claim 3, characterized by, Based on the muscle synergistic offset tensor, the asymmetric synergistic offset rate between channels is extracted, the channel pair weight factor is introduced in combination with the influence degree of different muscle groups in the rehabilitation training process, and the time weight factor is set in combination with the time decay, and the asymmetric synergistic offset rate is jointly weighted; According to the global maximum and minimum values of the muscle synergistic offset tensor, the weighted result is normalized to obtain a synergistic offset matrix; Based on the synergistic offset matrix, the second-order calculation is performed on the continuous difference of adjacent time steps, and normalization processing is performed to obtain a time sequence curvature sequence; According to the synergistic offset matrix and the time sequence curvature, the offset difference item of adjacent time steps is extracted, and the time sequence curvature is used as a calculation component, and the offset difference item is weighted and superimposed according to a preset weight relationship; In the weighting process, the channel coupling weight factor is introduced to distinguish the sensitivity of different channel pairs in the calculation, and the disturbance adjustment factor is combined to balance the relative proportion of the curvature component and the difference component, and an energy disturbance matrix is obtained. Through the energy disturbance matrix, a cumulative deviation embedding method is constructed, the energy disturbance of each channel pair is weighted and accumulated in a fixed time window W based on the historical average energy disturbance level, a cumulative deviation result is obtained, and is converged as a synergistic embedding feature.

5. The rehabilitation process-oriented muscle movement abnormality analysis method according to claim 4, characterized by, The synergistic feature relationship between any channel pairs is jointly modeled, the response degree between channels is dynamically regulated by introducing a disturbance weight vector, a set of adjustment mechanisms for enhancing numerical stability is combined to form a weight value for representing the relationship strength between channels, and an adaptive adjacency matrix is constructed. Based on the adaptive adjacency matrix, the bidirectional synergistic strength between channels is described differently, and an asymmetric adjustment factor is introduced to regulate the relative weight between different calculation components to form a disturbance operator matrix for representing asymmetric synergistic disturbance. Through the disturbance operator matrix, multi-level matrix information at historical time points is combined, a tensor outer product operation and an element-by-element multiplication operation are introduced at the matrix level, an interaction structure across time points and a historical association path are respectively established, a historical memory weight is set to regulate the joint degree between the current and past operators, and a tensor interaction superposition result is constructed. Based on the channel dimension of the tensor interaction superposition result, a channel offset factor is introduced for difference coupling operation on the corresponding elements at consecutive time points. For the change trend of the tensor in each channel dimension, the tensor mapping results of the current and the previous time point are combined to complete the response degree mapping across dimensions relying on multiple independent offset factors to generate an offset field coupling matrix.

6. The rehabilitation process-oriented muscle movement abnormality analysis method according to claim 5, characterized by, Based on the extraction of the disturbance performance of each channel dimension based on the offset field coupling matrix, a Gaussian disturbance kernel function with bandwidth adjustment capability is introduced to perform exponential mapping on the channel offset intensity, and the activation difference between the standard template activation pattern and the actual activation pattern of the patient in the corresponding channel dimension is combined to generate a bias sensitivity response vector dimension by dimension; Through the channel dimension fusion of the bias sensitivity response vector and the channel direction weight coefficient, followed by the bias coding of the difference direction of the reference template activation value and the patient activation value combined with the sign function, further superimposing the multi-channel direction projection component, completing the weighted normalization operation of the residual direction tensor, and finally generating the bias intensity tensor of the multi-dimensional channel bias response intensity; Based on the bias intensity value at the current moment, a normalization construction strategy with local logarithmic amplitude variation characteristics is introduced, and the bias intensity value at the previous moment is combined to form an abnormal score index.

7. The rehabilitation process-oriented muscle movement abnormality analysis method according to claim 6, characterized by, The abnormal score index is used to realize the muscle movement abnormality analysis through two layers of multilayer perception machine; Based on the cross entropy loss function of probability discrimination, the muscle movement abnormality analysis model is trained to realize the update of the muscle movement abnormality analysis model super parameter.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the steps of the muscle movement abnormality analysis method for rehabilitation process in any one of claims 1 to 7.

Images