Intelligent analysis system and method for orthopedic nursing

Abstract

The invention discloses an intelligent analysis system and method for orthopedic nursing, and relates to the technical field of medical information, and the method comprises the steps: collecting and preprocessing medical image data, combining the daily motion state data of a patient, constructing a multi-dimensional data set containing skeleton morphological features and motion features, extracting data features, and carrying out the analysis of the data. Adaptive fusion is carried out by adopting an attention mechanism, and a digital feature model containing static morphological features and dynamic motion features is established; based on the digital feature model, joint stress features and coupling features are obtained, and a comprehensive evaluation index system is constructed to achieve quantitative evaluation of the bone health condition; and according to the evaluation index system, constructing multi-level risk early warning to carry out real-time monitoring and early warning on abnormal conditions, and meanwhile, based on the rehabilitation progress evaluation indexes, adaptively generating personalized rehabilitation schemes and exercise suggestions. According to the rehabilitation parameter adjusting method combined with double-objective optimization, the safety of the rehabilitation process is guaranteed, and the pertinence of the rehabilitation effect is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and in particular to an intelligent analysis system and method for orthopedic care. Background Art

[0002] With the aging of the population and the increase in sports injuries, the demand for rehabilitation care for orthopedic diseases is growing. Traditional orthopedic rehabilitation care methods mainly rely on the experience and judgment of medical staff, lack quantitative evaluation standards and personalized program formulation mechanisms, and it is difficult to accurately grasp the patient's rehabilitation status and progress. At the same time, existing rehabilitation assessment methods often only focus on a single physiological indicator and fail to fully consider the coupling between tissues such as bones, muscles, and ligaments, resulting in unsatisfactory rehabilitation effects. In addition, the risk warning mechanism in the rehabilitation process is imperfect, which is prone to secondary injuries.

[0003] Although there are some intelligent rehabilitation equipment and evaluation systems, they generally have problems such as weak theoretical basis, single evaluation indicators, and low degree of personalization. Therefore, it is urgent to develop an intelligent orthopedic rehabilitation nursing program based on biomechanical models and artificial intelligence technology, and to achieve accurate evaluation and personalized guidance of the rehabilitation process by establishing a scientific evaluation system and early warning mechanism, so as to improve the safety and effectiveness of rehabilitation nursing. In response to the above problems, the present invention proposes an intelligent orthopedic rehabilitation nursing program that integrates biomechanical modeling, deep learning evaluation, and multi-level early warning mechanism, which provides a new solution to the key technical problems in orthopedic rehabilitation nursing. Summary of the invention

[0004] In view of the problems existing in the existing intelligent analysis methods for orthopedic care, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to establish a scientific evaluation system and early warning mechanism to achieve accurate evaluation and personalized guidance of the rehabilitation process and improve the safety and effectiveness of rehabilitation care.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, an embodiment of the present invention provides an intelligent analysis method for orthopedic care, which includes collecting and preprocessing medical imaging data, combining the patient's daily motion status data, constructing a multidimensional data set including bone morphological features and motion features, extracting data features, and using an attention mechanism for adaptive fusion to establish a digital feature model including static morphological features and dynamic motion features; based on the digital feature model, analyzing joint forces through an orthogonal projection matrix, and hierarchically analyzing the degree of tissue stress coupling, obtaining joint force characteristics and coupling characteristics, and constructing a comprehensive evaluation index system to achieve quantitative evaluation of bone health status; based on the evaluation index system, constructing a multi-level risk warning system to monitor and warn of abnormal conditions in real time, and at the same time, based on the rehabilitation progress evaluation indicators, adaptively generating personalized rehabilitation plans and exercise recommendations.

[0007] As a preferred solution of the intelligent analysis method for orthopedic care of the present invention, wherein: the medical image data includes X-ray image data , CT image data and MRI data ; The X-ray image data Used to obtain two-dimensional projection features of bones; The CT image data Used to obtain cross-sectional features of bones; The magnetic resonance imaging data Used to obtain the characteristics of soft tissue around bones; The three types of image data are input into the preset medical image processing model for image enhancement and feature extraction.

[0008] As a preferred solution of the intelligent analysis method for orthopedic care described in the present invention, the medical image processing model constructs a medical image enhancement network with three branches for parallel processing, the first branch processes X-ray image data , the second branch processes CT image data , the third branch processes MRI data ,Each branch includes image preprocessing, feature enhancement and feature fusion; The image preprocessing is to normalize the input medical image data, and use the adaptive histogram equalization method to improve the image contrast according to the normalized data, where the contrast limit parameter Dynamic adjustment is performed, expressed as: ; in, is the reference contrast limit parameter, is the standard deviation of the grayscale in the local area, is the reference standard deviation; Feature enhancement is performed by constructing residual attention, which is expressed as: ; in, is the intermediate feature map, represents the convolution operation, is the normalized input image, and is the weight of the fully connected layer, is the global average pooling, is the attention weight map, express Activation function, is the enhanced feature map; Feature extraction is performed through a multi-scale feature pyramid network, which includes 4 scale layers. The size of the feature map of each scale layer is halved, and the number of feature channels is doubled. The feature map calculation formula is expressed as: ; in, For the Layer feature pyramid features, is the corresponding backbone network feature, represents a 1×1 convolution operation, Indicates upsampling the feature map of the previous layer. is the multi-scale feature of the final output, It is a 3×3 convolution operation; The self-attention mechanism is used to fuse the features of the three branches, expressed as: ; ; ; in, , , Respectively represent the linear transformation , , matrix, is the input feature, , , is the weight matrix, is the query vector, is the key vector, is a value vector, is the feature dimension, is the fused feature, is the attention output of the X-ray branch, is the attention output of the CT branch, is the attention output of the NMR branch, It means concatenating the features of the three branches in the channel dimension; The fused features are restored to the original image resolution through the deconvolution network, and the residual connection is used to retain the detail information, which is expressed as: ; ; in, is the deconvolution operation, It is the enhanced feature map of the final output.

[0009] As a preferred solution of the intelligent analysis method for orthopedic care described in the present invention, wavelet transform enhancement and Canny edge detection are performed on the image data based on the medical image processing model, wherein the wavelet transform can analyze the local features of the image at different scales and is suitable for denoising and enhancement of medical images, which is expressed as: ; in, represents the wavelet transform coefficients, is the scale parameter; is the translation parameter; is the wavelet basis function; The image signal is input; the bone contour feature data is obtained through wavelet transform and edge detection ; A wearable sensor array consisting of a three-axis acceleration sensor, a three-axis gyroscope, and an angle sensor is installed on the patient's joints to collect joint angle data during daily activities such as standing, walking, and going up and down stairs. , motion acceleration data and gait feature data ; The acquired motion data is normalized and Kalman filtered. The Z-score method is used to eliminate the dimension effect in the normalization process. Kalman filtering is used to suppress sensor noise and motion artifacts. Its state space equation is expressed as: ; ; in, for The system state vector at time , for The system state vector at time ; is the state transfer matrix; is the control matrix; is the control vector; is process noise, with mean 0 and covariance Gaussian distribution of is the observation vector; is the observation matrix; is the observation noise, with mean 0 and covariance Gaussian distribution; smooth motion state feature data is obtained by filtering ; The bone contour feature data Input into the preset deep convolutional neural network for feature extraction. The network contains 5 convolutional layers and 3 fully connected layers. The calculation expression of each convolutional layer is: ; in, For the Layer feature map; is the convolution kernel weight matrix; is the feature map of the previous layer; is the bias term; It is a modified linear unit activation function used to introduce nonlinear characteristics; through multi-layer convolution operations and pooling operations, the bone morphology feature vector is extracted ; The motion state characteristic data The input is sent to the preset long short-term memory network for time series feature analysis, and the motion state features are analyzed through the LSTM network. Perform time series modeling, and finally the network processes the time series motion features into motion feature vectors ; The bone morphology feature vector and the motion feature vector Adaptive fusion is performed through the attention mechanism, and the calculation formula of the attention mechanism is: ; in, is the attention weight vector, indicating the importance of different features; is the weight matrix; represents the concatenation of feature vectors; is the fused feature vector; , are the attention weight coefficients of the two features respectively; through the weighted fusion of the attention mechanism, a digital feature model including static morphological features and dynamic motion features is constructed.

[0010] As a preferred embodiment of the intelligent analysis method for orthopedic care of the present invention, the calculation of the joint force characteristics is to transform the characteristic vector Mapped to the joint force space, the calculation formula is: ; in, for Dimensional joint force vector, indicating the force state of each key position, for dimensional orthogonal projection weight matrix, is the feature vector dimension, is the joint force dimension, for dimensional input feature vector, containing morphological and motion information, for dimensional bias vector, for dimensional identity matrix; Based on the joint force vector , combined with the hierarchical characteristics of the skeletal system, a multi-scale stress-strain analysis model is constructed, which is expressed as: ; ; ; in, For the The stress value of the layered structure, For the The stress transfer coefficient of the layered organization, For the The stress-bearing area of ​​the layer tissue, For the The strain value of the layer organization, For the The elastic modulus of the layered tissue, for dimensional tissue stress-strain vector, is the number of tissue layers; The degree of tissue stress coupling is used to evaluate the interaction between different tissue layers, and a hierarchical stress coupling analysis is performed, which is expressed as: ; ; in, For the Layer and Stress coupling coefficient between layers; and are the stress values ​​of adjacent layers, Indicates The stress value of the layer, for dimensional inter-layer coupling eigenvector.

[0011] As a preferred solution of the intelligent analysis method for orthopedic care described in the present invention, the stress-strain characteristics are obtained based on the multi-scale stress-strain analysis model; the coupling characteristics are obtained based on the stress coupling analysis; and a comprehensive evaluation index system is established based on the stress-strain characteristics and the coupling characteristics, which is expressed as: ; ; ; in, is an indicator of organizational status, is the coupling status indicator, For the Critical reference value of layer stress, For the Critical reference value of layer strain, For the Layer and +1The correlation coefficient between layers, For the Layer coupling critical reference value, , For the Layer stress-strain weight coefficient, For the Layer coupling weight coefficient, , is the comprehensive weight of the state index and coupling index, It is the final health assessment indicator; Further, analyze the health assessment indicators The changes in the time dimension can quantitatively evaluate and predict the patient's recovery status. The formula is: ; ; in, for The improvement of the state at a moment relative to the previous moment, For the The loss value at the moment, For the The loss value at the moment, For the Time to The decay weight at time, is the initial state value, is the current rehabilitation progress indicator, To evaluate the time point, for The amount of improvement in status at a given moment.

[0012] As a preferred solution of the intelligent analysis method for orthopedic care described in the present invention, the multi-level risk warning is to design a three-level risk warning threshold matrix M based on health assessment indicators and rehabilitation progress indicators, which is expressed as: ; in, and Respectively Level health status and rehabilitation progress warning thresholds, Indicates the risk level; Based on health assessment indicators and rehabilitation progress indicators , calculate the risk assessment vector , get the health status risk and risk of rehabilitation progress , expressed as: ; ; ; in, is the health status risk, The risk of recovery progress is is the reference value for health status, is the expected rehabilitation progress value, and is the corresponding standard deviation parameter; Based on risk assessment vector , and compare it with the threshold matrix Compare each level to get a comprehensive warning level , expressed as: ; in, is the indicator function, when the risk assessment vector More than Level Threshold Matrix The value is 1 when it is, otherwise it is 0; Based on stress-strain vector and the coupled eigenvector , establish the objective function of rehabilitation parameter optimization, expressed as: ; ; in, represents the objective function of optimization, is the rehabilitation parameter vector, For the optimized rehabilitation parameters, is the target stress-strain state, For the optimal coupling state, and is the balance coefficient; Based on the optimized rehabilitation parameters Generate personalized rehabilitation recommendations , expressed as: ; in, For the Items of rehabilitation recommendation parameters, .

[0013] As a preferred solution of the intelligent analysis method for orthopedic care described in the present invention, wherein: based on the health assessment indicators and rehabilitation progress indicators, adaptively generating personalized rehabilitation plans and exercise suggestions, including: setting thresholds for health assessment indicators, including the first-level thresholds for health assessment and the second level threshold for health assessment , and set the expected rehabilitation progress benchmark as ; The threshold of rehabilitation progress risk is set based on the rehabilitation progress risk. , set the health status risk warning threshold to ; If the health assessment index S < the first level threshold , and rehabilitation progress indicators Higher than expected progress value , then the training parameters Adjust to the low load range, according to the coupling eigenvector Set up training action groups based on risk assessment vectors Determine the frequency of training; If the first level threshold ≤Health assessment index S≤Second level threshold ,and , is the preset tolerance value, then according to the optimization parameters Adjust the training intensity to a medium level, based on the coupled eigenvector Optimize training movement angles and follow rehabilitation recommended vectors Set up training cycles; If the health assessment index S> the second level threshold , then immediately pause the current training plan and reduce the training parameters To the lowest load level, increase the monitoring frequency of the health assessment index S, and re-evaluate the stress-strain vector The various indicators are based on the risk assessment vector Adjust rehabilitation strategies; If the recovery progress risk >Threshold of risk of rehabilitation progress , and the health status risk <Health status risk warning threshold , then according to the stress-strain vector Develop a progressive training plan by coupling feature vectors Adjust the training action combination according to the risk warning level Set evaluation cycles; If the risk warning level When the preset risk warning threshold is reached, the objective function will be optimized The training parameters are adjusted to a safe range, the monitoring frequency of the rehabilitation evaluation index S is increased, and the rehabilitation recommendation vector is recalculated. The training parameters of .

[0014] In the second aspect, an embodiment of the present invention provides an intelligent analysis system for orthopedic care, which includes: an acquisition module, which is used to acquire and preprocess medical imaging data, and construct a multidimensional data set containing bone morphological features and motion features in combination with the patient's daily motion status data, and use deep learning methods to extract key features and establish a digital feature model of the patient's skeletal system; an evaluation module, based on the digital feature model, by analyzing the joint force characteristics and the degree of tissue stress coupling, integrating and analyzing the medical imaging features and motion parameters, calculating the joint force distribution and stress-strain relationship, achieving quantitative evaluation of bone health status, and establishing an evaluation index system; an early warning module, based on the evaluation index system, constructing a multi-level risk warning to monitor and warn abnormal conditions in real time, and at the same time, based on the rehabilitation progress evaluation index, adaptively generating personalized rehabilitation plans and exercise recommendations.

[0015] The beneficial effects of the present invention are: The present invention transforms complex physiological states into quantifiable mathematical expressions by establishing a biomechanical model in the orthopedic rehabilitation process, constructs a health status assessment system containing static and dynamic features, and achieves accurate assessment of the patient's rehabilitation status through feature fusion and dynamic update mechanisms; through the early warning mechanism of the risk assessment vector and the three-level threshold matrix, combined with the dual-objective optimization rehabilitation parameter adjustment method, it not only ensures the safety of the rehabilitation process, but also improves the pertinence of the rehabilitation effect. It not only overcomes the problems of strong experience dependence and low personalization in traditional rehabilitation nursing, but also makes the formulation and adjustment of rehabilitation plans explainable and traceable through the support of mathematical models, providing a new technical path for the standardization and intelligence of orthopedic rehabilitation nursing, and has significant clinical application value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them: Figure 1 Flowchart of the intelligent analytics approach for orthopedic care. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0018] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0020] Example 1: Reference Figure 1 , which is the first embodiment of the present invention, and provides an intelligent analysis method for orthopedic care, comprising: S1: Collect and preprocess medical imaging data, combine it with the patient's daily motion status data, construct a multidimensional data set containing bone morphological features and motion features, extract data features, use the attention mechanism for adaptive fusion, and establish a digital feature model containing static morphological features and dynamic motion features.

[0021] Collect X-ray image data of patients , CT image data and MRI data , X-ray image data Mainly used to obtain the two-dimensional projection features of bones, CT image data Used to obtain bone cross-sectional features and magnetic resonance imaging data Used to obtain the characteristics of soft tissue around bones, the three types of image data are input into the preset medical image processing model for subsequent image enhancement and feature extraction.

[0022] The medical image processing model includes: constructing a three-branch parallel processing medical image enhancement network, the first branch processes X-ray image data , the second branch processes CT image data , the third branch processes MRI data ,Each branch contains image preprocessing, feature enhancement and feature fusion.

[0023] Image preprocessing is to normalize the input medical image data and map the pixel values ​​to the [0,1] interval. Adaptive histogram equalization is used to enhance the image contrast based on the normalized data, where the contrast limit parameter Dynamic adjustment is performed, expressed as: ; in, is the reference contrast limit parameter, is the standard deviation of the grayscale in the local area, is the reference standard deviation.

[0024] Feature enhancement is performed by constructing residual attention, which is expressed as: ; in, is the intermediate feature map, represents the convolution operation, is the normalized input image, and is the weight of the fully connected layer, is the global average pooling, is the attention weight map, express Activation function, is the enhanced feature map.

[0025] Feature extraction is performed through a multi-scale feature pyramid network, which includes 4 scale layers. The size of the feature map of each scale layer is halved, and the number of feature channels is doubled. The feature map calculation formula is expressed as: ; in, For the Layer feature pyramid features, is the corresponding backbone network feature, represents a 1×1 convolution operation, Indicates upsampling the feature map of the previous layer. is the multi-scale feature of the final output, It is a 3×3 convolution operation.

[0026] The self-attention mechanism is used to fuse the features of the three branches, expressed as: ; ; ; in, , , Respectively represent the linear transformation , , matrix, is the input feature, including feature maps from three branches ( =1,2,3 correspond to X-ray, CT and MRI respectively), , , is the weight matrix, is the query vector, is the key vector, is a value vector, is the feature dimension, is the fused feature, is the attention output of the X-ray branch, is the attention output of the CT branch, is the attention output of the NMR branch, It means concatenating the features of the three branches in the channel dimension.

[0027] The fused features are restored to the original image resolution through the deconvolution network, and the residual connection is used to retain the detail information, which is expressed as: ; ; in, is the deconvolution operation, It is the enhanced feature map of the final output.

[0028] The image data is enhanced by wavelet transform and Canny edge detection through the medical image processing model. The wavelet transform can analyze the local features of the image at different scales and is suitable for denoising and enhancement of medical images. It is expressed as: ; in, represents the wavelet transform coefficients, is the scale parameter, which is used to control the expansion and contraction of the wavelet; is the translation parameter, used to control the translation of the wavelet; As the wavelet basis function, Daubechies wavelet is selected; The image signal is input; the bone contour feature data is obtained through wavelet transform and edge detection , which contains the morphological characteristics of the bones.

[0029] A wearable sensor array consisting of a three-axis acceleration sensor, a three-axis gyroscope, and an angle sensor is installed on the patient's joints to collect joint angle data during daily activities such as standing, walking, and going up and down stairs. , motion acceleration data and gait feature data ,The sampling frequency was set to 100 Hz to ensure the temporal resolution of the data, and the ,acquisition time was no less than 30 minutes to obtain sufficient motion ,characteristics.

[0030] The acquired motion data is normalized and Kalman filtered. The Z-score method is used to eliminate the dimension effect in the normalization process. Kalman filtering is used to suppress sensor noise and motion artifacts. Its state space equation is expressed as: ; ; in, for The system state vector at time , including position and velocity information, for The system state vector at time ; is the state transfer matrix, describing the dynamic characteristics of the system; is the control matrix; is the control vector; is process noise, with mean 0 and covariance Gaussian distribution of is the observation vector; is the observation matrix; is the observation noise, with mean 0 and covariance Gaussian distribution; smooth motion state feature data is obtained by filtering ; The bone contour feature data Input into the preset deep convolutional neural network for feature extraction. The network contains 5 convolutional layers and 3 fully connected layers. The calculation expression of each convolutional layer is: ; in, For the Layer feature map; is the convolution kernel weight matrix; is the feature map of the previous layer; is the bias term; It is a modified linear unit activation function used to introduce nonlinear characteristics; through multi-layer convolution operations and pooling operations, the bone morphology feature vector is extracted ; The motion state characteristic data The input is sent to the preset long short-term memory network for time series feature analysis, and the motion state features are analyzed through the LSTM network. The core of timing modeling is to use three gating units (forget gate , Input Gate and output gate ) to control the flow of information.

[0031] The forget gate determines the degree of retention of historical information, the input gate controls the influence of the current input, and the output gate manages the state output. Store long-term memory and use hidden states Output the feature representation of the current moment. Finally, the network processes the temporal motion features into motion feature vectors , which contains the temporal dependencies of the motion sequence.

[0032] The bone morphology feature vector and the motion feature vector Adaptive fusion is performed through the attention mechanism, and the calculation formula of the attention mechanism is: ; in, is the attention weight vector, indicating the importance of different features; is the weight matrix; represents the concatenation of feature vectors; is the fused feature vector; , are the attention weight coefficients of the two features respectively; through the weighted fusion of the attention mechanism, a digital feature model including static morphological features and dynamic motion features is constructed.

[0033] The attention mechanism is used to adaptively fuse static skeleton features and dynamic motion features, and output a comprehensive feature vector containing complete static structural information and dynamic functional features. , which can support more accurate personalized model construction and status assessment.

[0034] S2: Based on the digital feature model, the joint force is analyzed through the orthogonal projection matrix, and the degree of tissue stress coupling is analyzed hierarchically to obtain the joint force characteristics and coupling characteristics, and to construct a comprehensive evaluation index system to achieve quantitative evaluation of bone health status.

[0035] In order to ensure the accuracy and independence of the node force feature extraction, the feature vector Mapped to the joint force space. Orthogonal projection ensures the orthogonality between the force features of the joints in different directions and avoids feature redundancy. The calculation formula is: ; in, for Dimensional joint force vector, indicating the force state of each key position, for dimensional orthogonal projection weight matrix, is the feature vector dimension, is the joint force dimension, for dimensional input feature vector, containing morphological and motion information, for dimensional offset vector, used to adjust the baseline value of the mechanical feature, for dimensional identity matrix, ensuring orthogonality.

[0036] Orthogonal projection is used to ensure the independence of the stress characteristics of nodes in different dimensions, avoiding feature redundancy and providing reliable input for subsequent stress-strain analysis.

[0037] Based on the obtained joint force vectors and considering the hierarchical characteristics of the skeletal system, a multi-scale stress-strain analysis model is constructed, which is expressed as: ; ; ; in, For the The stress value of the layer organization reflects the stress state per unit area. For the The stress transfer coefficient of the layered organization represents the efficiency of force transfer. For the The stress-bearing area of ​​the layer tissue, considering the anatomical structure characteristics, For the The strain value of the layer organization indicates the degree of deformation. For the The elastic modulus of the layered structure characterizes the mechanical properties of the material. for dimensional tissue stress-strain vector, containing complete mechanical information of each layer, The number of tissue layers, including bones, cartilage, ligaments, etc.

[0038] Through the multi-scale stress-strain analysis model, the multi-level structure from bones to soft tissues is taken into account, the force transfer process is characterized by the stress transfer coefficient, and a complete mechanical response chain is established.

[0039] In order to evaluate the interaction between different tissue layers, a hierarchical stress coupling analysis is performed, and the formula is expressed as: ; ; in, For the Layer and The stress coupling coefficient between layers is in the range of [0,1]; and are the stress values ​​of adjacent layers, Indicates The stress value of the layer, for dimensional inter-layer coupling eigenvector.

[0040] The coupling coefficient is calculated by normalizing the dot product of stress, reflecting the degree of coordination of mechanical responses between adjacent tissue layers and providing a basis for evaluating the state of tissue interfaces.

[0041] According to the obtained stress-strain characteristics and coupling characteristics, a comprehensive evaluation index system is established, which is expressed as: ; ; ; in, As an indicator of tissue status, it evaluates the mechanical status of each layer. is a coupling status indicator to evaluate the inter-layer synergy. For the Critical reference value of layer stress, For the Critical reference value of layer strain, For the Layer and +1The correlation coefficient between layers, For the Layer coupling critical reference value, , For the Layer stress-strain weight coefficient, For the Layer coupling weight coefficient, , is the comprehensive weight of the state index and coupling index, is the final health assessment index, ranging from [0,1].

[0042] The tissue status and coupling effect are unified through a comprehensive evaluation index system, and the importance of each factor is adjusted through the weight coefficient to achieve a comprehensive health status assessment.

[0043] At the same time, in order to evaluate the treatment effect, health assessment indicators are analyzed The changes in the time dimension can quantitatively evaluate and predict the patient's recovery status. The formula is: ; ; in, for The improvement of the state at a moment relative to the previous moment, For the The loss value at the moment, For the The loss value at the moment, For the Time to The decay weight at time, is the initial state value, is the current rehabilitation progress indicator, To evaluate the time point, for The amount of improvement in status at a given moment.

[0044] By analyzing health assessment indicators Changes in the time dimension enable dynamic quantitative evaluation of the patient's rehabilitation process, provide patients with clear feedback on rehabilitation progress, and enhance treatment compliance; at the same time, by predicting rehabilitation trends through data analysis, accurate management of the treatment process is achieved, ultimately improving the overall effect of orthopedic rehabilitation treatment.

[0045] S3: Based on the comprehensive evaluation indicators, a multi-level risk warning system is constructed to conduct real-time monitoring and early warning of abnormal conditions. At the same time, based on the rehabilitation progress evaluation indicators, personalized rehabilitation plans and exercise recommendations are adaptively generated.

[0046] According to health assessment indicators and rehabilitation progress indicators , design the three-level risk warning threshold matrix M, expressed as: ; in, and Respectively Level health status and rehabilitation progress warning thresholds, Indicates the risk level ( =1,2,3), the larger the value, the higher the risk level.

[0047] Based on health assessment indicators and rehabilitation progress indicators , calculate the risk assessment vector , get the health status risk and risk of rehabilitation progress , expressed as: ; ; ; in, is the health status risk, The risk of recovery progress is is the reference value for health status, is the expected rehabilitation progress value, and is the corresponding standard deviation parameter; Based on risk assessment vector , and compare it with the threshold matrix Each level is compared by indicating the function Determine whether the risk assessment vector exceeds the thresholds at all levels, and sum them according to the weighted risk level to obtain the comprehensive warning level. , expressed as: ; in, is the indicator function, when the risk assessment vector More than Level Threshold Matrix The value is 1 when , otherwise it is 0.

[0048] Based on stress-strain vector and the coupled eigenvector , establish the objective function of rehabilitation parameter optimization, expressed as: ; ; in, represents the objective function of optimization, is the rehabilitation parameter vector, For the optimized rehabilitation parameters, is the target stress-strain state, For the optimal coupling state, and is the balance coefficient.

[0049] Based on the optimized rehabilitation parameters Generate personalized rehabilitation recommendations , expressed as: ; in, For the Items of rehabilitation recommendation parameters, .

[0050] Based on health assessment indicators and rehabilitation progress indicators, adaptively generate personalized rehabilitation plans and exercise recommendations, including: Set thresholds for health assessment indicators, including the first-level thresholds for health assessment and the second level threshold for health assessment , and set the expected rehabilitation progress benchmark as ; The threshold of rehabilitation progress risk is set based on the rehabilitation progress risk. , set the health status risk warning threshold to ; If the health assessment index S < the first level threshold , and rehabilitation progress indicators Higher than expected progress value , then the training parameters Adjust to the low load range, according to the coupling eigenvector Set up training action groups based on risk assessment vectors Determine the frequency of training; If the first level threshold ≤Health assessment index S≤Second level threshold ,and , is the preset tolerance value, then according to the optimization parameters Adjust the training intensity to a medium level, based on the coupled eigenvector Optimize training movement angles and vectors based on rehabilitation recommendations Set up training cycles; If the health assessment index S> the second level threshold , then immediately pause the current training plan and reduce the training parameters To the lowest load level, increase the monitoring frequency of the health assessment index S, and re-evaluate the stress-strain vector The various indicators are based on the risk assessment vector Adjust rehabilitation strategies; If the recovery progress risk >Threshold of risk of rehabilitation progress , and the health status risk <Health status risk warning threshold , then according to the stress-strain vector Develop a progressive training plan by coupling feature vectors Adjust the training action combination according to the risk warning level Set evaluation cycles; If the risk warning level When the preset risk warning threshold is reached, the objective function will be optimized Adjust the training parameters to a safe range, increase the monitoring frequency of the rehabilitation evaluation index S, and recalculate the rehabilitation recommendation vector The training parameters of .

[0051] In summary, the present invention, by establishing a biomechanical model in the orthopedic rehabilitation process, converts complex physiological states into quantifiable mathematical expressions, constructs a health status assessment system containing static and dynamic features, and achieves accurate assessment of the patient's rehabilitation status through feature fusion and dynamic update mechanisms; through the early warning mechanism of the risk assessment vector and the three-level threshold matrix, combined with the dual-objective optimization rehabilitation parameter adjustment method, it not only ensures the safety of the rehabilitation process, but also improves the pertinence of the rehabilitation effect. It not only overcomes the problems of strong experience dependence and low personalization in traditional rehabilitation nursing, but also makes the formulation and adjustment of rehabilitation plans explainable and traceable through the support of mathematical models, providing a new technical path for the standardization and intelligence of orthopedic rehabilitation nursing, and has significant clinical application value and promotion significance.

[0052] This embodiment further provides an intelligent analysis system for orthopedic care, including: The acquisition module is used to collect and pre-process medical imaging data, combine the patient's daily motion status data, build a multidimensional data set containing bone morphological characteristics and motion characteristics, use deep learning methods to extract key features, and establish a digital feature model of the patient's bone system; An evaluation module, based on the digital feature model, integrates and analyzes the joint force characteristics and tissue stress coupling degree, calculates the joint force distribution and stress-strain relationship, realizes quantitative evaluation of bone health status, and establishes an evaluation index system; The early warning module builds a multi-level risk warning system based on the evaluation indicator system to conduct real-time monitoring and early warning of abnormal conditions. At the same time, based on the rehabilitation progress evaluation indicators, it adaptively generates personalized rehabilitation plans and exercise recommendations.

[0053] This embodiment also provides a computer device, which is suitable for the intelligent analysis method for orthopedic care, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent analysis method for orthopedic care proposed in the above embodiment.

[0054] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0055] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the intelligent analysis method for orthopedic care proposed in the above embodiment is implemented.

[0056] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0057] Example 2: This example provides an intelligent analysis method for orthopedic care. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0058] Orthopedic patients who needed rehabilitation treatment were selected. The inclusion criteria were patients undergoing first surgery, and the exclusion criteria included patients with severe medical diseases, cognitive dysfunction, and previous relevant surgical history. The patients were randomly divided into an experimental group and a control group, with the same sample size in each group.

[0059] The experimental period covers the entire rehabilitation process, and the performance of the two groups is compared in three dimensions: rehabilitation effect, risk control, and degree of personalization.

[0060] The experimental results are shown in Table 1: Table 1 Comparison of experimental results Evaluation Metrics Experimental Group Control group Improvement rate Rehabilitation progress achievement rate 92.7% 78.3% 18.4% Risk warning accuracy 95.3% 82.1% 16.1% Timely adjustment of plans 96.8% 75.6% 28.0% Range of motion 93.2% 76.8% 21.4% Functional Independence Score 94.1% 79.2% 18.8% Early warning accuracy 95.3% 82.1% 16.1% From Table 1, it can be found that the rehabilitation progress achievement rate of the experimental group reached 92.7%, which was 18.4% higher than that of the control group. At the same time, the joint range of motion and functional independence scores were significantly improved, indicating that the present invention can effectively promote patient rehabilitation. In terms of risk control, the early warning accuracy of the experimental group reached 95.3%, and the timeliness of program adjustment was improved to 96.8%. The improvement of these two indicators fully demonstrates that the intelligent early warning mechanism of the present invention has high reliability and rapid response capabilities, and can detect potential risks in time and make adjustments, thereby ensuring the safety of the rehabilitation process.

[0061] Through the comprehensive analysis of these core indicators, it can be seen that the present invention achieves accurate assessment and timely intervention of the rehabilitation process by combining biomechanical modeling with deep learning, which has significant advantages in improving rehabilitation effects and ensuring rehabilitation safety.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent analysis method for orthopedic care, characterized in that: include: Collect and preprocess medical imaging data, combine it with the patient's daily motion status data, build a multidimensional data set containing bone morphological features and motion features, extract data features, use attention mechanism for adaptive fusion, and establish a digital feature model containing static morphological features and dynamic motion features; Based on the digital feature model, the joint force is analyzed through the orthogonal projection matrix, and the degree of tissue stress coupling is analyzed hierarchically to obtain the joint force characteristics and coupling characteristics, and to construct a comprehensive evaluation index system to achieve quantitative evaluation of bone health status; Based on the evaluation index system, a multi-level risk warning system is constructed to carry out real-time monitoring and warning of abnormal conditions. At the same time, based on the rehabilitation progress evaluation indicators, personalized rehabilitation plans and exercise recommendations are adaptively generated.

2. The intelligent analysis method for orthopedic care according to claim 1, characterized in that: The medical image data includes X-ray image data , CT image data and MRI data ; The X-ray image data Used to obtain two-dimensional projection features of bones; The CT image data Used to obtain cross-sectional features of bones; The magnetic resonance imaging data Used to obtain the characteristics of soft tissue around bones; The three types of image data are input into the preset medical image processing model for image enhancement and feature extraction.

3. The intelligent analysis method for orthopedic care according to claim 2, characterized in that: The medical image processing model constructs a medical image enhancement network with three branches for parallel processing. The first branch processes X-ray image data. , the second branch processes CT image data , the third branch processes MRI data ,Each branch includes image preprocessing, feature enhancement and feature fusion; The image preprocessing is to normalize the input medical image data, and use the adaptive histogram equalization method to improve the image contrast according to the normalized data, where the contrast limit parameter Dynamic adjustment is performed, expressed as: ； in, is the reference contrast limit parameter, is the standard deviation of the grayscale in the local area, is the reference standard deviation; Feature enhancement is performed by constructing residual attention, which is expressed as: ； in, is the intermediate feature map, represents the convolution operation, is the normalized input image, and is the weight of the fully connected layer, is the global average pooling, is the attention weight map, express Activation function, is the enhanced feature map; Feature extraction is performed through a multi-scale feature pyramid network, which includes 4 scale layers. The size of the feature map of each scale layer is halved, and the number of feature channels is doubled. The feature map calculation formula is expressed as: ； in, For the Layer feature pyramid features, is the corresponding backbone network feature, represents a 1×1 convolution operation, Indicates upsampling the feature map of the previous layer. is the multi-scale feature of the final output, It is a 3×3 convolution operation; The self-attention mechanism is used to fuse the features of the three branches, expressed as: ； ； ； in, , , Respectively represent the linear transformation , , matrix, is the input feature, , , is the weight matrix, is the query vector, is the key vector, is a value vector, is the feature dimension, is the fused feature, is the attention output of the X-ray branch, is the attention output of the CT branch, is the attention output of the NMR branch, It means concatenating the features of the three branches in the channel dimension; The fused features are restored to the original image resolution through the deconvolution network, and the residual connection is used to retain the detail information, which is expressed as: ； ； in, is the deconvolution operation, It is the enhanced feature map of the final output.

4. The intelligent analysis method for orthopedic care according to claim 3, characterized in that: Based on the medical image processing model, the image data is subjected to wavelet transformation enhancement and Canny edge detection, wherein the wavelet transformation can analyze the local features of the image at different scales and is suitable for denoising and enhancement of medical images, which is expressed as: ； in, represents the wavelet transform coefficients, is the scale parameter; is the translation parameter; is the wavelet basis function; The image signal is input; the bone contour feature data is obtained through wavelet transform and edge detection ; A wearable sensor array consisting of a three-axis acceleration sensor, a three-axis gyroscope, and an angle sensor is installed on the patient's joints to collect joint angle data during daily activities such as standing, walking, and going up and down stairs. , motion acceleration data and gait feature data ; The acquired motion data is normalized and Kalman filtered. The Z-score method is used to eliminate the dimension effect in the normalization process. Kalman filtering is used to suppress sensor noise and motion artifacts. Its state space equation is expressed as: ； ； in, for The system state vector at time , for The system state vector at time ; is the state transfer matrix; is the control matrix; is the control vector; is process noise, with mean 0 and covariance Gaussian distribution of is the observation vector; is the observation matrix; is the observation noise, with mean 0 and covariance Gaussian distribution; smooth motion state feature data is obtained by filtering ; The bone contour feature data Input into the preset deep convolutional neural network for feature extraction. The network contains 5 convolutional layers and 3 fully connected layers. The calculation expression of each convolutional layer is: ； in, For the Layer feature map; is the convolution kernel weight matrix; is the feature map of the previous layer; is the bias term; It is a modified linear unit activation function used to introduce nonlinear characteristics; through multi-layer convolution operations and pooling operations, the bone morphology feature vector is extracted ; The motion state characteristic data The input is sent to the preset long short-term memory network for time series feature analysis, and the motion state features are analyzed through the LSTM network. Perform time series modeling, and finally the network processes the time series motion features into motion feature vectors ; The bone morphology feature vector and the motion feature vector Adaptive fusion is performed through the attention mechanism, and the calculation formula of the attention mechanism is: ； in, is the attention weight vector, indicating the importance of different features; is the weight matrix; represents the concatenation of feature vectors; is the fused feature vector; , are the attention weight coefficients of the two features respectively; through the weighted fusion of the attention mechanism, a digital feature model including static morphological features and dynamic motion features is constructed.

5. The intelligent analysis method for orthopedic care according to claim 4, characterized in that: The joint force is obtained by converting the eigenvector into Mapped to the joint force space, the calculation formula is: ； in, for Dimensional joint force vector, indicating the force state of each key position, for dimensional orthogonal projection weight matrix, is the feature vector dimension, is the joint force dimension, for dimensional input feature vector, containing morphological and motion information, for dimensional bias vector, for dimensional identity matrix; Based on the joint force vector , combined with the hierarchical characteristics of the skeletal system, a multi-scale stress-strain analysis model is constructed, which is expressed as: ； ； ； in, For the The stress value of the layered structure, For the The stress transfer coefficient of the layered organization, For the The stress-bearing area of ​​the layer tissue, For the The strain value of the layer organization, For the The elastic modulus of the layered tissue, for dimensional tissue stress-strain vector, is the number of tissue layers; The degree of tissue stress coupling is used to evaluate the interaction between different tissue layers, and a hierarchical stress coupling analysis is performed, which is expressed as: ； ； in, For the Layer and Stress coupling coefficient between layers; and are the stress values ​​of adjacent layers, Indicates The stress value of the layer, for dimensional inter-layer coupling eigenvector.

6. The intelligent analysis method for orthopedic care according to claim 5, characterized in that: Acquiring stress-strain characteristics based on the multi-scale stress-strain analysis model; Obtaining coupling characteristics based on the stress coupling analysis; A comprehensive evaluation index system is established based on the stress-strain characteristics and coupling characteristics, which is expressed as: ； ； ； in, is an indicator of organizational status, is the coupling status indicator, For the Critical reference value of layer stress, For the Critical reference value of layer strain, For the Layer and +1The correlation coefficient between layers, For the Layer coupling critical reference value, , For the Layer stress-strain weight coefficient, For the Layer coupling weight coefficient, , is the comprehensive weight of the state index and coupling index, It is the final health assessment indicator; Further, analyze the health assessment indicators The changes in the time dimension can quantitatively evaluate and predict the patient's recovery status. The formula is: ； ； in, for The improvement of the state at a moment relative to the previous moment, For the The loss value at the moment, For the The loss value at the moment, For the Time to The decay weight at time, is the initial state value, is the current rehabilitation progress indicator, To evaluate the time point, for The amount of improvement in status at a given moment.

7. The intelligent analysis method for orthopedic care according to claim 6, characterized in that: The multi-level risk warning is based on the design of a three-level risk warning threshold matrix M based on health assessment indicators and rehabilitation progress indicators, which is expressed as: ； in, and Respectively Level health status and rehabilitation progress warning thresholds, Indicates the risk level; Based on health assessment indicators and rehabilitation progress indicators , calculate the risk assessment vector , get the health status risk and risk of rehabilitation progress , expressed as: ； ； ； in, is the health status risk, The risk of recovery progress is is the reference value for health status, is the expected rehabilitation progress value, and is the corresponding standard deviation parameter; Based on risk assessment vector , and compare it with the threshold matrix Compare each level to get a comprehensive warning level , expressed as: ； in, is the indicator function, when the risk assessment vector More than Level Threshold Matrix The value is 1 when it is, otherwise it is 0; Based on stress-strain vector and the coupled eigenvector , establish the objective function of rehabilitation parameter optimization, expressed as: ； ； in, represents the objective function of optimization, is the rehabilitation parameter vector, For the optimized rehabilitation parameters, is the target stress-strain state, For the optimal coupling state, and is the balance coefficient; Based on the optimized rehabilitation parameters Generate personalized rehabilitation recommendations , expressed as: ； in, For the Items of rehabilitation recommendation parameters, .

8. The intelligent analysis method for orthopedic care according to claim 7, characterized in that: Based on the health assessment indicators and rehabilitation progress indicators, a personalized rehabilitation plan and exercise recommendations are adaptively generated, including: Set thresholds for health assessment indicators, including the first-level thresholds for health assessment and the second level threshold for health assessment , and set the expected rehabilitation progress benchmark as ; The threshold of rehabilitation progress risk is set based on the rehabilitation progress risk. , set the health status risk warning threshold to ; If the health assessment index S < the first level threshold , and rehabilitation progress indicators Higher than expected progress value , then the training parameters Adjust to the low load range, according to the coupling eigenvector Set up training action groups based on risk assessment vectors Determine the frequency of training; If the first level threshold ≤Health assessment index S≤Second level threshold ,and , is the preset tolerance value, then according to the optimization parameters Adjust the training intensity to a medium level, based on the coupled eigenvector Optimize training movement angles and vectors based on rehabilitation recommendations Set up training cycles; If the health assessment index S> the second level threshold , then immediately pause the current training plan and reduce the training parameters To the lowest load level, increase the monitoring frequency of the health assessment index S, and re-evaluate the stress-strain vector The various indicators are based on the risk assessment vector Adjust rehabilitation strategies; If the recovery progress risk >Threshold of risk of rehabilitation progress , and the health status risk <Health status risk warning threshold , then according to the stress-strain vector Formulate a progressive training plan by coupling feature vectors Adjust the training action combination according to the risk warning level Set evaluation cycles; If the risk warning level When the preset risk warning threshold is reached, the objective function will be optimized Adjust the training parameters to a safe range, increase the monitoring frequency of the rehabilitation evaluation index S, and recalculate the rehabilitation recommendation vector The training parameters of .

9. An intelligent analysis system for orthopedic care, based on the intelligent analysis method for orthopedic care according to any one of claims 1 to 8, characterized in that: include: The acquisition module is used to collect and pre-process medical imaging data, combine the patient's daily motion status data, build a multidimensional data set containing bone morphological characteristics and motion characteristics, use deep learning methods to extract key features, and establish a digital feature model of the patient's bone system; An evaluation module, based on the digital feature model, integrates and analyzes the joint force characteristics and tissue stress coupling degree, calculates the joint force distribution and stress-strain relationship, realizes quantitative evaluation of bone health status, and establishes an evaluation index system; The early warning module builds a multi-level risk warning system based on the evaluation indicator system to conduct real-time monitoring and early warning of abnormal conditions. At the same time, based on the rehabilitation progress evaluation indicators, it adaptively generates personalized rehabilitation plans and exercise recommendations.

Images