Spine load monitoring system for dynamically adjusting rehabilitation plan after operation

Abstract

The invention discloses a spine load monitoring system for dynamically adjusting a rehabilitation plan after an operation, and relates to the technical field of spine load monitoring, a monitoring data acquisition module collects preoperative and postoperative multi-modal data, and provides a comprehensive basis for the rehabilitation plan; a rehabilitation training initial parameter setting module determines expected recovery effect data and initial parameters of each stage, so that rehabilitation training is more targeted; in the rehabilitation process, the recovery state evaluation module evaluates training and recovery effects in real time and discovers problems in time; the rehabilitation training initial parameter adjusting module adjusts parameters when the recovery effect does not reach the standard, it is ensured that the rehabilitation plan always conforms to the condition of the patient, and the problem that in the prior art, a personalized rehabilitation plan cannot be made accurately according to preoperative and postoperative multi-modal monitoring data of the patient is solved. The training effect is difficult to dynamically evaluate in the rehabilitation process, and rehabilitation training initial parameters are difficult to adjust in time to guarantee the effectiveness and safety of the rehabilitation process.

Description

Technical Field

[0001] The present invention relates to the technical field of spinal load monitoring, and specifically to a spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery. Background Art

[0002] Currently, the existing technologies for spinal load monitoring mainly include the following. Motion sensors and analysis systems use wearable accelerometers, gyroscopes, etc., combined with algorithms and analysis software, to estimate the spinal force by monitoring the body movement patterns and postural changes. Biomechanical models and simulations establish models based on individual body parameters and simulate and calculate the spinal pressure distribution and bearing capacity in combination with motion data. Pressure sensors can directly measure the pressure at the spinal part. Electromyogram monitoring indirectly infers the spinal force condition by measuring the electrical activity of the lumbar muscles. In addition, imaging technologies such as MRI and CT can also indirectly reflect the spinal pressure by observing the changes in the spinal morphological structure.

[0003] However, there are problems in the existing technologies that it is impossible to accurately formulate a personalized rehabilitation plan based on the multi-modal monitoring data before and after surgery of patients, and it is difficult to dynamically evaluate the training effect during the rehabilitation process and timely adjust the initial rehabilitation training parameters to ensure the effectiveness and safety of the rehabilitation process. Summary of the Invention

[0004] In view of the deficiencies of the existing technologies, the present invention provides a spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery, which solves the problems in the existing technologies that it is impossible to accurately formulate a personalized rehabilitation plan based on the multi-modal monitoring data before and after surgery of patients, and it is difficult to dynamically evaluate the training effect during the rehabilitation process and timely adjust the initial rehabilitation training parameters to ensure the effectiveness and safety of the rehabilitation process.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery, including a monitoring data acquisition module, an initial rehabilitation training parameter setting module, a recovery status evaluation module during the rehabilitation process, and an initial rehabilitation training parameter adjustment module, wherein: The monitoring data acquisition module is used to acquire multi-modal monitoring data, including multi-modal monitoring data before surgery and multi-modal monitoring data after surgery; The initial rehabilitation training parameter setting module is used to obtain the expected recovery effect data of each stage of the rehabilitation plan based on the multi-modal monitoring data and determine the initial rehabilitation training parameters of each stage; The recovery status evaluation module during the rehabilitation process is used to evaluate the training effect and recovery effect data of each stage of the patient based on the initial rehabilitation training parameters of each stage and the expected recovery effect data of each stage; The initial rehabilitation training parameter adjustment module is used to adjust the initial rehabilitation training parameters of a certain stage when it is detected that the recovery effect data of a certain stage of the patient does not meet the standard.

[0006] Further, the multimodal monitoring data includes patient basic information, spinal imaging data, physical function assessment data, vertebral body pressure distribution data, and electromyography data, where: the vertebral body pressure distribution data is obtained based on the array piezoresistive sensors on the wearable vest, and the array piezoresistive sensors cover the projection area of the T1-L5 vertebral bodies; the electromyography data is obtained based on the flexible sEMG electrodes attached to the wearable vest.

[0007] Further, the initial rehabilitation training parameter setting module includes an expected recovery effect data determination unit for each stage and an initial rehabilitation training parameter determination unit for each stage, where: the expected recovery effect data determination unit for each stage is used to compare and analyze the preoperative multimodal monitoring data pmm and the postoperative multimodal monitoring data pom to obtain the preoperative and postoperative comparison data ppc, including the changes in basic information before and after surgery, spinal recovery effect data, physical function improvement data, and biomechanical characteristic change data; the preoperative multimodal monitoring data pmm, the postoperative multimodal monitoring data pom, and the preoperative and postoperative comparison data ppc are used as the inputs of the trained random forest model, and the expected recovery effect data ero for each stage of the rehabilitation plan is output;

[0008] The initial rehabilitation training parameter determination unit for each stage is used to obtain the defined data sets stored in the database, including the preoperative multimodal monitoring defined data Spmm i , the postoperative multimodal monitoring defined data Spom i , the preoperative and postoperative comparison defined data Sppc i and the expected defined recovery effect Sero for each stage of the rehabilitation plan i ; the preoperative multimodal monitoring data pmm, the postoperative multimodal monitoring data pom, the preoperative and postoperative comparison data ppc, the expected recovery effect data ero for each stage of the rehabilitation plan, and the defined data sets are used as the inputs of the matching model to obtain the fitness coefficients corresponding to the defined data sets; the defined data set corresponding to the maximum fitness coefficient is determined, and the initial rehabilitation training parameters for each stage of the rehabilitation plan stored in the database corresponding to the defined data set are obtained from the database.

[0009] Further, the matching model is expressed as:

[0010]

[0011] where Pmod i is the fitness coefficient, f1, f2, and f3 are all transfer functions, α1, α2, and α3 are all weight factors, and σ(·) is the cosine similarity function.

[0012] Further, the multimodal monitoring data includes three-dimensional motion data and physiological data, where: the three-dimensional motion data is obtained based on the IMU sensors set on the wearable vest, including cervical vertebra IMU sensors, thoracic vertebra IMU sensors, and lumbar vertebra IMU sensors; the physiological data includes temperature Ws, humidity Hs, and heart rate Xs, and is obtained based on the temperature sensor, humidity sensor, and heart rate sensor set on the wearable vest;

[0013] During the rehabilitation process, the recovery status evaluation module includes a training process quality monitoring unit and a recovery effect data evaluation unit, where:

[0014] The training process quality monitoring unit is used to evaluate the training effect of the patient at each stage: obtain the physiological parameter determination data during the training process from the database based on the patient's basic information and the initial rehabilitation training parameters at the current rehabilitation stage, including the temperature determination range, humidity determination range, and heart rate determination range; judge whether each physiological parameter in the physiological data is within the corresponding determination range. If any one of the physiological parameters is not within the determination range, the alarm is triggered through the alarm and the training is stopped; if all are within the determination range, the physiological state is determined based on the comprehensive physiological evaluation coefficient ZPx; if the comprehensive physiological evaluation coefficient is greater than the evaluation threshold, the physiological state is abnormal, the alarm is triggered through the alarm and the training is stopped; if the comprehensive physiological evaluation coefficient is not greater than the evaluation threshold, the training action compliance XdF is obtained based on the cone pressure distribution data vpd, electromyogram data emg, and three-dimensional motion data dmd during the training process, which is used to represent the training effect of the patient at each stage; if the training action compliance is greater than the set action compliance threshold, it means the training action is qualified; if the training action compliance is not greater than the set action compliance threshold, it means the training action is unqualified;

[0015] The expected recovery effect data evaluation unit is used to obtain the spinal imaging data, body function evaluation data Jz2 after the end of the training phase, obtain the spinal feature parameters Jz1 after training based on the spinal imaging data, and is also used to obtain the cone pressure distribution data Jz3, electromyogram data Jz4 and three-dimensional motion data Jz5 when the patient makes an evaluation action; obtain the expected recovery effect data after the end of the training phase of this training stage, including expected spinal feature parameters YJz1, expected body function evaluation data YJz2, expected cone pressure distribution data YJz3, expected electromyogram data YJz4 and expected three-dimensional motion data YJz5; based on the expected effect evaluation model, obtain the expected effect evaluation coefficient; if the expected effect evaluation coefficient is greater than the set effect evaluation threshold, it is considered that the training effect of the current stage reaches the standard and the next stage of rehabilitation training can be carried out; if the expected effect evaluation coefficient is not greater than the set effect evaluation threshold, obtain the compliance of each training action during the training process of this stage, record the compliance of each training action as the set label, compare the set label with each set label stored in the database, determine the set label that is consistent with the set label, and obtain the effect correction factor corresponding to the set label stored in the database; obtain the product of the expected effect evaluation coefficient and the effect correction factor, and if the product is greater than the set product threshold, it is considered that the initial parameters of the rehabilitation training in the current stage are in line and the rehabilitation training in this stage continues; if the product is not greater than the set product threshold, it is considered that the training effect of the current stage does not reach the standard.

[0016] Furthermore, the method for obtaining the comprehensive physiological evaluation coefficient ZPx is as follows:

[0017]

[0018] Among them, cW is the temperature intermediate value, which is the mean value of the maximum and minimum values of the temperature reference range, cH is the humidity intermediate value, which is the mean value of the maximum and minimum values of the humidity reference range, and cX is the heart rate intermediate value, which is the mean value of the maximum and minimum values of the heart rate reference range;

[0019] The method for obtaining the training action compliance XdF is as follows:

[0020] XdF = σ(vpd, Svpd) + σ(emg, Semg) + σ(dmd, Sdmd);

[0021] Among them, Svpd is the initial cone pressure distribution data in the initial parameters of the stage rehabilitation training in the current stage, Semg is the initial electromyogram data in the initial parameters of the stage rehabilitation training in the current stage, Sdmd is the initial three-dimensional motion data in the initial parameters of the stage rehabilitation training in the current stage, and σ(·) is the cosine similarity function.

[0022] Furthermore, the method for obtaining the expected effect evaluation coefficient is as follows:

[0023]

[0024] Among them, RS a is the recovery effect score, DS a is the deviation from the expected score, CRI is the expected effect evaluation coefficient, γ is the weight coefficient, F(Jz a , YJz a ) is specifically: the absolute value of the difference between each parameter in Jz a and each parameter in YJz a divided by the normal range of the parameter, and after summing up each ratio, the average value is processed. min(||Jz a ||, ||YJz a ||) is the minimum value of the norms of Jz a and YJz a , and max(||Jz a ||, ||YJz a ||) is the maximum value of the norms of Jz a and YJz a .

[0025] Furthermore, adjust the initial parameters of the rehabilitation training, including the following steps: construct the objective function:

[0026]

[0027] Among them, Mb is the objective function value, is the complex function of the change between the adjusted training parameters and the initial parameters of the rehabilitation training, is the rehabilitation training cost function of the adjusted training parameters, the expected effect evaluation function of the adjusted training parameters, C tol is the evaluation value of the patient's physical tolerance, is the adjusted training parameter, λ1 is 's weight factor, λ2 is 's weight factor, λ3 is 's weight factor, λ4 is C tol 's weight factor; under the constraint conditions, starting from the initial parameters of the rehabilitation training at the current stage, adjust the initial parameters of the rehabilitation training through the gradient descent method until the objective function converges to the minimum value.

[0028] Furthermore, is the square of the Euclidean distance between the vector representation of the initial parameters of the rehabilitation training and the vector representation of the adjusted training parameters; Obtained based on the trained linear regression model: taking the adjusted training parameter as the independent variable of the trained linear regression model, and recording the function value of the trained linear regression model as Based on the trained machine learning model: The input of the trained machine learning model is the adjusted training parameters, and the output is the predicted rehabilitation effect data. The predicted rehabilitation effect data is compared and analyzed with the expected recovery effect data at the current stage to obtain the function value.

[0029] Furthermore, the patient's physical tolerance evaluation value C tol is obtained as follows: Obtain the physical function evaluation data of the patient at the current training stage, including the ODI score P ODI and the VAS pain intensity score P VAS ; Based on the physiological data during the current training stage, obtain the patient's physiological characteristic change rate data, including the maximum temperature change rate Wyb, the maximum humidity change rate Hyb, and the maximum heart rate change rate Xyb; Input the physical function evaluation data and the patient's physiological characteristic change rate data into the physical condition evaluation model to obtain the patient's physical tolerance evaluation value:

[0030] where, δ1 is the first weight factor, and δ2 is the second weight factor.

[0031] The present invention has the following beneficial effects:

[0032] The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery collects preoperative and postoperative multimodal data through the monitoring data acquisition module, providing a comprehensive basis for the rehabilitation plan; the initial parameter setting module for rehabilitation training determines the expected recovery effect data and initial parameters for each stage accordingly, making the rehabilitation training more targeted; the recovery status evaluation module during the rehabilitation process evaluates the training and recovery effects in real time, and discovers problems in a timely manner; the initial parameter adjustment module for rehabilitation training adjusts the parameters when the recovery effect does not meet the standard, ensuring that the rehabilitation plan always fits the patient's condition, solving the problems in the prior art that it is impossible to accurately formulate a personalized rehabilitation plan based on preoperative and postoperative multimodal monitoring data of patients, and it is difficult to dynamically evaluate the training effect during the rehabilitation process and timely adjust the initial parameters of the rehabilitation training to ensure the effectiveness and safety of the rehabilitation process, and realizing the dynamic and accurate adjustment of the rehabilitation plan for spinal surgery patients.

[0033] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the flowchart of the spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery of the present invention.

[0035] Figure 2 is the schematic diagram of the position of the arrayed piezoresistive sensors of the spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery of the present invention.

[0036] In the figure, 1 is an array piezoresistive sensor; 2 is a flexible sEMG electrode; 3 is a cervical vertebra IMU sensor; 4 is a thoracic vertebra IMU sensor; 5 is a lumbar vertebra IMU sensor; 6 is a temperature sensor; 7 is a humidity sensor; 8 is a heart rate sensor. Detailed implementation manners

[0037] Please refer to Figure 1 , the embodiments of the present invention provide a technical solution: a spinal load monitoring system for dynamically adjusting a rehabilitation plan after surgery, including the following steps: including a monitoring data acquisition module, an initial parameter setting module for rehabilitation training, a recovery status evaluation module during rehabilitation, and an initial parameter adjustment module for rehabilitation training, where:

[0038] The monitoring data acquisition module is used to acquire multimodal monitoring data, including preoperative multimodal monitoring data and postoperative multimodal monitoring data. According to needs, corresponding parameters are subjected to dimensionality reduction processing or normalization processing before calculation.

[0039] The multimodal monitoring data includes patient basic information (including but not limited to height, gender, age), spinal imaging data, body function evaluation data (including but not limited to ODI score P ODI and VAS pain intensity score P VAS ), vertebral body pressure distribution data, and electromyography data, where: the vertebral body pressure distribution data is obtained based on the array piezoresistive sensor 1 on the wearable vest, and the array piezoresistive sensor 1 covers the projection area of the T1-L5 vertebral bodies; the electromyography data is obtained based on the flexible sEMG electrode 2 attached to the wearable vest, as Figure 2 shown.

[0040] The initial parameter setting module for rehabilitation training is used to obtain the expected recovery effect data for each stage of the rehabilitation plan based on the multimodal monitoring data, and determine the initial parameters for rehabilitation training in each stage.

[0041] The initial parameter setting module for rehabilitation training includes an expected recovery effect data determination unit for each stage and an initial parameter determination unit for rehabilitation training in each stage, where:

[0042] The unit for determining the expected recovery effect data at each stage is used to compare and analyze the preoperative multimodal monitoring data pmm and the postoperative multimodal monitoring data pom to obtain the preoperative and postoperative comparison data ppc, including the changes in basic information before and after surgery, spinal recovery effect data, physical function improvement, and biomechanical characteristic change data (including changes in cone pressure distribution data and electromyographic data); comparing and analyzing the preoperative multimodal monitoring data and the postoperative multimodal monitoring data can fully understand the changes in the patient's physical condition, including changes in basic information, spinal recovery effect data, physical function improvement, and biomechanical characteristic change data. These comparative data can accurately reflect the impact of the surgery on the patient's body and provide a key basis for the subsequent formulation of a rehabilitation plan that meets the patient's actual recovery needs.

[0043] The preoperative multimodal monitoring data pmm, the postoperative multimodal monitoring data pom, and the preoperative and postoperative comparison data ppc are used as the input of the trained random forest model, and the expected recovery effect data ero of each stage of the rehabilitation plan is output. The preoperative and postoperative multimodal monitoring data and the comparison data are used as the input of the trained random forest model, and the expected recovery effect data of each stage of the rehabilitation plan is output. The random forest model has strong learning and prediction capabilities, and can comprehensively analyze complex multimodal data and explore the potential laws and relationships therein. The expected recovery effect data of each stage predicted based on these data can set scientific and reasonable goals for rehabilitation training, avoid the blindness of rehabilitation plans, and ensure that rehabilitation training is carried out in the direction of promoting the patient's physical recovery.

[0044] The initial parameter determination unit for each stage of rehabilitation training is used to obtain each defined data set stored in the database, including the preoperative multimodal monitoring defined data Spmm i , postoperative multimodal monitoring defines data Spom i , preoperative and postoperative comparison definition data Sppc i and rehabilitation program stages to define the expected recovery effect Sero i ; Use the preoperative multimodal monitoring data pmm, postoperative multimodal monitoring data pom, preoperative and postoperative comparison data ppc, expected recovery effect data ero at each stage of the rehabilitation plan and each defined data set as the input of the matching model to obtain the fitness coefficient corresponding to each defined data set; obtain each defined data set stored in the database, including preoperative multimodal monitoring definition data, postoperative multimodal monitoring definition data, preoperative and postoperative comparison definition data and expected defined recovery effect at each stage of the rehabilitation plan. These defined data sets are standard data summarized based on a large amount of clinical experience and research, and provide a reference framework for determining the initial parameters of rehabilitation training. By comparing and matching with the actual data of the patient, the initial parameters of rehabilitation training that best suit the individual patient's situation can be found.

[0045] Determine the defined data set corresponding to the maximum fitness coefficient, and obtain the initial parameters of the rehabilitation training at each stage of the rehabilitation plan stored in the database corresponding to this defined data set from the database.

[0046] The matching model is expressed as:

[0047]

[0048] Among them, Pmod i is the fitness coefficient, f1, f2, and f3 are all transfer functions, α1, α2, and α3 are all weight factors, and σ(·) is the cosine similarity function. When determining the initial parameters of the rehabilitation training, this coefficient can be used to accurately measure the degree of fit between the current patient data and different defined data sets in the database. For example, when selecting the initial parameters of the rehabilitation training for a certain patient, the model calculates the fitness coefficients of each defined data set. The larger the value, the higher the degree of match between the defined data set and the actual situation of the patient. Then, the most suitable initial parameters of the rehabilitation training are found to achieve the precise customization of the rehabilitation plan and solve the problem that the rehabilitation plan is difficult to fit individual differences.

[0049] Take the preoperative multimodal monitoring data, postoperative multimodal monitoring data, preoperative and postoperative comparison data, the expected recovery effect data at each stage of the rehabilitation plan, and each defined data set as the input of the matching model to obtain the fitness coefficients corresponding to each defined data set. Determine the defined data set corresponding to the maximum fitness coefficient, and obtain the initial parameters of the rehabilitation training at each stage of the rehabilitation plan stored in the database corresponding to this defined data set from the database. This method uses the matching model, fully considering the individual differences of patients and various factors affecting rehabilitation, making the determined initial parameters of the rehabilitation training more in line with the actual situation of patients, improving the pertinence and effectiveness of the rehabilitation plan, and then solving the problems of lack of personalization in the rehabilitation plan and difficulty in precisely meeting the rehabilitation needs of patients.

[0050] During the rehabilitation process, the recovery status evaluation module is used to evaluate the training effect and recovery effect data of each stage of the patient based on the initial parameters of the rehabilitation training at each stage and the expected recovery effect data at each stage.

[0051] The multimodal monitoring data also includes three-dimensional motion data and physiological data, where: the three-dimensional motion data is obtained based on the IMU sensors set on the wearable vest, including the cervical vertebra IMU sensor 3, the thoracic vertebra IMU sensor 4, and the lumbar vertebra IMU sensor 5; the physiological data includes temperature Ws, humidity Hs, and heart rate Xs, which are obtained based on the temperature sensor 6, humidity sensor 7, and heart rate sensor 8 set on the wearable vest.

[0052] During the rehabilitation process, the recovery status evaluation module includes a training process quality monitoring unit and a recovery effect data evaluation unit, where:

[0053] The training process quality monitoring unit is used to evaluate the training effects at each stage of the patient: obtain the physiological parameter determination data during the training process from the database based on the patient's basic information and the initial rehabilitation training parameters at the current rehabilitation stage, including the temperature determination range, humidity determination range, and heart rate determination range (combine the patient's basic information and the initial rehabilitation training parameters at the current rehabilitation stage into a matching label, compare the matching label with each pointing label stored in the database one by one, determine the pointing label identical to the matching label, and obtain the physiological parameter determination data during the training process stored in the database corresponding to the pointing label); judge whether each physiological parameter in the physiological data is within the corresponding determination range. If any one physiological parameter is not within the determination range, alarm through the alarm and stop the training; if all are within the determination range, determine the physiological state based on the comprehensive physiological evaluation coefficient ZPx; if the comprehensive physiological evaluation coefficient is greater than the evaluation threshold, the physiological state is abnormal, alarm through the alarm and stop the training.

[0054] Incorporate three-dimensional motion data and physiological data into the multi-modal monitoring category. Obtain the three-dimensional motion data of the cervical vertebra, thoracic vertebra, and lumbar vertebra based on the IMU sensors on the wearable vest, and be able to grasp in real time information such as the motion trajectory and posture changes of the spine in space. For example, during rehabilitation training, it can accurately judge whether the patient's bending and twisting movements are standard, and avoid aggravating the burden on the spine due to improper movements. At the same time, use the physiological data obtained by the temperature sensor, humidity sensor, and heart rate sensor to reflect the internal state of the patient's body during training. When the heart rate is too high, or the body temperature or humidity changes abnormally, it indicates that the training intensity is too high or the patient's body is uncomfortable, providing a basis for timely adjustment of the training plan.

[0055] Obtain the physiological parameter determination data during the training process from the database according to the patient's basic information and the initial rehabilitation training parameters, and monitor the physiological parameters of the patient during training in real time. Once any physiological parameter is found to exceed the determination range, immediately alarm and stop the training to ensure the safety of the patient. If the physiological parameters are within the range, further determine the physiological state through the comprehensive physiological evaluation coefficient. The calculation of the comprehensive physiological evaluation coefficient takes into account the intermediate values of temperature, humidity, and heart rate, and can more comprehensively reflect the physiological state of the patient. If this coefficient is greater than the evaluation threshold, also alarm and stop the training to prevent potential risks. For the situation where the physiological state is normal, evaluate the training effect through the compliance of the training actions, which is calculated based on the cosine similarity between the cone pressure distribution data, electromyography data, and three-dimensional motion data and the initial set data, and judge whether the patient's training actions are qualified, and correct the wrong actions in time to ensure the training quality.

[0056] If the comprehensive physiological evaluation coefficient is not greater than the evaluation threshold, the training movement compliance XdF is obtained based on the cone pressure distribution data vpd, electromyography data emg and three-dimensional motion data dmd during the training process, which is used to indicate the training effect of the patient at each stage; if the training movement compliance is greater than the set movement compliance threshold, it means that the training movement is qualified; if the training movement compliance is not greater than the set movement compliance threshold, it means that the training movement is unqualified.

[0057] The method for obtaining the comprehensive physiological assessment coefficient ZPx is as follows:

[0058]

[0059] Among them, cW is the middle value of temperature, the average of the maximum and minimum values of the temperature parameter range, cH is the middle value of humidity, the average of the maximum and minimum values of the humidity parameter range, and cX is the middle value of heart rate, the average of the maximum and minimum values of the heart rate parameter range. The changes in the three key physiological parameters of temperature, humidity and heart rate during training are fully integrated. During rehabilitation training, this coefficient can accurately quantify the patient's physiological state. If the coefficient is greater than the evaluation threshold, it indicates that the patient's physiological state is abnormal, which may be due to excessive training intensity that the body cannot adapt. At this time, the system alarms and stops training to prevent patients from being injured due to overtraining, ensuring the safety of patients during rehabilitation training.

[0060] The method for obtaining the training action compliance XdF is as follows:

[0061] XdF=σ(vpd,Svpd)+σ(emg,Semg)+σ(dmd,Sdmd);

[0062] Among them, Svpd is the initial cone pressure distribution data in the initial parameters of the current stage of stage rehabilitation training, Semg is the initial electromyographic data in the initial parameters of the current stage of stage rehabilitation training, Sdmd is the initial three-dimensional motion data in the initial parameters of the current stage of stage rehabilitation training, and σ(·) is the cosine similarity function. Comprehensively evaluate the consistency of training movements with expected movements from multiple dimensions (cone pressure, muscle electrical activity, and movement state). When the training movement compliance is greater than the set movement compliance threshold, it means that the training movement is qualified, and the system can promptly detect the patient's incorrect movements during training, and prompt the patient and rehabilitation personnel to correct them. This helps to improve the training effect, avoid obstruction of rehabilitation progress or potential damage due to incorrect movements, and ensure that rehabilitation training can be carried out according to the predetermined plan and goals, thereby better promoting patient recovery.

[0063] The expected recovery effect data evaluation unit is used to obtain the spinal image data, body function evaluation data Jz2 after the end of the training phase, obtain the spinal feature parameters Jz1 after training based on the spinal image data, and is also used to obtain the vertebral body pressure distribution data Jz3, electromyogram data Jz4 and three-dimensional motion data Jz5 when the patient makes an evaluation action; obtain the expected recovery effect data after the end of this training phase, including expected spinal feature parameters YJz1, expected body function evaluation data YJz2, expected vertebral body pressure distribution data YJz3, expected electromyogram data YJz4 and expected three-dimensional motion data YJz5;

[0064] Based on the expected effect evaluation model, obtain the expected effect evaluation coefficient; if the expected effect evaluation coefficient is greater than the set effect evaluation threshold, it is considered that the training effect of the current phase meets the standard and the next phase of rehabilitation training can be carried out; if the expected effect evaluation coefficient is not greater than the set effect evaluation threshold, obtain the compliance of each training action during the training process of this phase, record the compliance of each training action as the set label, compare the set label with the set labels stored in the database, determine the set label that is consistent with the set label, and obtain the effect correction factor corresponding to the set label stored in the database; obtain the product of the expected effect evaluation coefficient and the effect correction factor, if the product is greater than the set product threshold, it is considered that the initial parameters of the rehabilitation training in the current phase are in line and continue the rehabilitation training in this phase; if the product is not greater than the set product threshold, it is considered that the training effect of the current phase does not meet the standard.

[0065] After the end of the training phase, this unit obtains multi-faceted information such as spinal image data and body function evaluation data, combines the spinal feature parameters after training and the expected recovery effect data, and uses the expected effect evaluation model to calculate the expected effect evaluation coefficient. If the coefficient is greater than the set threshold, it indicates that the training effect meets the standard and the next phase of rehabilitation training can be entered; if it does not meet the standard, obtain the compliance of the training actions during the training process, compare with the set labels in the database, and find the corresponding effect correction factor. By multiplying the expected effect evaluation coefficient by the effect correction factor, it is judged whether the initial parameters of the rehabilitation training are in line. If the product is greater than the set product threshold, continue the training in the current phase; otherwise, it is determined that the training effect does not meet the standard, triggering the process of adjusting the initial parameters of the rehabilitation training to ensure that the rehabilitation plan always fits the patient's recovery progress and physical condition.

[0066] The method for obtaining the expected effect evaluation coefficient is as follows:

[0067]

[0068] Among them, RS a is the recovery effect score, DS a is the deviation from the expected score, CRI is the expected effect evaluation coefficient, γ is the weight coefficient, F(Jz a ,YJza )Specifically: Jz a The absolute value of the difference between each parameter in and the corresponding parameter in YJz a is divided by the normal range of the parameter, and the sums of these ratios are averaged. min(||Jz a ||,||YJz a ||) is the minimum value of the norms of Jz a and YJz a , and max(||Jz a ||,||YJz a ||) is the maximum value of the norms of Jz a and YJz a . The recovery effect score reflects the actual improvement of the patient in terms of spinal characteristics, physical function, etc. after a stage of rehabilitation training; the deviation from the expected score measures the gap between the actual recovery and the expected recovery effect. Through such a calculation method, the effect of rehabilitation training can be comprehensively and accurately evaluated. For example, if after a certain stage of training, the scoliosis angle of the patient improves (the recovery effect score increases), and the gap with the expected recovery effect is small (the deviation from the expected score is low), then the expected effect evaluation coefficient will be high, intuitively reflecting that the rehabilitation training effect at this stage is good

[0069] The initial parameter adjustment module for rehabilitation training is used to adjust the initial parameters of the rehabilitation training at a certain stage when it is monitored that the recovery effect data of the patient does not meet the standard.

[0070] Adjusting the initial parameters of the rehabilitation training includes the following steps:

[0071] Construct the objective function:

[0072]

[0073] where Mb is the value of the objective function, is the complex function of the change between the adjusted training parameters and the initial parameters of the rehabilitation training, is the rehabilitation training cost function of the adjusted training parameters, is the expected effect evaluation function of the adjusted training parameters, C tol is the evaluation value of the patient's physical tolerance, is the adjusted training parameter, λ1 is the weight factor, λ2 is the weight factor, λ3 is the weight factor, and λ4 is the tol weight factor of C; under the constraint conditions (ensuring that each parameter is adjusted within the set range), starting from the initial parameters of the current stage of rehabilitation training, the initial parameters of the rehabilitation training are adjusted by the gradient descent method until the objective function converges to the minimum value.

[0074] It is the square of the Euclidean distance between the vector representation of the initial parameters for rehabilitation training and the vector representation of the adjusted training parameters; A smaller value means that the adjusted parameters are not very different from the initial parameters, the adjustment process is relatively simple, and the overall impact on the rehabilitation plan is small; conversely, if this value is large, it indicates a large adjustment range of the parameters, which may bring more uncertainties and risks. By clarifying this functional form, when constructing the objective function, the adjustment range and degree of rehabilitation training parameters can be more precisely controlled, avoiding adverse effects on patients caused by excessive adjustment, and ensuring the stability and safety of the rehabilitation plan adjustment.

[0075] Obtained based on the trained linear regression model: Taking the adjusted training parameters as the independent variable of the trained linear regression model, and denoting the function value output by the trained linear regression model as Obtained based on the trained machine learning model: The input of the trained machine learning model is the adjusted training parameters, and the output is the predicted rehabilitation effect data. Comparing and analyzing the predicted rehabilitation effect data with the expected recovery effect data at the current stage to obtain The function value of (the calculation method is the same as that of the expected effect evaluation coefficient).

[0076] Patient's physical tolerance evaluation value C tol The acquisition method is as follows: Obtain the physical function evaluation data of the patient at the current training stage, including the ODI score P ODI and the VAS pain intensity score P VAS ; Based on the physiological data during the current training stage, obtain the patient's physiological characteristic variation rate data, including the maximum temperature variation rate Wyb, the maximum humidity variation rate Hyb, and the maximum heart rate variation rate Xyb; Input the physical function evaluation data and the patient's physiological characteristic variation rate data into the physical condition evaluation model to obtain the patient's physical tolerance evaluation value:

[0077] where δ1 is the first weight factor and δ2 is the second weight factor. By comprehensively considering the physical function evaluation data (including the ODI score and the VAS pain intensity score) and the physiological characteristic variation rate data (the maximum temperature variation rate, the maximum humidity variation rate, and the maximum heart rate variation rate) of the patient at the current training stage, it comprehensively reflects the patient's physical condition during the rehabilitation training process. The ODI score can reflect the degree of influence of low back pain on the patient's daily living activities, and the VAS pain intensity score directly reflects the patient's pain perception. The two show the patient's physical function state from the aspects of subjective feelings and functional limitations. The physiological characteristic variation rate data, on the other hand, reflects the patient's body's stress response during training from the perspective of objective physiological index changes.

[0078] An electronic device, comprising: a processor; and a memory in which computer program instructions are stored, and when the computer program instructions are run by the processor, the processor is caused to execute the spinal load monitoring system for post-operative dynamic adjustment of a rehabilitation plan as described above.

[0079] A computer-readable storage medium for storing a program, which when executed by a processor implements the spinal load monitoring system for post-operative dynamic adjustment of a rehabilitation plan as described above.

[0080] Those skilled in the art should understand that the embodiments of the present invention may be provided as a method, a system, or a computer program product. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] The present invention is described with reference to the flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or a plurality of flows and / or blocks

[0082] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the flows Figure 1 or a plurality of flows and / or blocks

[0083] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are performed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the functions specified in Figure 1 one or more of the flows Figure 1Steps of the functions specified in one or more boxes.

[0084] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A spinal load monitoring system for dynamically adjusting a rehabilitation plan after surgery, characterized in that, It includes a monitoring data acquisition module, an initial rehabilitation training parameter setting module, a recovery status evaluation module during the rehabilitation process, and an initial rehabilitation training parameter adjustment module, where: The monitoring data acquisition module is used to acquire multi-modal monitoring data, including pre-operative multi-modal monitoring data and post-operative multi-modal monitoring data; The initial rehabilitation training parameter setting module is used to obtain the expected recovery effect data for each stage of the rehabilitation plan based on the multi-modal monitoring data, and determine the initial rehabilitation training parameters for each stage; The recovery status evaluation module during the rehabilitation process is used to evaluate the training effect and recovery effect data of the patient at each stage based on the initial rehabilitation training parameters for each stage and the expected recovery effect data for each stage; The initial rehabilitation training parameter adjustment module is used to adjust the initial rehabilitation training parameters for a certain stage when it is detected that the recovery effect data of the patient in a certain stage does not meet the standard.

2. The spinal load monitoring system for dynamically adjusting a rehabilitation plan after surgery according to claim 1, wherein The multi-modal monitoring data includes patient basic information, spinal imaging data, body function evaluation data, vertebral body pressure distribution data, and electromyography data, where: The vertebral body pressure distribution data is obtained based on the array piezoresistive sensor (1) on the wearable vest, and the array piezoresistive sensor (1) covers the projection area of the T1-L5 vertebral bodies; The electromyography data is obtained based on the flexible sEMG electrodes (2) attached to the wearable vest.

3. The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery according to claim 2, wherein, The initial rehabilitation training parameter setting module includes an expected recovery effect data determination unit for each stage and an initial rehabilitation training parameter determination unit for each stage, where: The expected recovery effect data determination unit for each stage is used to compare and analyze the pre-operative multi-modal monitoring data pmm and the post-operative multi-modal monitoring data pom to obtain the pre- and post-operative comparison data ppc, including the changes in basic information before and after surgery, spinal recovery effect data, body function improvement data, and biomechanical characteristic change data; Taking the pre-operative multi-modal monitoring data pmm, the post-operative multi-modal monitoring data pom, and the pre- and post-operative comparison data ppc as the input of the trained random forest model, and outputting the expected recovery effect data ero for each stage of the rehabilitation plan; The initial parameter determination unit for each stage of rehabilitation training is used to obtain the defined data sets stored in the database, including the multi-modal monitoring defined data Spmm before surgery i , the multi-modal monitoring defined data Spom after surgery i , the comparison defined data Sppc before and after surgery i and the expected defined recovery effects Sero for each stage of the rehabilitation plan i ; Taking the pre-operative multi-modal monitoring data pmm, the post-operative multi-modal monitoring data pom, the pre- and post-operative comparison data ppc, the expected recovery effect data ero for each stage of the rehabilitation plan, and each defined data set as the input of the matching model to obtain the fitness coefficient corresponding to each defined data set; Determine the defined data set corresponding to the maximum fitness coefficient, and obtain the initial rehabilitation training parameters for each stage of the rehabilitation plan stored in the database corresponding to this defined data set from the database.

4. The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery according to claim 3, wherein, The matching model is expressed as: Among them, Pmod i is the fitness coefficient, f1, f2, and f3 are all transfer functions, α1, α2, and α3 are all weight factors, and σ(·) is the cosine similarity function.

5. The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery according to claim 2, wherein The multi-modal monitoring data also includes three-dimensional motion data and physiological data, where: The three-dimensional motion data is obtained based on the IMU sensors set on the wearable vest, including a cervical vertebra IMU sensor (3), a thoracic vertebra IMU sensor (4), and a lumbar vertebra IMU sensor (5); The physiological data includes temperature Ws, humidity Hs, and heart rate Xs, and is obtained based on the temperature sensor (6), humidity sensor (7), and heart rate sensor (8) set on the wearable vest; The recovery status evaluation module during the rehabilitation process includes a training process quality monitoring unit and a recovery effect data evaluation unit, where: The training process quality monitoring unit is used to evaluate the training effects of patients at each stage: Obtain the physiological parameter determination data during training from the database based on the patient's basic information and the initial rehabilitation training parameters at the current rehabilitation stage, including the temperature determination range, humidity determination range, and heart rate determination range; Judge whether each physiological parameter in the physiological data is within the corresponding determination range. If any physiological parameter is not within the determination range, alarm through the alarm and stop the training; If all are within the determination range, determine the physiological state based on the comprehensive physiological evaluation coefficient ZPx; If the comprehensive physiological evaluation coefficient is greater than the evaluation threshold, the physiological state is abnormal, alarm through the alarm and stop the training; If the comprehensive physiological evaluation coefficient is not greater than the evaluation threshold, obtain the training action compliance XdF based on the cone pressure distribution data vpd, electromyogram data emg, and three-dimensional motion data dmd during the training process, which is used to represent the training effects of patients at each stage; If the training action compliance is greater than the set action compliance threshold, it means that the training action is qualified; If the training action compliance is not greater than the set action compliance threshold, it means that the training action is unqualified; The expected recovery effect data evaluation unit is used to obtain the spinal image data and body function evaluation data Jz2 after the end of the training stage, obtain the spinal characteristic parameters Jz1 after training based on the spinal image data, and is also used to obtain the cone pressure distribution data Jz3, electromyogram data Jz4, and three-dimensional motion data Jz5 when the patient makes the evaluation action; Obtain the expected recovery effect data after the end of this training stage, including the expected spinal characteristic parameters YJz1, expected body function evaluation data YJz2, expected cone pressure distribution data YJz3, expected electromyogram data YJz4, and expected three-dimensional motion data YJz5; Based on the expected effect evaluation model, obtain the expected effect evaluation coefficient; If the expected effect evaluation coefficient is greater than the set effect evaluation threshold, it is considered that the training effect of the current stage reaches the standard and the next stage of rehabilitation training can be carried out; If the expected effect evaluation coefficient is not greater than the set effect evaluation threshold, obtain the training action compliance of each time during the training process of this stage, record the training action compliance of each time as the set label, compare the set label with each set label stored in the database, determine the set label that is consistent with the set label, and obtain the effect correction factor corresponding to the set label stored in the database; Obtain the product of the expected effect evaluation coefficient and the effect correction factor. If the product is greater than the set product threshold, it is considered that the initial rehabilitation training parameters of the current stage are compliant and continue the rehabilitation training of this stage; If the product is not greater than the set product threshold, it is considered that the training effect of the current stage does not reach the standard.

6. The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery according to claim 5, characterized in that The method for obtaining the comprehensive physiological evaluation coefficient ZPx is as follows: Among them, cW is the temperature median, the mean of the maximum and minimum values of the temperature determination range, cH is the humidity median, the mean of the maximum and minimum values of the humidity determination range, and cX is the heart rate median, the mean of the maximum and minimum values of the heart rate determination range; The method for obtaining the training action compliance XdF is as follows: XdF = σ(vpd, Svpd) + σ(emg, Semg) + σ(dmd, Sdmd); Among them, Svpd is the initial cone pressure distribution data in the initial parameters of the stage rehabilitation training in the current stage, Semg is the initial electromyogram data in the initial parameters of the stage rehabilitation training in the current stage, Sdmd is the initial three-dimensional motion data in the initial parameters of the stage rehabilitation training in the current stage, and σ(·) is the cosine similarity function.

7. The spinal load monitoring system for post-operative dynamic adjustment of rehabilitation plan according to claim 5, wherein The method for obtaining the expected effect evaluation coefficient is as follows: Among them, RS a is the recovery effect score, DS a is the deviation from the expected score, CRI is the expected effect evaluation coefficient, γ is the weight coefficient, F(Jz a , YJz a ) is specifically: the ratio of the absolute value of the difference between each parameter in Jz a and each parameter in YJz a to the normal range of the parameter. After summing up each ratio, the average value is processed. min(||Jz a ||, ||YJz a ||) is the minimum value in the norms of Jz a and YJz a , and max(||Jz a ||, ||YJz a ||) is the maximum value in the norms of Jz a and YJz a .

8. The spinal load monitoring system for dynamically adjusting a rehabilitation plan after surgery according to claim 1, wherein Adjust the initial parameters of the rehabilitation training, including the following steps: Construct the objective function: Among them, Mb is the objective function value, is the complex function of the change between the adjusted training parameters and the initial parameters of the rehabilitation training, is the rehabilitation training cost function of the adjusted training parameters, The expected effect evaluation function of the adjusted training parameters, C tol is the evaluation value of the patient's physical tolerance, are the adjusted training parameters, is the weight factor of, λ2 is is the weight factor of, λ3 is is the weight factor of, λ4 is C tol is the weight factor; Under the constraint conditions, starting from the initial parameters of the rehabilitation training in the current stage, adjust the initial parameters of the rehabilitation training by the gradient descent method until the objective function converges to the minimum value.

9. The spinal load monitoring system for dynamically adjusting a rehabilitation plan after surgery according to claim 7, wherein It is the square of the Euclidean distance between the vector representation of the initial parameters for rehabilitation training and the vector representation of the adjusted training parameters; Obtained based on the trained linear regression model: Take the adjusted training parameters as the independent variables of the trained linear regression model, and denote the function value of the trained linear regression model as Obtained based on a trained machine learning model: The input of the trained machine learning model is the adjusted training parameters, and the output is the predicted rehabilitation effect data. The predicted rehabilitation effect data is compared and analyzed with the expected recovery effect data at the current stage to obtain the function value.

10. The spinal load monitoring system for dynamically adjusting the rehabilitation plan after surgery according to claim 8, wherein Patient's physical condition assessment value C tol The acquisition method is as follows: Obtain the physical function assessment data of the patient at the current training stage, including the ODI score P ODI and the VAS pain intensity score P VAS ; Based on the physiological data during the current training stage, obtain the data of the abnormal change rate of the patient's physiological characteristics, including the maximum temperature change rate Wyb, the maximum humidity change rate Hyb, and the maximum heart rate change rate Xyb; Input the body function evaluation data and the data of the abnormal change rate of the patient's physiological characteristics into the body condition evaluation model to obtain the evaluation value of the patient's body tolerance: Among them, δ1 is the first weight factor, and δ2 is the second weight factor.

Images