Method, apparatus and device for dynamically supervising rehabilitation of patient, and readable storage medium

Abstract

Provided are a method, apparatus and device for dynamically supervising rehabilitation of a patient, and a readable storage medium. The method comprises: using an optimized machine learning algorithm to analyze regularly collected and uploaded body rehabilitation data of different body parts of a patient after an operation, to obtain the rehabilitation degree type of the operation of the patient in the current rehabilitation stage; on the basis of the rehabilitation degree type of the patient, determining a rapid body rehabilitation strategy of the patient in a new rehabilitation stage, and sending same to a terminal of primary responsible medical care personnel for adjustment and confirmation; sending the adjusted and confirmed rapid body rehabilitation strategy to a terminal of the patient for checking; and analyzing rehabilitation questionnaire response information uploaded by the terminal of the patient in the new rehabilitation stage, and sending analysis results to the terminal of the primary responsible medical care personnel. The postoperative rehabilitation condition of the patient can be effectively supervised, and targeted rehabilitation advices can be given in time.

Description

A method, device, equipment and readable storage medium for dynamically monitoring patient recovery Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and readable storage medium for dynamically monitoring patient recovery. Background Art

[0002] With the continuous development of medical technology, surgery has become a common treatment method, but post-operative recovery directly affects the patient's physical and mental health. The current monitoring of patients' post-operative recovery mainly relies on the patient's subjective feedback and remote face-to-face consultations to assess the recovery status. However, these existing methods of monitoring patients' post-operative recovery, because they mainly rely on patients' subjective feedback and simple face-to-face consultations to assess the recovery status, cannot intuitively and comprehensively understand the patient's actual recovery progress. This makes it difficult to formulate and adjust rehabilitation plans, resulting in uncertainty in rehabilitation effects; at the same time, potential rehabilitation problems or complications cannot be discovered and treated in a timely manner; and there is often no targeted provision of personalized rehabilitation advice based on the unique situation of each patient. Universal rehabilitation plans cannot meet the special needs and adaptive changes of different patients. Therefore, based on the defects of monitoring patients' post-operative recovery based on existing technologies, a new method is needed to solve these problems.

[0003] Summary of the Invention

[0004] The embodiments of the present invention provide a method, apparatus, device and readable storage medium for dynamically monitoring patient rehabilitation, which effectively monitors the patient's postoperative rehabilitation and provides targeted rehabilitation advice in a timely manner, helping the patient to recover more effectively and quickly.

[0005] An embodiment of the present invention provides a method for dynamically monitoring patient recovery, comprising:

[0006] Utilizing an optimized machine learning algorithm, regularly collected and uploaded rehabilitation data of different body parts of patients after such surgery are analyzed to determine the type of rehabilitation level of the patient at the current rehabilitation stage; the body parts include the surgical site, and include at least one of the following: a left surgical site and a right surgical site;

[0007] Determine the patient's rapid physical recovery strategy in the new recovery stage based on the patient's recovery level type and send it to the terminal of the primary medical staff for adjustment and confirmation;

[0008] Archiving the adjusted and confirmed rapid physical recovery strategy and sending it to the patient's terminal for review;

[0009] The rehabilitation questionnaire response information uploaded by the patient's terminal in the new rehabilitation stage is analyzed to determine whether the patient has followed the doctor's instructions for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and the analysis results are sent to the terminal of the responsible medical staff.

[0010] As an improvement to the above solution, the method further includes:

[0011] Receive the satisfaction evaluation of the responsible medical care uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical care.

[0012] As an improvement to the above solution, the optimized machine learning algorithm is used to analyze the regularly collected and uploaded rehabilitation data of different body parts of patients after such surgery to obtain the type of recovery level of the patient at the current rehabilitation stage, including:

[0013] Extracting features from the collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery; the physical rehabilitation data includes pain level AR, muscle potential MS, and myoelectric frequency PD;

[0014] The extracted features are input into the trained SVM model for prediction to obtain the type of recovery level of the patient in the current recovery stage for this type of surgery.

[0015] As an improvement to the above solution, the training process of the SVM model includes:

[0016] Step a: Obtain sample rehabilitation data of n patients undergoing similar surgeries and sample labels of the recovery degree type of the similar surgeries corresponding to the sample rehabilitation data of each patient, and perform feature fusion calculation on the sample rehabilitation data of the patients according to Formula 1 to obtain sample data fusion features; the sample rehabilitation data include pain degree samples AR, muscle potential samples MS, and myoelectric frequency samples PD of N different body parts of the patient after surgery; 2≤N≤3, the body parts include the surgical site, and include at least one of the following: the left side of the surgical site and the right side of the surgical site;

[0017] Z = w1*ΣARi+w2*ΣMSi+w3*ΣPDi+b, Formula 1; w1, w2, and w3 are the weights of the corresponding parameters, b is the bias term, ARi represents the pain level sample of the patient's i-th body part, MSi represents the muscle potential sample of the patient's i-th body part, and PDi represents the myoelectric frequency sample of the patient's i-th body part; Σ represents the summation;

[0018] Step b: Each sample data in the training sample set is fused with features and the corresponding sample label to form a feature vector Zk as the training sample set. The training sample set includes n feature vectors Z1, Z2, ..., Zn. For any two samples Zi and Zj, their similarity in the feature space is calculated according to formula 2:

[0019] P(Zi,Zj)=exp(-g*||Zi-Zj||^2), Equation 2; where g is the decay rate parameter;

[0020] The similarity calculation results of each pair of samples in the training sample set are used to form a kernel matrix P, whose element P(i,j) represents the similarity between the i-th and j-th samples, thereby realizing feature conversion of sample data fusion features and mapping them into a higher-dimensional feature space;

[0021] Step c: based on the kernel matrix P and in the feature space, find an optimal hyperplane that maximizes the interval between positive and negative samples in the training sample set;

[0022] Step d: Initialize the Lagrange multiplier α of all samples in the training sample set to 0, and randomly select two multipliers α for optimization; in each iteration, select two multipliers αi and αj as the multiplier pair to be optimized according to the maximum step size strategy; under the constraints of the objective function, fix the values ​​of other multipliers, optimize the selected multiplier pairs αi and αj, and thus update the weights corresponding to the two multipliers; in each iteration, check whether the optimization conditions are met according to the KKT condition (Karush-Kuhn-Tucker), and if so, adjust the threshold of the model according to the updated multiplier pair; decide whether to terminate the iteration according to a pre-set termination condition, which includes reaching the maximum number of iterations or meeting certain convergence conditions; repeat the previous steps in step d until the termination condition is met, thereby gradually adjusting the weights and thresholds corresponding to the multipliers, and continuously optimizing the support vectors and corresponding weights in the SVM model using the weights and thresholds corresponding to the multipliers, and finding the best parameter combination that minimizes the objective function to obtain the optimal segmentation hyperplane;

[0023] Among them, the objective function is: minL(α)=Σα-1 / 2*Σ(Σα*α*yi*yj*P(Zi,Zj)), the constraints are: Σα*yi=0; 0<=α<=C; α is the Lagrange multiplier, which is the weight value corresponding to each sample; yi refers to the label of the i-th training sample, which is used to indicate the recovery level category of the i-th sample; yj refers to the label of the j-th training sample, which is used to indicate the recovery level category of the j-th sample; yi and yj are 1 or -1, representing two categories respectively; P(Zi,Zj) maps the two samples Zi and Zj in the input feature space to a high-dimensional feature space and calculates their similarity; Zi is the i-th training sample in the input feature space, consisting of a set of feature vectors; Zj is the j-th training sample in the input feature space, also consisting of a set of feature vectors; C is the regularization parameter, which is used to control the complexity of the model;

[0024] Step e: Save the calculated support vectors and weight parameters as the trained SVM model.

[0025] As an improvement to the above solution, the feature extraction of the collected and uploaded body rehabilitation data of different body parts of the patient after such surgery includes:

[0026] The collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery are subjected to feature fusion calculation according to Formula 3 to obtain data fusion features; the data fusion features include pain level samples AR', muscle potential samples MS' and myoelectric frequency samples PD' of N different body parts of the patient after surgery; 2≤N≤3;

[0027] Z'=w1*ΣARi'+w2*ΣMSi'+w3*ΣPDi'+b, Formula 3; w1, w2, and w3 are the weights of the corresponding parameters, b is the bias term, ARi' represents the pain severity data of the i-th body part of the patient, MSi' represents the muscle potential data of the i-th body part of the patient, and PDi' represents the myoelectric frequency data of the i-th body part of the patient; Σ represents the summation.

[0028] As an improvement to the above solution, determining the patient's rapid physical recovery strategy in a new recovery stage according to the patient's recovery level type includes:

[0029] According to the patient's rehabilitation level type and based on the preset mapping relationship between the rehabilitation level type and the physical rapid recovery strategy, the patient's physical rapid recovery strategy in the new rehabilitation stage is determined; the physical rapid recovery strategy includes: rehabilitation measures, and the treatment implementation time range for each rehabilitation measure; the rehabilitation measures include medications required for rehabilitation, rehabilitation diet precautions and rehabilitation exercise training plans.

[0030] Another embodiment of the present invention provides a device for dynamically monitoring patient recovery, which is applied to a cloud server and includes:

[0031] A first analysis module is configured to analyze the collected and uploaded post-operative rehabilitation data of different body parts of the patient using an optimized machine learning algorithm to determine the type of rehabilitation level of the patient at the current rehabilitation stage for the type of surgery; the body parts include the surgical site, and include at least one of the following: a left surgical site and a right surgical site;

[0032] An optimization module is used to determine a rapid physical recovery strategy for the patient in a new recovery stage according to the type of recovery level of the patient and send the strategy to the terminal of the primary medical staff for adjustment and confirmation;

[0033] A sending module, configured to archive the adjusted and confirmed rapid physical recovery strategy and send it to the patient's terminal for review;

[0034] The second analysis module is used to analyze the rehabilitation questionnaire response information in the new rehabilitation stage uploaded by the patient's terminal, to analyze whether the patient complies with the doctor's orders for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and to send the analysis results to the terminal of the responsible medical staff.

[0035] As an improvement to the above solution, the device further includes:

[0036] The receiving module is used to receive the satisfaction evaluation of the responsible medical care uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical care.

[0037] Another embodiment of the present invention provides a device for dynamically monitoring patient rehabilitation, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the method for dynamically monitoring patient rehabilitation described in the above-mentioned embodiment of the invention.

[0038] Another embodiment of the present invention provides a storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for dynamically monitoring patient rehabilitation described in the above-mentioned embodiment of the invention.

[0039] Compared with the existing technology, one of the above technical solutions has the following advantages:

[0040] By using an optimized machine learning algorithm to analyze the physical rehabilitation data of different body parts of patients after such surgery that are regularly collected and uploaded, the type of rehabilitation degree of the patient in the current rehabilitation stage is obtained, so that the patient's postoperative rehabilitation can be effectively supervised; at the same time, according to the type of rehabilitation degree of the patient, the patient's physical rapid rehabilitation strategy in the new rehabilitation stage is determined and sent to the terminal of the main medical staff for adjustment and confirmation, so that targeted rehabilitation suggestions can be given in time according to the patient's postoperative rehabilitation situation; the adjusted and confirmed physical rapid rehabilitation strategy is archived and sent to the patient's terminal for review, so that the patient knows in time how to conduct rapid rehabilitation treatment in the next stage; the rehabilitation questionnaire response information uploaded by the patient's terminal in the new rehabilitation stage is analyzed to analyze whether the patient complies with the doctor's advice for rehabilitation treatment and whether the patient's rehabilitation situation is abnormal, and the analysis results are sent to the terminal of the main medical staff, so that the main medical staff can understand the patient's rehabilitation situation in time and realize the full process tracking of the patient's rehabilitation situation. It can be seen that this application can effectively supervise the patient's postoperative rehabilitation situation and give targeted rehabilitation suggestions in time, helping patients to recover more effectively and quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] FIG1 is a flow chart of a method for dynamically monitoring patient rehabilitation provided by one embodiment of the present invention;

[0042] FIG2 is a schematic structural diagram of a device for dynamically monitoring patient recovery provided by one embodiment of the present invention;

[0043] FIG3 is a schematic structural diagram of a device for dynamically monitoring patient recovery provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] 1 is a flow chart of a method for dynamically monitoring patient rehabilitation provided by an embodiment of the present invention. The method for dynamically monitoring patient rehabilitation, applied in a cloud server, comprises steps S10 to S13:

[0046] S10, using an optimized machine learning algorithm to analyze the physical rehabilitation data of different body parts of the patient after this type of surgery that are regularly collected and uploaded, to obtain the type of recovery level of the patient for this type of surgery in the current rehabilitation stage; the body parts include the surgical site, and include at least one of the following: the left side of the surgery, the right side of the surgery.

[0047] Among them, patients can go to any nearby medical point or community point with relevant physical physiological data detection equipment to collect physical rehabilitation data on a regular basis. They do not need to go to the hospital responsible for supervising the patient's rehabilitation for regular physical examinations. The collected physical rehabilitation data can be uploaded to the cloud server through the user terminal. Of course, patients can also purchase relevant physical physiological data detection equipment at home, and after completing their own examinations, they can upload the collected physical rehabilitation data to the cloud server through the user terminal. Among them, when uploading relevant physical rehabilitation data, the type of surgery can be uploaded at the same time. In addition, in order to improve the accuracy and comprehensiveness of physical rehabilitation analysis, when collecting physical rehabilitation data, physical rehabilitation data can be collected for the surgical site and some surrounding areas of the surgery. For example, if a patient has undergone leg surgery, the body parts that need to collect data can be: the leg surgery site and its surrounding areas, which can be used to evaluate the physical rehabilitation data of the recovery situation.

[0048] The cloud server uses the machine learning algorithm optimized by this application to analyze the physical rehabilitation data of different body parts of the patient after surgery, so as to timely and accurately obtain the type of recovery level of the patient's surgery at the current recovery stage.

[0049] S11, according to the type of the patient's recovery level, determine the patient's rapid physical recovery strategy in the new recovery stage and send it to the terminal of the responsible medical staff for adjustment and confirmation.

[0050] Specifically, different physical rapid recovery strategies will be formulated in advance according to different types of rehabilitation degrees, and the mapping relationship between the two will be established and saved in advance. In this way, when dynamically supervising the rehabilitation of patients, the physical rapid recovery strategy of the patient in the new rehabilitation stage can be determined in a timely and targeted manner according to the rehabilitation degree type of the patient and based on the preset mapping relationship between the rehabilitation degree type and the physical rapid recovery strategy; the physical rapid recovery strategy includes: rehabilitation measures, the treatment implementation time range of each rehabilitation measure (that is, which rehabilitation measure should be implemented in which time period and the duration of each implementation of the rehabilitation measure); the rehabilitation measures include the drugs required for rehabilitation, rehabilitation diet precautions and rehabilitation exercise training programs (for example, what type of rehabilitation exercise needs to be done). For example, for different parts of the body such as legs, hands, head, feet, etc., the rehabilitation degree types can be complete rehabilitation, high rehabilitation, moderate rehabilitation, mild rehabilitation and other types. If it is complete rehabilitation, then the rehabilitation measures include no need for rehabilitation drugs, no rehabilitation diet precautions, and no rehabilitation exercise training program.

[0051] In addition, in order to make the rapid physical recovery strategy more reasonable, the rapid physical recovery strategy determined by the system can be sent to the terminal of the responsible medical staff for adjustment and confirmation.

[0052] S12, archiving the adjusted and confirmed rapid physical recovery strategy and sending it to the patient's terminal for review.

[0053] Among them, by archiving the rapid physical recovery strategy, it is convenient to trace the patient's entire recovery process and find the corresponding responsible medical staff, so that problems can be accurately found and accurately held accountable after they occur.

[0054] S13, analyzing the rehabilitation questionnaire response information in the new rehabilitation stage uploaded by the patient's terminal, analyzing whether the patient complies with the doctor's orders for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and sending the analysis results to the terminal of the responsible medical staff.

[0055] The rehabilitation questionnaire may be pushed to the patient's terminal by the system at regular intervals according to the current rehabilitation stage, or may be sent to the patient in advance by the primary medical staff for archiving.

[0056] This embodiment uses an optimized machine learning algorithm to analyze the physical rehabilitation data of different body parts of patients after such surgery that are regularly collected and uploaded, and obtains the type of rehabilitation degree of the patient in the current rehabilitation stage, thereby effectively monitoring the patient's postoperative rehabilitation; at the same time, based on the patient's rehabilitation degree type, the patient's physical rapid rehabilitation strategy in the new rehabilitation stage is determined and sent to the terminal of the primary medical staff for adjustment and confirmation, so as to give targeted rehabilitation suggestions in time according to the patient's postoperative rehabilitation situation; the adjusted and confirmed physical rapid rehabilitation strategy is archived and sent to the patient's terminal for review, so that the patient knows in time how to conduct rapid rehabilitation treatment in the next stage; the rehabilitation questionnaire response information uploaded by the patient's terminal in the new rehabilitation stage is analyzed to determine whether the patient complies with the doctor's instructions for rehabilitation treatment and whether the patient's rehabilitation situation is abnormal, and the analysis results are sent to the terminal of the primary medical staff, so that the primary medical staff can understand the patient's rehabilitation situation in time and realize the full process tracking of the patient's rehabilitation situation. It can be seen that this application can effectively monitor the patient's postoperative rehabilitation situation and give targeted rehabilitation suggestions in time, helping patients to recover more effectively and quickly.

[0057] In one embodiment, the method further comprises:

[0058] Receive the satisfaction evaluation of the responsible medical care uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical care.

[0059] In this embodiment, patients can evaluate their satisfaction with the responsible medical staff. After receiving the evaluation, the cloud server will send it to the responsible medical staff's terminal for viewing, so that the responsible medical staff can understand the patient's evaluation of themselves. If there are any problems, the responsible medical staff can also adjust and optimize their service work attitude and service work methods in time, thereby ultimately improving the patient's user experience.

[0060] In one embodiment, the optimized machine learning algorithm is used to analyze the regularly collected and uploaded physical rehabilitation data of different body parts of patients after such surgery to obtain the type of recovery level of the patient at the current recovery stage, including:

[0061] Extracting features from the collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery; the physical rehabilitation data includes pain level AR, muscle potential MS, and myoelectric frequency PD;

[0062] The extracted features are input into the trained SVM model for prediction to obtain the type of recovery level of the patient in the current recovery stage for this type of surgery.

[0063] In this embodiment, by selecting the following three types of physical rehabilitation data: pain intensity (AR), muscle potential (MS), and myoelectric frequency (PD), the patient's postoperative recovery status in different body parts can be accurately and comprehensively reflected. Pain intensity (AR) refers to the intensity or degree of pain experienced by the patient. Pain intensity is typically quantified within a standard assessment range, such as a numerical rating scale (e.g., 0-10) or a facial expression rating scale (e.g., 5-level expression). Pain intensity (AR) is one of the most important physiological indicators in measuring a patient's postoperative recovery process, reflecting wound healing, pain management effectiveness, and the patient's overall recovery. Myoelectric frequency (MS) refers to changes in the electrical potential difference between cells within muscle tissue. Myoelectric potential is generated by measuring muscle electrical signals to observe the muscle's response and activity under stimulation. It provides information about muscle contraction and relaxation, thereby assessing the patient's muscle function and recovery progress. Myoelectric frequency (PD) refers to the frequency change of muscle electrical signals during muscle contraction. Under normal circumstances, the myoelectric frequency of muscle contraction falls within a certain range. By monitoring myoelectric frequency, the intensity and coordination of muscle contractions during a patient's postoperative recovery can be understood. Changes in electromyographic frequency can be used as one of the important indicators for evaluating patients' rehabilitation progress and muscle function.

[0064] Specifically, monitoring pain intensity, muscle potential, and EMG frequency can objectively assess the effectiveness of a patient's rehabilitation. Changes in these parameters can reflect the recovery of the surgical site, the effectiveness of pain management, and the degree of improvement in muscle function. Accurately assessing rehabilitation outcomes helps adjust rehabilitation plans, identify problems promptly, and implement appropriate measures. Furthermore, since postoperative rehabilitation needs and progress may vary among patients, monitoring parameters such as pain intensity, muscle potential, and EMG frequency allows for customized rehabilitation plans tailored to each patient's characteristics. Accurately assessing these parameters helps determine rehabilitation goals and appropriate measures for specific patients, thereby improving rehabilitation outcomes. Furthermore, monitoring pain intensity, muscle potential, and EMG frequency allows for real-time observation of changes during a patient's rehabilitation process. Abnormalities in these parameters can quickly identify and address rehabilitation issues, such as infection, congestion, and muscle weakness. Timely intervention can help avoid adverse consequences and accelerate recovery.

[0065] In one embodiment, in order to accurately analyze the recovery level of a patient at the current stage of recovery for a certain type of surgery, the present application uses an optimized SVM machine learning algorithm. The training process of the optimized SVM machine learning algorithm includes:

[0066] Step a: Obtain sample rehabilitation data of n patients undergoing similar surgeries and sample labels of the recovery degree type of the similar surgeries corresponding to the sample rehabilitation data of each patient, and perform feature fusion calculation on the sample rehabilitation data of the patients according to Formula 1 to obtain sample data fusion features; the sample rehabilitation data include pain degree samples AR, muscle potential samples MS, and myoelectric frequency samples PD of N different body parts of the patient after surgery; 2≤N≤3, the body parts include the surgical site, and include at least one of the following: the left side of the surgical site and the right side of the surgical site;

[0067] Z = w1*ΣARi+w2*ΣMSi+w3*ΣPDi+b, Formula 1; w1, w2, and w3 are the weights of the corresponding parameters, b is the bias term, ARi represents the pain level sample of the patient's i-th body part, MSi represents the muscle potential sample of the patient's i-th body part, and PDi represents the myoelectric frequency sample of the patient's i-th body part; Σ represents the summation;

[0068] Step b: Each sample data in the training sample set is fused with features and the corresponding sample label to form a feature vector Zk as the training sample set. The training sample set includes n feature vectors Z1, Z2, ..., Zn. For any two samples Zi and Zj, their similarity in the feature space is calculated according to formula 2:

[0069] P(Zi,Zj)=exp(-g*||Zi-Zj||^2), Equation 2; where g is the decay rate parameter;

[0070] The similarity calculation results of each pair of samples in the training sample set are used to form a kernel matrix P, whose element P(i,j) represents the similarity between the i-th and j-th samples, thereby realizing feature conversion of sample data fusion features and mapping them to a higher-dimensional feature space, so as to better perform classification;

[0071] Step c: Based on the kernel matrix P and in the feature space, find an optimal hyperplane to maximize the interval between positive and negative samples in the training sample set and maintain the accuracy of sample classification;

[0072] Step d: Initialize the Lagrange multiplier α of all samples in the training sample set to 0, and randomly select two multipliers α for optimization; in each iteration, select two multipliers αi and αj as the multiplier pair to be optimized according to the maximum step size strategy; under the constraints of the objective function, fix the values ​​of other multipliers, optimize the selected multiplier pairs αi and αj, and thus update the weights corresponding to the two multipliers; in each iteration, check whether the optimization conditions are met according to the KKT condition (Karush-Kuhn-Tucker), and if so, adjust the threshold of the model according to the updated multiplier pair; decide whether to terminate the iteration according to a pre-set termination condition, which includes reaching the maximum number of iterations or meeting certain convergence conditions; repeat the previous steps in step d until the termination condition is met, thereby gradually adjusting the weights and thresholds corresponding to the multipliers, and continuously optimizing the support vectors and corresponding weights in the SVM model using the weights and thresholds corresponding to the multipliers, and finding the best parameter combination that minimizes the objective function to obtain the optimal segmentation hyperplane;

[0073] Among them, the objective function is: minL(α)=Σα-1 / 2*Σ(Σα*α*yi*yj*P(Zi,Zj)), and the constraints are: Σα*yi=0; 0<=α<=C; α is the Lagrange multiplier, which is the weight value corresponding to each sample; yi refers to the label of the i-th training sample, which is used to indicate the recovery level category of the i-th sample; yj refers to the label of the j-th training sample, which is used to indicate the recovery level category of the j-th sample; yi and yj are 1 or -1, representing two categories respectively; P(Zi,Zj) maps the two samples Zi and Zj in the input feature space to the high-dimensional feature space and calculates their similarity; Zi is the i-th training sample in the input feature space, consisting of a set of feature vectors; Zj is the j-th training sample in the input feature space, also consisting of a set of feature vectors; C is the regularization parameter, which is used to control the complexity of the model;

[0074] Step e: Save the calculated support vectors and weight parameters as the trained SVM model.

[0075] In this embodiment, an optimized feature fusion algorithm is used to fuse sample data features, so that the features of the physical rehabilitation data of the surgical site or the surrounding area can be reasonably fused, so that the extracted features can more comprehensively reflect the patient's physical rehabilitation after such surgery; on this basis, by forming a feature vector Zk by combining the fusion features of each sample data in the training sample set and the corresponding sample label, and calculating their similarity in the feature space, the samples can be well classified. At the same time, the similarity calculation results of each pair of samples in the training sample set are formed into a kernel matrix P, and the sample data fusion features are transformed and mapped to a higher-dimensional feature space, so as to better classify; and by finding an optimal hyperplane in the feature space, the interval between the positive and negative samples in the training sample set is maximized, and the accuracy of the sample classification is maintained; finally, through the above-mentioned sample parameter optimization process, it is ensured that the appropriate support vector and corresponding weight of the SVM model can be selected, so as to obtain the optimal segmentation hyperplane, so that an SVM model that can more accurately identify the type of recovery degree based on the fusion features of the three parameters of pain degree, muscle potential and electromyographic frequency can be trained.

[0076] The objective function is calculated by summing the Lagrange multipliers α for all training samples, where the product Σ(Σα*α*yi*yj*P(Zi,Zj)) expresses the similarity or distance between two samples. By minimizing this objective function, the optimal set of α values ​​can be found to determine the splitting hyperplane. The constraint Σα*yi=0 means that the sum of the Lagrange multiplier α multiplied by the corresponding label yi is 0, which ensures that the classifier satisfies the Karush-Kuhn-Tucker (KKT) conditions. This condition ensures a good relationship between the predicted value and the true label for each sample. 0<=α<=C is a range constraint on the Lagrange multiplier α, ensuring that α is between 0 and C. This is because α represents a weight and must be within positive and negative bounds. Furthermore, the method for finding an optimal hyperplane in feature space can be referenced in existing methods for finding the optimal hyperplane, and this feature is not discussed here.

[0077] In one embodiment, corresponding to the training process, when the model is specifically applied, the collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery is subjected to feature extraction, including:

[0078] The collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery are subjected to feature fusion calculation according to Formula 3 to obtain data fusion features; the data fusion features include pain level samples AR', muscle potential samples MS' and myoelectric frequency samples PD' of N different body parts of the patient after surgery; 2≤N≤3;

[0079] Z'=w1*ΣARi'+w2*ΣMSi'+w3*ΣPDi'+b, Formula 3; w1, w2, and w3 are the weights of the corresponding parameters, b is the bias term, ARi' represents the pain severity data of the i-th body part of the patient, MSi' represents the muscle potential data of the i-th body part of the patient, and PDi' represents the myoelectric frequency data of the i-th body part of the patient; Σ represents the summation.

[0080] In this embodiment, an optimized feature fusion algorithm is used to fuse sample data features, so that the features of the physical rehabilitation data of the surgical site or the surrounding areas can be reasonably fused, so that the extracted features can more comprehensively reflect the patient's physical rehabilitation after such surgery; on this basis, the above-mentioned optimized SVM model is used to more accurately identify the type of rehabilitation degree based on the fusion features of the three parameters of pain degree, muscle potential and electromyographic frequency.

[0081] FIG2 is a schematic diagram of a device for dynamically monitoring patient recovery according to an embodiment of the present invention. The device for dynamically monitoring patient recovery includes:

[0082] A first analysis module 10 is configured to analyze the collected and uploaded post-operative rehabilitation data of different body parts of the patient using an optimized machine learning algorithm to determine the type of rehabilitation level of the patient at the current rehabilitation stage for the type of surgery; the body part includes the surgical site, and includes at least one of the following: a left surgical site and a right surgical site;

[0083] An optimization module 11 is configured to determine a rapid physical recovery strategy for the patient in a new recovery stage according to the type of recovery level of the patient and send the strategy to a terminal of the primary medical staff for adjustment and confirmation;

[0084] A sending module 12 is used to archive the adjusted and confirmed rapid physical recovery strategy and send it to the patient's terminal for review;

[0085] The second analysis module 13 is used to analyze the rehabilitation questionnaire response information in the new rehabilitation stage uploaded by the patient's terminal, analyze whether the patient complies with the doctor's advice for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and send the analysis results to the terminal of the responsible medical staff.

[0086] In one embodiment, the apparatus further comprises:

[0087] The receiving module is used to receive the satisfaction evaluation of the responsible medical care uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical care.

[0088] This embodiment uses an optimized machine learning algorithm to analyze the physical rehabilitation data of different body parts of patients after such surgery that are regularly collected and uploaded, and obtains the type of rehabilitation degree of the patient in the current rehabilitation stage, thereby effectively monitoring the patient's postoperative rehabilitation; at the same time, based on the patient's rehabilitation degree type, the patient's physical rapid rehabilitation strategy in the new rehabilitation stage is determined and sent to the terminal of the primary medical staff for adjustment and confirmation, so as to give targeted rehabilitation suggestions in time according to the patient's postoperative rehabilitation situation; the adjusted and confirmed physical rapid rehabilitation strategy is archived and sent to the patient's terminal for review, so that the patient knows in time how to conduct rapid rehabilitation treatment in the next stage; the rehabilitation questionnaire response information uploaded by the patient's terminal in the new rehabilitation stage is analyzed to determine whether the patient complies with the doctor's instructions for rehabilitation treatment and whether the patient's rehabilitation situation is abnormal, and the analysis results are sent to the terminal of the primary medical staff, so that the primary medical staff can understand the patient's rehabilitation situation in time and realize the full process tracking of the patient's rehabilitation situation. It can be seen that this application can effectively monitor the patient's postoperative rehabilitation situation and give targeted rehabilitation suggestions in time, helping patients to recover more effectively and quickly.

[0089] It should be noted that other contents of the above-mentioned embodiment of the apparatus for dynamically monitoring patient rehabilitation can also refer to the relevant contents of the above-mentioned embodiment of the method for dynamically monitoring patient rehabilitation, which will not be repeated here.

[0090] Referring to FIG3 , which is a schematic diagram of an apparatus for dynamically monitoring patient rehabilitation provided by an embodiment of the present invention. The apparatus for dynamically monitoring patient rehabilitation of this embodiment includes: a processor 100, a memory 101, and a computer program stored in the memory 101 and executable on the processor 100, such as a program for dynamically monitoring patient rehabilitation. When the processor 100 executes the computer program, the steps of each of the above-mentioned method embodiments for dynamically monitoring patient rehabilitation are implemented. Alternatively, when the processor 100 executes the computer program, the functions of each module / unit in each of the above-mentioned device embodiments are implemented.

[0091] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 101 and executed by the processor 100 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the device for dynamically monitoring patient rehabilitation.

[0092] The device for dynamically monitoring patient rehabilitation can be a computing device such as a cloud server. The device for dynamically monitoring patient rehabilitation can include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the schematic diagram is merely an example of a device for dynamically monitoring patient rehabilitation and does not constitute a limitation on the device for dynamically monitoring patient rehabilitation. The device can include more or fewer components than shown in the diagram, or a combination of certain components, or different components. For example, the device for dynamically monitoring patient rehabilitation can also include input and output devices, network access devices, buses, and the like.

[0093] The processor may be a central processing unit (CPU), or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the device for dynamically monitoring patient rehabilitation, and utilizes various interfaces and lines to connect various parts of the device for dynamically monitoring patient rehabilitation.

[0094] The memory can be used to store the computer programs and / or modules, and the processor implements the various functions of the device for dynamically monitoring patient rehabilitation by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0095] Wherein, if the module / unit integrated in the device for dynamically monitoring patient rehabilitation is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, it can implement the steps of the above-mentioned various method embodiments. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practices in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practices, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0096] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0097] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for dynamically monitoring patient rehabilitation, characterized in that: Applied to cloud servers, including: The physical rehabilitation data of different body parts of the patient after such surgery that are regularly collected and uploaded are analyzed by using an optimized machine learning algorithm to obtain the type of rehabilitation degree of the patient for such surgery at the current rehabilitation stage; the body parts include the surgical site, and include at least one of the following: the left side of the surgery site, and the right side of the surgery site; According to the type of the patient's recovery level, determine the patient's physical rapid recovery strategy in the new recovery stage and send it to the terminal of the responsible medical staff for adjustment and confirmation; Archiving the adjusted and confirmed physical rapid recovery strategy and sending it to the patient's terminal for review; The rehabilitation questionnaire response information uploaded by the patient's terminal in the new rehabilitation stage is analyzed to determine whether the patient has followed the doctor's orders for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and the analysis results are sent to the terminal of the responsible medical staff.

2. The method for dynamically monitoring patient rehabilitation according to claim 1, characterized in that: The method further comprises: Receive the satisfaction evaluation of the responsible medical care uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical care.

3. The method for dynamically monitoring patient rehabilitation according to claim 1, characterized in that: The optimized machine learning algorithm is used to analyze the physical rehabilitation data of different body parts of patients after such surgery that are regularly collected and uploaded, and the type of rehabilitation degree of the patient's such surgery at the current rehabilitation stage is obtained, including: Extracting features from the collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery; the physical rehabilitation data includes pain level AR, muscle potential MS and myoelectric frequency PD; The extracted features are input into the trained SVM model for prediction to obtain the patient's current rehabilitation stage. The type of recovery that the patient will experience from this type of surgery.

4. The method for dynamically monitoring patient rehabilitation according to claim 3, characterized in that: The training process of the SVM model includes: Step a: Obtain sample rehabilitation data of similar surgeries of n patients and sample labels of the type of rehabilitation degree of similar surgeries corresponding to the sample rehabilitation data of each patient, and perform feature fusion calculation on the sample rehabilitation data of the patient according to Formula 1 to obtain sample data fusion features; the sample rehabilitation data includes pain degree samples AR, muscle potential samples MS and myoelectric frequency samples PD of N different body parts of the patient after surgery; 2≤N≤3, the body parts include surgical sites, and include at least one of the following: surgical left site, surgical right site; Z=w1*∑ARi+w2*∑MSi+w3*∑PDi+b, Formula 1; w1, w2 and w3 are weights of corresponding parameters respectively, b is bias term, ARi represents the pain degree sample of the i-th body part of the patient, MSi represents the muscle potential sample of the i-th body part of the patient, PDi represents the myoelectric frequency sample of the i-th body part of the patient; ∑ represents summation; Step b: The features of each sample data in the training sample set are fused with the corresponding sample label to form a feature vector Zk as the training sample set. The training sample set includes n feature vectors Z1, Z2, ..., Zn; for any two samples Zi and Zj, their similarity in the feature space is calculated according to Formula 2: P(Zi, Zj) = exp(-g*||Zi-Zj||^2), Formula 2; where g is the decay rate parameter; The similarity calculation results of each pair of samples in the training sample set form a kernel matrix P, whose element P(i, j) represents the similarity between the i-th and j-th samples, so as to realize feature conversion of sample data fusion features and map them to a higher-dimensional feature space; Step c: based on the kernel matrix P and in the feature space, find an optimal hyperplane so as to maximize the interval between the positive and negative samples in the training sample set; Step d: Initialize the Lagrange multipliers α of all samples in the training sample set to 0, and randomly select two multipliers α for optimization; in each iteration, select two multipliers αi according to the maximum step size strategy and αj as the multiplier pair to be optimized; under the constraint of the objective function, fix the values ​​of other multipliers, optimize the selected multiplier pair αi and αj, so as to update the weights corresponding to the two multipliers; in each iteration, check whether the optimization condition is met according to the KKT condition (Karush-Kuhn-Tucker), if so, adjust the threshold of the model according to the updated multiplier pair; decide whether to terminate the iteration according to the pre-set termination condition, which includes reaching the maximum number of iterations or satisfying certain convergence conditions; repeat the previous steps in step d until the termination condition is met, so as to gradually adjust the weights and thresholds corresponding to the multipliers, and use the weights and thresholds corresponding to the multipliers to continuously optimize the support vectors and corresponding weights in the SVM model, find the best parameter combination that minimizes the objective function, and obtain the optimal segmentation hyperplane; Wherein, the objective function is: minL(α)=∑α-1 / 2*∑(∑α*α*yi*yj*P(Zi, Zj)), and the constraints are: ∑α*yi=0; 0<=α<=C; α is the Lagrange multiplier, which is the weight value corresponding to each sample; yi refers to the label of the i-th training sample, which is used to indicate the recovery degree category of the i-th sample; yj refers to the label of the j-th training sample, which is used to indicate the recovery degree category of the j-th sample; yi and yj are 1 or -1, representing two categories respectively; P(Zi, Zj) maps the two samples Zi and Zj in the input feature space to the high-dimensional feature space and calculates their similarity; Zi is the i-th training sample in the input feature space, which consists of a set of feature vectors; Zj is the j-th training sample in the input feature space, which also consists of a set of feature vectors; C is the regularization parameter, which is used to control the complexity of the model; Step e: Save the calculated support vector and weight parameters as the trained SVM model.

5. The method for dynamically monitoring patient rehabilitation according to claim 4, characterized in that: The collected and uploaded body rehabilitation data of different body parts of the patient after such surgery are subjected to feature extraction, including: The collected and uploaded physical rehabilitation data of different body parts of the patient after such surgery are subjected to feature fusion calculation according to Formula 3 to obtain data fusion features; the data fusion features include pain level samples AR', muscle potential samples MS' and myoelectric frequency samples PD' of N different body parts of the patient after surgery; 2≤N≤3; Z'=w1*∑ARi'+w2*∑MSi'+w3*∑PDi'+b, Formula 3; w1, w2 and w3 are respectively The weight of the response parameter, b is the bias term, ARi' represents the pain severity data of the i-th body part of the patient, MSi' represents the muscle potential data of the i-th body part of the patient, PDi' represents the myoelectric frequency data of the i-th body part of the patient; ∑ represents the sum.

6. The method for dynamically monitoring patient rehabilitation according to claim 1, characterized in that: Determining a rapid physical recovery strategy for the patient in a new recovery stage according to the type of recovery level of the patient includes: According to the patient's rehabilitation level type and based on the preset mapping relationship between the rehabilitation level type and the physical rapid recovery strategy, the patient's physical rapid recovery strategy in the new rehabilitation stage is determined; the physical rapid recovery strategy includes: rehabilitation measures, and the treatment implementation time range of each rehabilitation measure; the rehabilitation measures include medications required for rehabilitation, rehabilitation diet precautions and rehabilitation exercise training plans.

7. A device for dynamically monitoring patient recovery, characterized in that: Applied to cloud servers, including: The first analysis module is used to analyze the collected and uploaded physical rehabilitation data of different body parts of the patient after surgery using an optimized machine learning algorithm to obtain the type of rehabilitation degree of the patient for this type of surgery at the current rehabilitation stage; the body parts include the surgical site, and include at least one of the following: the left side of the surgery site, and the right side of the surgery site; An optimization module is used to determine the patient's physical rapid recovery strategy in the new recovery stage according to the type of recovery level of the patient and send it to the terminal of the responsible medical staff for adjustment and confirmation; A sending module, used for archiving the adjusted and confirmed rapid physical recovery strategy and sending it to the patient's terminal for viewing; The second analysis module is used to analyze the rehabilitation questionnaire response information in the new rehabilitation stage uploaded by the patient's terminal, to analyze whether the patient complies with the doctor's orders for rehabilitation treatment and whether the patient's rehabilitation condition is abnormal, and to send the analysis results to the terminal of the responsible medical staff.

8. The device for dynamically monitoring patient rehabilitation as claimed in claim 7, characterized in that: The device also includes: The receiving module is used to receive the satisfaction evaluation of the responsible medical staff uploaded by the patient's terminal, archive the satisfaction evaluation and send it to the terminal of the responsible medical staff.

9. A device for dynamically monitoring patient rehabilitation, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for dynamically supervising patient rehabilitation as claimed in any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for dynamically monitoring patient rehabilitation as described in any one of claims 1 to 6.

Images