A remote rehabilitation management method and system based on virtual reality

Abstract

The present invention belongs to the field of remote virtual rehabilitation training, and in particular discloses a remote rehabilitation management method and system based on virtual reality. The method first integrates and converts the electronic medical records of the rehabilitated person into structured data, and uses a comprehensive evaluation algorithm and machine learning to predict the rehabilitated person's current health status, expected rehabilitation goals and required cycles. Secondly, based on the evaluation results, personalized training plans are intelligently recommended from the virtual rehabilitation training strategy library. During the training process, sensors continuously monitor the rehabilitated person's physiological indicators and micro-expression reactions, evaluate the rehabilitation progress through real-time data analysis, and promptly feedback to the training strategy library, dynamically adjust the training intensity and cycle, and ensure that the training is closely matched with the rehabilitated person's actual health status. The present invention realizes precise and intelligent management of rehabilitation training by integrating electronic medical record processing, virtual reality technology, machine learning prediction and visual emotion recognition, significantly improving rehabilitation efficiency and experience.

Description

Technical Field

[0001] The present invention belongs to the technical field of communication security, and in particular discloses a remote rehabilitation management method and system based on virtual reality. Background Art

[0002] With the rapid development of science and technology, particularly in the healthcare sector, virtual reality (VR) technology, with its unique immersive experience and high interactivity, is becoming a key driver of innovation in rehabilitation treatment. As a branch of modern healthcare, remote rehabilitation management aims to break through geographical limitations through information technology and provide patients with convenient and efficient rehabilitation services.

[0003] For example, the patent with authorization announcement number CN111312363B discloses a two-hand coordination enhancement system based on virtual reality, which includes a training mode management module, a model management module, a scene management module, a training management module, a data analysis module, and a system configuration management module; wherein the training mode management module is used to provide a mode of two-hand coordination control capability; the scene selection module is used to provide different application scenarios; the model management module is used to provide different entity models; the data analysis module is used to compare and analyze different indicator data collected in the training management module; and the system configuration management module is used to manage file output paths and configuration parameters.

[0004] For example, the patent with publication number CN115985461A discloses a rehabilitation training system based on virtual reality, which includes: an information acquisition module, a data processing module, a control module, a virtual scene module and an evaluation and suggestion module; the information acquisition module is used to collect the user's patient bone data and the user's active input information; the data processing module is used to process the patient's bone data to obtain data processing results; the control module is used to generate motion control instructions for synchronizing the patient and the character model based on the bone data processing results; the virtual scene module is used to establish a virtual scene based on the user's active input information and control the character model in the virtual scene according to the motion control instructions; the evaluation and suggestion module is used to score the user's movement behavior and output the next rehabilitation training suggestion.

[0005] The above existing technologies have the following problems: 1) Remote rehabilitation resources are unevenly distributed among different regions and different medical institutions, and there is a lack of unified treatment standards and effect evaluation systems, which affects the quality and comparability of rehabilitation services; 2) Current methods often focus on the single application of VR technology, such as simulating sports rehabilitation scenarios, but lack deep integration with other advanced technologies such as biosensors and artificial intelligence. This results in the VR rehabilitation process being unable to make timely adjustments to the rehabilitation process and environment according to the current status of the rehabilitator, which limits the formulation of personalized treatment plans; in order to solve the above problems, the present invention provides a remote rehabilitation management method and system based on virtual reality. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention proposes a remote rehabilitation management method and system based on virtual reality. The method first integrates and converts the electronic medical records of the rehabilitated person into structured data, and uses a comprehensive evaluation algorithm and machine learning to predict the rehabilitated person's current health status, expected rehabilitation goals and required cycles. Secondly, based on the evaluation results, personalized training plans are intelligently recommended from the virtual rehabilitation training strategy library. During the training process, sensors continuously monitor the rehabilitated person's physiological indicators and micro-expression reactions, evaluate the rehabilitation progress through real-time data analysis, and provide timely feedback to the training strategy library to dynamically adjust the training intensity and cycle to ensure that the training is closely matched with the rehabilitated person's actual health status. The present invention integrates electronic medical record processing, virtual reality technology, machine learning prediction and visual emotion recognition to achieve precise and intelligent management of rehabilitation training, significantly improving rehabilitation efficiency and experience.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A remote rehabilitation management method based on virtual reality, comprising:

[0009] S1. Obtain the electronic medical record information of the recovered patients through the electronic medical record subsystem integration, and perform structural transformation on the obtained electronic medical record information to obtain a structured electronic medical record database;

[0010] S2. Based on the acquired structured electronic medical record database, the virtual reality subsystem's built-in comprehensive assessment algorithm is used to obtain the patient's initial first health status assessment score and health status level. A machine learning algorithm is then used to predict the patient's target rehabilitation status and corresponding rehabilitation period based on the acquired electronic medical record data and the initial first health status assessment score.

[0011] S3, based on the obtained initial first health status assessment score, target rehabilitation status and corresponding rehabilitation period, recommend corresponding rehabilitation training strategies to the patient through the virtual rehabilitation training strategy library and recommendation algorithm built into the virtual reality subsystem;

[0012] S4. Using sensors to monitor the physiological signal data of the rehabilitated person in real time during the training process according to the recommended rehabilitation training strategy, and using a built-in comprehensive evaluation algorithm to evaluate the physiological signal data obtained in real time to obtain a real-time first health status assessment score and a corresponding health status level;

[0013] S5. Compare the obtained real-time first health status assessment score and the corresponding health status level with the predicted target rehabilitation status, and feed the comparison result back to the virtual rehabilitation training strategy library to adjust the recommended rehabilitation training strategy in real time;

[0014] S6. Using the built-in visual monitoring algorithm of the virtual reality subsystem, monitor in real time the changes in the patient's micro-expressions during the recommended rehabilitation training strategy. Based on the acquired micro-expression changes, adjust the intensity of the recommended rehabilitation training strategy in real time, so that the intensity of the training strategy corresponds to the training intensity that the patient can tolerate.

[0015] The health status level in S4 includes a first health status level, a second health status level, a third health status level, a fourth health status level and a fifth health status level, wherein the first health status level, the second health status level, the third health status level, the fourth health status level and the fifth health status level correspond to the very serious, serious, general, relatively healthy and healthy indicators in the evaluation indicator set obtained by the expert experience method, respectively;

[0016] Specific steps for real-time adjustments to recommended rehabilitation strategies include:

[0017] S501, comparing the real-time first health state assessment score of the recovered person obtained in S4 with the score corresponding to the target rehabilitation state and the initial first health state assessment score, to obtain a score difference δ1 between the initial first health state assessment score and the real-time first health state assessment score, and a score difference δ2 between the real-time first health state assessment score and the score corresponding to the target rehabilitation state;

[0018] S502, setting rehabilitation discrimination thresholds α1 and α2, and based on the obtained score difference δ1 and score difference δ2, when δ1>α1 and δ2<α2, maintaining the currently recommended rehabilitation training strategy to conduct rehabilitation training for the patient until the target rehabilitation state is reached;

[0019] S503. When δ1>α1 and δ2≥α2, the obtained real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment, and the training intensity corresponding to the recommended rehabilitation training strategy is adjusted according to the initial first health status assessment score and health status level at the current moment;

[0020] S504. When δ1≤α1, the obtained real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment. A new rehabilitation training strategy is re-matched and recommended from the rehabilitation training strategies using a recommendation algorithm based on the initial first health status assessment score and health status level at the current moment, so that after training with the newly recommended rehabilitation training strategy, the real-time first health status assessment score of the rehabilitated person satisfies δ1>α1 and δ2<α2;

[0021] The specific steps of adjusting the training intensity in S503 include:

[0022] S5031: Build an intensity adjustment model based on the PID control algorithm, and initialize the PID control algorithm using the difference between the real-time first health status assessment score and the midpoint of the health status grade score interval corresponding to the current training intensity level;

[0023] S5032. If the difference between the real-time first health status assessment score and the midpoint of the health status rating interval corresponding to the current training intensity level is greater than half the radius of the rating interval, then using the intensity adjustment model to adjust the training intensity to the training intensity level corresponding to the health status level to which the real-time first health status assessment score belongs based on the real-time first health status assessment score;

[0024] S5033: If the difference between the real-time first health status assessment score and the midpoint of the health status rating interval corresponding to the current training intensity level is less than or equal to half of the rating interval radius, the training intensity of the corresponding rehabilitation training strategy is not adjusted.

[0025] Specifically, the specific steps for obtaining the initial first health status assessment score of the recovered person include:

[0026] S201. According to the rehabilitation training needs of the rehabilitated person, extract the electronic medical record data of the rehabilitated person's corresponding rehabilitation limb from the structured electronic medical record database in S1, and set it as the main target;

[0027] S202: Based on the acquired electronic medical record data corresponding to the rehabilitated limb part and the set primary goal, the acquired electronic medical record data corresponding to the rehabilitated limb part is input into a principal component analysis algorithm to calculate the cumulative contribution rate of each physiological parameter variable in the electronic medical record data corresponding to the rehabilitated limb part to the set primary goal;

[0028] S203: Setting a cumulative contribution rate threshold, and retaining the physiological parameter variables whose cumulative contribution rates are greater than the cumulative contribution rate threshold as evaluation variables for the first health status evaluation;

[0029] S204: According to the acquired first health status evaluation variable, set the comprehensive fuzzy evaluation first-level evaluation factor set s=(s1…s i …s n ) and the corresponding secondary evaluation factor set s i =(s i1 …s ij …s im ), and set the corresponding evaluation index set u=(u1…u k …u K ), where s i represents the i-th first-level evaluation factor in the first-level evaluation factor set, s ij Indicates the jth secondary evaluation factor corresponding to the i-th first-level evaluation factor, u k represents the kth evaluation index.

[0030] Specifically, the specific steps of obtaining the initial first health status assessment score of the recovered person also include:

[0031] S205, through the expert experience method, obtain the membership of the corresponding evaluation factors in the first-level evaluation factor set and the second-level evaluation factor set and the corresponding evaluation indicators in the evaluation indicator set, and use the obtained membership to obtain the corresponding first-level fuzzy evaluation matrix A nK And the second-level fuzzy evaluation matrix A mK , and at the same time, the first-level fuzzy evaluation matrix A is obtained nK And the second-level fuzzy evaluation matrix A mK Obtain the corresponding first-level indicator weight vector w and second-level indicator weight vector w through the entropy weight method i ;

[0032] S206, according to the obtained first-level fuzzy evaluation matrix A nK and the secondary fuzzy evaluation matrix A mK ,The validity of the corresponding evaluation matrix is ​​tested through the consistency test formula to ,determine the validity of the first level fuzzy evaluation matrix and the second level fuzzy evaluation matrix obtained through ,expert scoring;

[0033] S207. Setting a consistency ratio threshold. When the calculated consistency ratios of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are both less than or equal to the consistency ratio threshold, the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are considered valid. If the consistency ratio of one of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix is ​​greater than the consistency ratio threshold, the corresponding fuzzy evaluation matrix fails the test, and the corresponding matrix is ​​re-scored to obtain an updated fuzzy evaluation matrix and a corresponding indicator weight vector.

[0034] S208, the obtained effective first-level fuzzy evaluation matrix and second-level fuzzy evaluation matrix and their corresponding first-level index weight vector w and second-level index weight vector w i The data are input into the comprehensive fuzzy algorithm to calculate the initial first health status assessment score corresponding to the individual rehabilitated person, and the corresponding health status level is obtained according to the calculated initial first health status assessment score.

[0035] A remote rehabilitation management system based on virtual reality, including: a medical record integration module, a health status estimation module;

[0036] The medical record integration module includes a data acquisition unit and a structured conversion unit; the data acquisition unit is used to obtain the electronic medical record information of the recovered patients in the electronic medical record subsystem; the structured conversion unit is used to perform structural conversion on the obtained electronic medical record information to generate a structured electronic medical record database;

[0037] The health status estimation module includes a comprehensive evaluation unit and a target prediction unit; the comprehensive evaluation unit is used to use the built-in comprehensive evaluation algorithm to obtain the initial first health status assessment score and health status level of the rehabilitated person; the target prediction unit is used to combine the machine learning algorithm to predict the target rehabilitation status and corresponding rehabilitation cycle of the rehabilitated person based on the electronic medical record data and the initial first health status assessment score.

[0038] Specifically, the rehabilitation management system also includes a recommendation module and a physiological monitoring and adjustment module;

[0039] The recommendation module includes a rehabilitation training strategy library unit and a recommendation algorithm unit; the rehabilitation training strategy library unit is used to construct a rehabilitation training strategy library based on the acquired historical rehabilitation training process using the knowledge graph algorithm, and to update the constructed rehabilitation training strategy library in real time; the recommendation algorithm unit is used to select and recommend matching rehabilitation training strategies from the strategy library using the recommendation algorithm based on the initial first health status assessment score, target rehabilitation status, and rehabilitation cycle;

[0040] The physiological monitoring adjustment module includes a real-time evaluation unit and a strategy adjustment unit; the real-time evaluation unit is used to evaluate the physiological signal data acquired in real time based on the real-time monitoring of the rehabilitator's physiological signal data using a built-in comprehensive evaluation algorithm to obtain a real-time first health status evaluation score and a corresponding health status level; the strategy adjustment unit is used to compare the real-time evaluation results with the predicted target rehabilitation status, and to make real-time adjustments to the recommended rehabilitation training strategy and rehabilitation cycle based on the comparison results.

[0041] Specifically, the rehabilitation management system also includes an intensity adjustment module;

[0042] The intensity adjustment module includes a micro-expression monitoring unit and an intensity adjustment unit; the micro-expression monitoring unit is used to monitor the micro-expression changes of the rehabilitator in real time through the visual monitoring algorithm built into the virtual reality subsystem, and to score the micro-expression changes; the intensity adjustment unit is used to adjust the recommended rehabilitation training strategy and the corresponding training intensity in the training strategy in real time according to the obtained micro-expression change score and the health status level change obtained by the real-time evaluation unit.

[0043] A computer-readable storage medium stores computer instructions, which, when executed, execute a remote rehabilitation management method based on virtual reality.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. The present invention addresses the deficiencies of existing technologies. By integrating electronic medical record subsystems, the present invention achieves the structuring and standardization of rehabilitation patient information, which helps to establish unified rehabilitation service standards and effect evaluation systems across different regions and medical institutions, thereby improving the quality and comparability of rehabilitation services. Secondly, by deeply integrating advanced technologies such as virtual reality, biosensors, and artificial intelligence, the rehabilitation process can be personalized according to the real-time status of the rehabilitation patient, optimizing the formulation of treatment plans. In particular, by real-time monitoring of the rehabilitation patient's physiological signals and micro-expression changes, the present invention can accurately assess the rehabilitation patient's health status and dynamically adjust training strategies and intensity accordingly, ensuring the scientific nature and effectiveness of the rehabilitation process. The present invention;

[0046] 2. In response to the shortcomings of existing technologies, the present invention evaluates the health status of the rehabilitated person in real time and compares it with the target rehabilitation status. It can intelligently adjust the recommended rehabilitation training strategy and rehabilitation cycle to ensure the efficiency and personalization of the rehabilitation process. In particular, through the built-in five different training intensity levels and intensity adjustment model, the present invention can adaptively adjust the training intensity according to the real-time status of the rehabilitated person, further improving the rehabilitation effect. In addition, when the rehabilitation training strategy is ineffective, the system can re-match a new training strategy to ensure the scientific nature and pertinence of the treatment plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of a remote rehabilitation management method based on virtual reality according to Example 1 of the present invention;

[0048] Figure 2 This is a flow chart of the calculation process of the initial first health status assessment score of a recovered person in Example 1 of the present invention;

[0049] Figure 3 This is a structural diagram of the micro-expression recognition evaluation model in Example 1 of the present invention;

[0050] Figure 4 This is a module diagram of a remote rehabilitation management system based on virtual reality in Example 2 of the present invention. DETAILED DESCRIPTION

[0051] Example 1

[0052] See also Figure 1 The present invention provides an embodiment of a remote rehabilitation management method based on virtual reality, which specifically comprises the following steps:

[0053] S1. The electronic medical record information of the recovered patient is obtained through the electronic medical record subsystem integration, and the obtained electronic medical record information is structurally transformed to obtain a structured electronic medical record database. Furthermore, the electronic medical record information of the recovered patient includes historical electronic medical records, real-time physical physiological parameter examination reports, and corresponding electronic medical order data issued during historical rehabilitation treatment. The specific steps of structurally transforming the obtained electronic medical record information in this embodiment include:

[0054] S101. Automatically extract key medical information from unstructured or semi-structured electronic medical record text using a natural language processing algorithm or regular expression algorithm configured within the electronic medical record subsystem, and perform rule-based permission encryption on the portion of the extracted data involving the patient's basic information using an information encryption algorithm. In this embodiment, the key medical information includes the patient's basic information (name, age, gender), diagnosis records, rehabilitation training process, examination results, and rehabilitation physician's instructions and notes. In this embodiment, the encryption algorithm includes a permission encryption algorithm, meaning that the extracted patient's basic information is only displayed to administrators who have been granted viewing permission. For those who are not granted permission, the information is displayed using a pseudonym or information hiding.

[0055] S102. Standardize the extracted information, including but not limited to standardizing medical terms into standard codes in medical vocabulary (such as ICD-10 disease codes and SNOMED CT clinical terms), and unifying drug names and examination item names to ensure data consistency and compatibility;

[0056] S103. Save the structured electronic medical record data of the recovered patients that have been standardized in S102 into a graph database to construct a structured medical record database of the recovered patients.

[0057] S2. Based on the acquired structured electronic medical record database, the virtual reality subsystem's built-in comprehensive assessment algorithm is used to obtain the patient's initial first health status assessment score and health status level. A machine learning algorithm is then used to predict the patient's target rehabilitation status and corresponding rehabilitation period based on the acquired electronic medical record data and the initial first health status assessment score.

[0058] For further information, see Figure 2 In this embodiment, the specific steps for obtaining the initial first health status assessment score of the recovered person include:

[0059] S201. Based on the rehabilitation training needs of the rehabilitated patient, extract the electronic medical record data of the corresponding rehabilitation limb part from the structured electronic medical record database in S1 and set it as the main target; for example, Achilles tendon rehabilitation, hand rehabilitation, etc.

[0060] S202: Based on the acquired electronic medical record data corresponding to the rehabilitated limb part and the set primary goal, the acquired electronic medical record data corresponding to the rehabilitated limb part is input into a principal component analysis algorithm to calculate the cumulative contribution rate of each physiological parameter variable in the electronic medical record data corresponding to the rehabilitated limb part to the set primary goal;

[0061] S203: Setting a cumulative contribution rate threshold, and retaining the physiological parameter variables whose cumulative contribution rates are greater than the cumulative contribution rate threshold as evaluation variables for the first health status evaluation;

[0062] S204: According to the acquired first health status evaluation variable, set the comprehensive fuzzy evaluation first-level evaluation factor set s=(s1…s i …s n ) and the corresponding secondary evaluation factor set s i =(s i1 …s ij …s im ), and set the corresponding evaluation index set u=(u1…u k …u K ), where s i represents the i-th first-level evaluation factor in the first-level evaluation factor set, s ijIndicates the jth secondary evaluation factor corresponding to the ith first-level evaluation factor, and uk indicates the kth evaluation index; further, in order to better illustrate the comprehensive evaluation method of this embodiment, this embodiment takes the rehabilitation factors of the Achilles tendon as an example of evaluation variables, and the first-level evaluation factor set includes the degree of recovery of physiological function, pain management, and daily activity ability; the second-level evaluation factors corresponding to the degree of recovery of physiological function include muscle strength, range of joint motion, balance ability and coordination; the second-level evaluation factors corresponding to pain management include pain intensity, pain frequency and the effect of pain relief measures; the second-level evaluation factors corresponding to daily activity ability include walking ability, going up and down stairs and daily self-care ability; among them, muscle strength is the strength recovery of the muscle groups around the Achilles tendon, the range of joint motion is the flexibility and range of motion of the ankle joint, and the balance ability is the effect of pain management. The pain intensity is assessed using a VAS (Visual Analog Scale), the pain frequency is the frequency of pain, and the pain relief effect is the effect of relief measures such as medication and physical therapy. Walking ability is a gait analysis, such as walking speed, step length, and gait stability. Stair climbing and descending is the ability and efficiency to complete stair climbing and descending. The evaluation index set set in this embodiment is set by professional rehabilitation physicians in this field as very severe, severe, general, relatively healthy, and healthy, where the corresponding score interval for very severe is [8, 10], the corresponding score interval for severe is [6, 8), the corresponding score interval for general is [4, 6), the corresponding score interval for relatively healthy is [2, 4), and the corresponding score interval for healthy is [0, 2).

[0063] S205, through the expert experience method, obtain the membership of the corresponding evaluation factors in the first-level evaluation factor set and the second-level evaluation factor set and the corresponding evaluation indicators in the evaluation indicator set, and use the obtained membership to obtain the corresponding first-level fuzzy evaluation matrix A nK And the second-level fuzzy evaluation matrix A mK , and at the same time, the first-level fuzzy evaluation matrix A is obtained nK And the second-level fuzzy evaluation matrix A mK Obtain the corresponding first-level indicator weight vector w and second-level indicator weight vector w through the entropy weight method i For example, the degree of membership of the physiological function recovery degree in the first-level evaluation factor set and the severity evaluation index in the evaluation index set; the degree of membership of the muscle strength in the second-level evaluation factor corresponding to the physiological function recovery degree and the severity evaluation index;

[0064] S206, according to the obtained first-level fuzzy evaluation matrix A nK and the secondary fuzzy evaluation matrix A mK, the validity of the corresponding evaluation matrix is ​​tested through the consistency test formula to determine the validity of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix obtained through expert scoring; in this embodiment, the consistency test can effectively prevent the evaluation results from being affected by the subjectivity of experts, making the comprehensive evaluation results more effective;

[0065] Furthermore, the consistency check formula in this embodiment is specifically: Wherein, CR represents the consistency ratio. In this embodiment, the consistency ratio threshold is set to 0.1. When CR is less than 0.1, it proves that the corresponding fuzzy evaluation matrix has passed the consistency verification and the fuzzy evaluation matrix obtained by expert scoring is valid. Otherwise, it is invalid. RI represents the expected value of the random correlation coefficient between each variable in the fuzzy evaluation matrix; λ max Represents the maximum eigenvalue of the fuzzy evaluation matrix, and r represents the order of the corresponding fuzzy evaluation matrix; for example, to verify the first-level fuzzy evaluation matrix, r = n, λ max A nK The largest characteristic root of

[0066] S207. Setting a consistency ratio threshold. When the calculated consistency ratios of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are both less than or equal to the consistency ratio threshold, the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are considered valid. If the consistency ratio of one of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix is ​​greater than the consistency ratio threshold, the corresponding fuzzy evaluation matrix fails the test, and the corresponding matrix is ​​re-scored to obtain an updated fuzzy evaluation matrix and a corresponding indicator weight vector.

[0067] S208, the obtained effective first-level fuzzy evaluation matrix and second-level fuzzy evaluation matrix and their corresponding first-level index weight vector w and second-level index weight vector w i The data are input into the comprehensive fuzzy algorithm to calculate the initial first health status assessment score corresponding to the individual rehabilitated person, and the corresponding health status level is obtained according to the calculated initial first health status assessment score.

[0068] Furthermore, the specific steps of obtaining the target rehabilitation status and corresponding rehabilitation period of the rehabilitation patient in this embodiment are as follows:

[0069] S209. Based on the structured electronic medical record data, a principal component analysis algorithm is used to obtain electronic medical record data features with a contribution rate greater than 0.8 to the rehabilitation process as input to the machine learning model. Rehabilitation physicians evaluate and annotate the target rehabilitation status and corresponding rehabilitation cycle based on the obtained electronic medical record data features to obtain a complete annotated prediction data set.

[0070] S210: Build a corresponding prediction model using a machine learning algorithm, and input the corresponding prediction data set into the prediction model for training, thereby obtaining a trained prediction model;

[0071] S211. The trained prediction model is built into the virtual reality subsystem to predict the corresponding target rehabilitation state and the corresponding rehabilitation period based on the physiological parameter data input by the sensor and the rehabilitator in real time.

[0072] In this embodiment, the target rehabilitation state corresponds to the health status level and is also obtained by predicting the first health status assessment score obtained by the rehabilitator after rehabilitation training. For example, if the predicted target rehabilitation state score is 3 points, the corresponding rehabilitation goal for the rehabilitator is relatively healthy.

[0073] S3. Based on the obtained initial first health status assessment score, target rehabilitation status and corresponding rehabilitation cycle, the corresponding rehabilitation training strategy is recommended to the rehabilitated person through the virtual rehabilitation training strategy library and recommendation algorithm built into the virtual reality subsystem; further, in this embodiment, the virtual rehabilitation training strategy library is constructed based on the obtained historical rehabilitation process training strategy and the corresponding virtual reality environment through visual algorithms and natural language algorithms, using the method steps of constructing a knowledge graph; and in the constructed virtual rehabilitation training strategy library, a built-in reasoning model is used to enable the existing virtual rehabilitation training strategy library to generate corresponding rehabilitation training videos and rehabilitation text suggestions based on the initial first health status assessment score, target rehabilitation status and corresponding rehabilitation cycle input by the rehabilitated person through video and text generation algorithms; in this embodiment, the reasoning models include Chinchilla, PaLM, TAPAS and Llama3 models; the video generation algorithms include Sora model, Phenaki model, Imagen Video; the text generation algorithms include Chinese pre-trained Bert model, Chat-GPT series model and T5 model.

[0074] S4. Real-time monitoring of the physiological signal data of the rehabilitated person during the training process according to the recommended rehabilitation training strategy is carried out through sensors, and the real-time acquired physiological signal data is evaluated through the built-in comprehensive evaluation algorithm to obtain a real-time first health status assessment score and a corresponding health status level; further, in this embodiment, the health status level includes a first health status level, a second health status level, a third health status level, a fourth health status level and a fifth health status level, and the first health status level, the second health status level, the third health status level, the fourth health status level and the fifth health status level respectively correspond to the very serious, serious, general, relatively healthy and healthy indicators in the evaluation index set, and the corresponding scoring intervals also correspond one to one.

[0075] S5. Compare the obtained real-time first health status assessment score and the corresponding health status level with the predicted target rehabilitation status, and feed the comparison result back to the virtual rehabilitation training strategy library to adjust the recommended rehabilitation training strategy in real time;

[0076] Furthermore, in this embodiment, the specific steps of adjusting the recommended rehabilitation training strategy in real time include:

[0077] S501. Compare the real-time first health status assessment score of the rehabilitated person obtained by the assessment in S4 with the score corresponding to the target rehabilitation state and the initial first health status assessment score to obtain a score difference δ1 between the initial first health status assessment score and the real-time first health status assessment score and a score difference δ2 between the real-time first health status assessment score and the score corresponding to the target rehabilitation state. In this embodiment, according to the score interval corresponding to the evaluation index set, the lower the first health status assessment score, the better the corresponding rehabilitation state.

[0078] S502, setting rehabilitation discrimination thresholds α1 and α2, according to the obtained score difference δ1 and score difference δ2, when δ1>α1 and δ2<α2, then maintaining the currently recommended rehabilitation training strategy to conduct rehabilitation training for the rehabilitated person until the target rehabilitation state is reached; in this embodiment, δ1>α1 and δ2<α2 indicate that the difference between the initial first health state assessment score and the real-time first health state assessment score is greater than the rehabilitation discrimination threshold α1, and at the same time, after training, the real-time first health state assessment score corresponding to the rehabilitated person becomes farther away from the initial first health state assessment score and closer to the target rehabilitation state. For example, if the initial first health state assessment score is 9, and after training with the currently recommended rehabilitation training strategy, the real-time first health state assessment score becomes 4, and the score corresponding to the target rehabilitation state is 3, and the set α1 is 4 and α2 is 2, then it is indicated that the rehabilitation training strategy is an effective rehabilitation training strategy; in this embodiment, the threshold is set by a person skilled in the art according to the score corresponding to the optimal target rehabilitation state obtained after historical training of the corresponding limb part of the rehabilitated person;

[0079] S503. When δ1>α1 and δ2≥α2, the acquired real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment, and the training intensity corresponding to the recommended rehabilitation training strategy is adjusted according to the initial first health status assessment score and health status level at the current moment; further, in this embodiment, by building five different training intensity levels into the configured virtual reality subsystem, including the first training intensity level, the second training intensity level, the third training intensity level, the fourth training intensity level and the fifth training intensity level; the corresponding intensity levels are adaptively adjusted through the configured intensity adjustment model; the first training intensity level, the second training intensity level, the third training intensity level, the fourth training intensity level and the fifth training intensity level The training intensity levels correspond one-to-one to the first health status level, the second health status level, the third health status level, the fourth health status level, and the fifth health status level obtained through comprehensive assessment of the rehabilitated person, and the first training intensity level corresponds to the first health status level and has the lowest training intensity, while the fifth training intensity level corresponds to the fifth health status level and has the highest training intensity. The five training intensity levels are set by those skilled in the art based on the average value of the historical training intensity of the limb part trained by the rehabilitated person at each health status level. For example, the setting of the training intensity for Achilles tendon rehabilitation is obtained by collecting the average value of the historical rehabilitation training intensity corresponding to S rehabilitated persons at each health status level: the first health status level, the second health status level, the third health status level, the fourth health status level, and the fifth health status level.

[0080] Furthermore, in this embodiment, the step of obtaining the strength adjustment model includes:

[0081] S5031: Build an intensity adjustment model based on the PID control algorithm, and initialize the PID control algorithm using the difference between the real-time first health status assessment score and the midpoint of the health status grade score interval corresponding to the current training intensity level;

[0082] S5032. If the difference between the real-time first health status assessment score and the midpoint of the health status grade scoring interval corresponding to the current training intensity level is greater than one-half of the scoring interval radius, then using the intensity adjustment model to adjust the training intensity to the training intensity level corresponding to the health status grade to which the real-time first health status assessment score belongs based on the real-time first health status assessment score; for example, if the current training intensity level is the third training intensity level and the real-time first health status assessment score at the current moment is within the scoring interval corresponding to the second health status grade, then the training intensity level is adjusted to the second training intensity level corresponding to the second health status grade;

[0083] S5033: If the difference between the real-time first health status assessment score and the midpoint of the health status rating interval corresponding to the current training intensity level is less than or equal to half of the rating interval radius, the training intensity of the corresponding rehabilitation training strategy is not adjusted.

[0084] S504. When δ1≤α1, the obtained real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment, and a new rehabilitation training strategy is re-matched and recommended from the rehabilitation training strategy using the recommendation algorithm based on the initial first health status assessment score and health status level at the current moment, so that after training with the newly recommended rehabilitation training strategy, the real-time first health status assessment score of the rehabilitated person satisfies δ1>α1 and δ2<α2.

[0085] S6. Using the built-in visual monitoring algorithm of the virtual reality subsystem, monitor in real time the changes in the patient's micro-expressions during the recommended rehabilitation training strategy. Based on the acquired micro-expression changes, adjust the intensity of the recommended rehabilitation training strategy in real time, so that the intensity of the training strategy corresponds to the training intensity that the patient can tolerate.

[0086] Further, see Figure 3 In this embodiment, the specific steps of monitoring the micro-expression change state include:

[0087] S601. Collect historical raw video data of the rehabilitated person during rehabilitation training, extract frames from the collected video data using the OPENCV toolkit, obtain facial expression image data of the rehabilitated person, and annotate the collected facial expression data with neutral, joyful, uncomfortable, and painful labels. Set scoring intervals for the set labels, with neutral micro-expressions corresponding to 0 points, joyful micro-expressions corresponding to scores greater than 0 points, discomfort corresponding to a scoring interval of [-5, -1], and pain corresponding to a scoring interval of [-6, -10];

[0088] S602: Divide the labeled image into seven parts according to eyebrows, eyes, nose, mouth, and facial contour, and label the corresponding part nodes in the order of eyebrows, eyes, nose, mouth, and facial contour; wherein eyebrows and eyes are divided into two left and right nodes;

[0089] S603, calibrate the range of each of the eyebrows, eyes, nose, mouth and facial contour parts using 8 nodes, and construct the adjacency matrix A using the connection relationship corresponding to the nodes. uv =1 indicates that there is a connection between node u and node v. uv =0, it means there is no connection between node u and node v;

[0090] S604: Construct a micro-expression recognition and evaluation model, input the obtained adjacency matrix A into the local node extraction layer of the micro-expression recognition and evaluation model, extract local facial node features, input the labeled facial expression image data into the global feature extraction layer, obtain global facial expression features, input the historical original video data of the rehabilitated person during rehabilitation training into the time information extraction layer, and obtain the variation characteristics of facial expressions in the time dimension; further, in this embodiment, the local node extraction layer is constructed using a convolutional network with a convolution kernel size of 3×3; the global feature extraction layer is constructed using a pre-trained lightweight VIT model; and the time information extraction layer is constructed using a bidirectional dilated convolutional neural network;

[0091] S605, cascade-fusing the obtained local facial node features and the global facial expression features to obtain facial local-global fusion features, and inputting the obtained facial local-global fusion features and the facial expression change features in the time dimension into a cross attention layer for secondary fusion to obtain facial local-global fusion features in the time dimension; the cross attention layer is constructed using a cross attention network;

[0092] S606: Input the acquired facial local-global fusion features in the time dimension into the classification layer constructed by the softmax function and the score regression layer constructed by the fully connected network, and input the facial micro-expression classification results and corresponding scores;

[0093] S607: Setting a training loss threshold, constructing a classification error and a scoring error using the classification results and scoring results in S606, and constructing a training loss function corresponding to the micro-expression recognition evaluation model using the classification error and scoring error. The micro-expression recognition evaluation model is trained using the constructed training loss function and training loss threshold to obtain a trained micro-expression recognition evaluation model.

[0094] S608, embedding the trained micro-expression recognition and evaluation model into the visual sensor corresponding to the virtual reality subsystem, using the visual sensor to monitor the facial images of the rehabilitated person in real time while training under the rehabilitation training strategy recommended and adjusted in S4 and S5, and using the embedded micro-expression recognition and evaluation model to recognize and score the monitored facial images based on the acquired facial images;

[0095] S609: When a micro-expression corresponding to discomfort is detected during the rehabilitation training process, the training intensity level is lowered based on the existing training intensity level; when a micro-expression corresponding to pain is detected, the current rehabilitation training process is stopped; further, lowering the training intensity level based on the existing training intensity level specifically includes:

[0096] Assuming that the current training intensity level is the fourth training intensity level and the corresponding evaluation score is -2, the corresponding training intensity level will be adjusted to the second training intensity level. That is, the difference between the corresponding score of the neutral expression and the corresponding score of the current uncomfortable micro-expression is a few points, and the training intensity level will be reduced by a few points until it reaches the first training intensity level and stops decreasing.

[0097] S7. The virtual reality subsystems corresponding to the rehabilitators who are trained in the same rehabilitation training strategy or environment are connected through the Internet of Things technology, and the rehabilitation training process of the rehabilitators with high real-time first health status assessment scores after rehabilitation training is shared with the rehabilitators with low real-time first health status assessment scores through information sharing technology, so as to realize real-time sharing of rehabilitation experience.

[0098] Example 2

[0099] See also Figure 4 , another embodiment provided by the present invention: a remote rehabilitation management system based on virtual reality, comprising: a medical record integration module, a health status estimation module, a recommendation module, a physiological monitoring adjustment module and an intensity adjustment module;

[0100] The medical record integration module is used to integrate and obtain the electronic medical record information of the recovered patients and perform structural transformation. The medical record integration module includes a data acquisition unit and a structural transformation unit. The data acquisition unit is used to obtain the electronic medical record information of the recovered patients in the electronic medical record subsystem. The structural transformation unit is used to perform structural transformation on the obtained electronic medical record information to generate a structured electronic medical record database.

[0101] The health status estimation module is used to evaluate the initial health status of the rehabilitated patient based on structured electronic medical record data and predict their target rehabilitation status and period. The health status estimation module includes a comprehensive assessment unit and a target prediction unit. The comprehensive assessment unit is used to obtain the rehabilitated patient's initial first health status assessment score and health status level using a built-in comprehensive assessment algorithm. The target prediction unit is used to combine a machine learning algorithm with the electronic medical record data and the initial first health status assessment score to predict the rehabilitated patient's target rehabilitation status and corresponding rehabilitation period.

[0102] The recommendation module is used to recommend corresponding rehabilitation training strategies based on the initial health status, target rehabilitation status and cycle of the rehabilitation patient, and to carry out rehabilitation training for the rehabilitation patient. The recommendation module includes a rehabilitation training strategy library unit and a recommendation algorithm unit. The rehabilitation training strategy library unit is used to construct a rehabilitation training strategy library based on the acquired historical rehabilitation training process using the knowledge graph algorithm, and to update the constructed rehabilitation training strategy library in real time. The recommendation algorithm unit is used to select and recommend matching rehabilitation training strategies from the strategy library using the recommendation algorithm based on the initial first health status assessment score, target rehabilitation status and rehabilitation cycle.

[0103] The physiological monitoring and adjustment module is used to monitor the physiological signal data of the rehabilitator in real time during training, evaluate the real-time health status, and adjust the strategy according to the training needs. The physiological monitoring and adjustment module includes a real-time evaluation unit and a strategy adjustment unit. The real-time evaluation unit is used to evaluate the real-time acquired physiological signal data based on the real-time monitoring of the rehabilitator's physiological signal data using a built-in comprehensive evaluation algorithm to obtain a real-time first health status evaluation score and a corresponding health status level. The strategy adjustment unit is used to compare the real-time evaluation result with the predicted target rehabilitation state and adjust the recommended rehabilitation training strategy in real time based on the comparison result.

[0104] The intensity adjustment module is used to monitor the changes in the rehabilitator's micro-expressions and rehabilitation levels in real time during the training process, and adjust the intensity of the training strategy according to the changes in micro-expressions and rehabilitation levels; the intensity adjustment module includes a micro-expression monitoring unit and an intensity adjustment unit; the micro-expression monitoring unit is used to monitor the rehabilitator's micro-expression changes in real time through the visual monitoring algorithm built into the virtual reality subsystem, and score the micro-expression changes; the intensity adjustment unit is used to adjust the recommended rehabilitation training strategy and the corresponding training intensity in the training strategy in real time according to the obtained micro-expression change score and the health status level change obtained by the real-time evaluation unit, to ensure that the training strategy intensity corresponds to the health status level in real time.

[0105] Example 3

[0106] A computer-readable storage medium stores computer instructions, which, when executed, execute a remote rehabilitation management method based on virtual reality.

[0107] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are protected by the present invention.

Claims

1. A remote rehabilitation management method based on virtual reality, characterized in that: include: S1. Obtain the electronic medical record information of the recovered patients through the electronic medical record subsystem integration, and perform structural transformation on the obtained electronic medical record information to obtain a structured electronic medical record database; S2. Based on the acquired structured electronic medical record database, the virtual reality subsystem's built-in comprehensive assessment algorithm is used to obtain the patient's initial first health status assessment score and health status level. A machine learning algorithm is then used to predict the patient's target rehabilitation status and corresponding rehabilitation period based on the acquired electronic medical record data and the initial first health status assessment score. S3, based on the obtained initial first health status assessment score, target rehabilitation status and corresponding rehabilitation period, recommend corresponding rehabilitation training strategies to the patient through the virtual rehabilitation training strategy library and recommendation algorithm built into the virtual reality subsystem; S4. Using sensors to monitor the physiological signal data of the patient during the training process according to the recommended rehabilitation training strategy in real time, and using a built-in comprehensive evaluation algorithm to evaluate the physiological signal data obtained in real time to obtain a real-time first health status assessment score and a corresponding health status level; S5. Compare the obtained real-time first health status assessment score and the corresponding health status level with the predicted target rehabilitation status, and feed the comparison result back to the virtual rehabilitation training strategy library to adjust the recommended rehabilitation training strategy in real time; S6. Using the built-in visual monitoring algorithm of the virtual reality subsystem, monitor in real time the changes in the patient's micro-expressions during the training according to the recommended rehabilitation training strategy, and adjust the intensity of the training strategy in the recommended rehabilitation training strategy in real time based on the acquired micro-expression changes; The health status level in S4 includes a first health status level, a second health status level, a third health status level, a fourth health status level and a fifth health status level, wherein the first health status level, the second health status level, the third health status level, the fourth health status level and the fifth health status level correspond to the very serious, serious, general, relatively healthy and healthy indicators in the evaluation indicator set obtained by the expert experience method, respectively; The specific steps of adjusting the recommended rehabilitation training strategy in real time include: S501, comparing the real-time first health state assessment score of the recovered person obtained in S4 with the score corresponding to the target rehabilitation state and the initial first health state assessment score, to obtain a score difference δ1 between the initial first health state assessment score and the real-time first health state assessment score, and a score difference δ2 between the real-time first health state assessment score and the score corresponding to the target rehabilitation state; S502, setting rehabilitation discrimination thresholds α1 and α2, and based on the obtained score difference δ1 and score difference δ2, when δ1>α1 and δ2<α2, maintaining the currently recommended rehabilitation training strategy to conduct rehabilitation training for the patient until the target rehabilitation state is reached; S503. When δ1>α1 and δ2≥α2, the obtained real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment, and the training intensity corresponding to the recommended rehabilitation training strategy is adjusted according to the initial first health status assessment score and health status level at the current moment; S504. When δ1≤α1, the obtained real-time first health status assessment score and its corresponding health status level are used as the initial first health status assessment score and health status level at the current moment. A new rehabilitation training strategy is re-matched and recommended from the rehabilitation training strategies using a recommendation algorithm based on the initial first health status assessment score and health status level at the current moment, so that after training with the newly recommended rehabilitation training strategy, the real-time first health status assessment score of the rehabilitated person satisfies δ1>α1 and δ2<α2; The specific steps of adjusting the training intensity in S503 include: S5031: Build an intensity adjustment model based on the PID control algorithm, and initialize the PID control algorithm using the difference between the real-time first health status assessment score and the midpoint of the health status grade score interval corresponding to the current training intensity level; S5032. If the difference between the real-time first health status assessment score and the midpoint of the health status rating interval corresponding to the current training intensity level is greater than half the radius of the rating interval, then using the intensity adjustment model to adjust the training intensity to the training intensity level corresponding to the health status level to which the real-time first health status assessment score belongs based on the real-time first health status assessment score; S5033: If the difference between the real-time first health status assessment score and the midpoint of the health status rating interval corresponding to the current training intensity level is less than or equal to half of the rating interval radius, the training intensity of the corresponding rehabilitation training strategy is not adjusted.

2. A remote rehabilitation management method based on virtual reality according to claim 1, characterized in that: The specific steps of obtaining the initial first health status assessment score of the recovered person include: S201. According to the rehabilitation training needs of the rehabilitated person, extract the electronic medical record data of the rehabilitated person's corresponding rehabilitation limb from the structured electronic medical record database in S1, and set it as the main target; S202: Based on the acquired electronic medical record data corresponding to the rehabilitated limb part and the set primary goal, the acquired electronic medical record data corresponding to the rehabilitated limb part is input into a principal component analysis algorithm to calculate the cumulative contribution rate of each physiological parameter variable in the electronic medical record data corresponding to the rehabilitated limb part to the set primary goal; S203: Setting a cumulative contribution rate threshold, and retaining the physiological parameter variables whose cumulative contribution rates are greater than the cumulative contribution rate threshold as evaluation variables for the first health status evaluation; S204: According to the acquired first health status evaluation variable, set the comprehensive fuzzy evaluation first-level evaluation factor set s=(s1…s i …s n ) and the corresponding secondary evaluation factor set s i =(s i1 …s ij …s im ), and set the corresponding evaluation index set u=(u1…u k …u K ), where s i represents the i-th first-level evaluation factor in the first-level evaluation factor set, s ij It represents the jth secondary evaluation factor corresponding to the i-th primary evaluation factor, and uk represents the k-th evaluation index.

3. A remote rehabilitation management method based on virtual reality according to claim 2, characterized in that: The specific steps of obtaining the initial first health status assessment score of the recovered person also include: S205, through the expert experience method, obtain the membership of the corresponding evaluation factors in the first-level evaluation factor set and the second-level evaluation factor set and the corresponding evaluation indicators in the evaluation indicator set, and use the obtained membership to obtain the corresponding first-level fuzzy evaluation matrix A nK And the second-level fuzzy evaluation matrix A mK , and at the same time, the first-level fuzzy evaluation matrix A is obtained nK And the second-level fuzzy evaluation matrix A mK Obtain the corresponding first-level indicator weight vector w and second-level indicator weight vector w through the entropy weight method i ; S206, according to the obtained first-level fuzzy evaluation matrix A nK and the secondary fuzzy evaluation matrix A mK ,The validity of the corresponding evaluation matrix is ​​tested through the consistency test formula to ,determine the validity of the first level fuzzy evaluation matrix and the second level fuzzy evaluation matrix obtained through ,expert scoring; S207. Setting a consistency ratio threshold. When the calculated consistency ratios of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are both less than or equal to the consistency ratio threshold, the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix are considered valid. If the consistency ratio of one of the first-level fuzzy evaluation matrix and the second-level fuzzy evaluation matrix is ​​greater than the consistency ratio threshold, the corresponding fuzzy evaluation matrix fails the test, and the corresponding matrix is ​​re-scored to obtain an updated fuzzy evaluation matrix and a corresponding indicator weight vector. S208, the obtained effective first-level fuzzy evaluation matrix and second-level fuzzy evaluation matrix and their corresponding first-level index weight vector w and second-level index weight vector w i The data are input into the comprehensive fuzzy algorithm to calculate the initial first health status assessment score corresponding to the individual rehabilitated person, and the corresponding health status level is obtained according to the calculated initial first health status assessment score.

4. A remote rehabilitation management system based on virtual reality, which is used to implement a remote rehabilitation management method based on virtual reality according to any one of claims 1 to 3, characterized in that: include: Medical record integration module, health status estimation module; The medical record integration module includes a data acquisition unit and a structured conversion unit; the data acquisition unit is used to obtain the electronic medical record information of the recovered patient in the electronic medical record subsystem; the structured conversion unit is used to perform structural conversion on the obtained electronic medical record information to generate a structured electronic medical record database; The health status estimation module includes a comprehensive evaluation unit and a target prediction unit; The comprehensive evaluation unit is used to use the built-in comprehensive evaluation algorithm to obtain the initial first health status assessment score and health status level of the rehabilitated person; the target prediction unit is used to combine the machine learning algorithm to predict the target rehabilitation status and corresponding rehabilitation cycle of the rehabilitated person based on the electronic medical record data and the initial first health status assessment score.

5. A remote rehabilitation management system based on virtual reality according to claim 4, characterized in that: The rehabilitation management system also includes a recommendation module and a physiological monitoring and adjustment module; The recommendation module includes a rehabilitation training strategy library unit and a recommendation algorithm unit; the rehabilitation training strategy library unit is used to construct a rehabilitation training strategy library based on the acquired historical rehabilitation training process using a knowledge graph algorithm, and to update the constructed rehabilitation training strategy library in real time; the recommendation algorithm unit is used to select and recommend a matching rehabilitation training strategy from the strategy library using a recommendation algorithm based on the initial first health status assessment score, the target rehabilitation status, and the rehabilitation cycle; The physiological monitoring adjustment module includes a real-time evaluation unit and a strategy adjustment unit; The real-time evaluation unit is used to evaluate the physiological signal data acquired in real time based on real-time monitoring of the rehabilitator's physiological signal data using a built-in comprehensive evaluation algorithm to obtain a real-time first health status evaluation score and a corresponding health status level; the strategy adjustment unit is used to compare the real-time evaluation results with the predicted target rehabilitation status, and to make real-time adjustments to the recommended rehabilitation training strategy and rehabilitation cycle based on the comparison results.

6. A remote rehabilitation management system based on virtual reality according to claim 5, characterized in that: The rehabilitation management system further includes an intensity adjustment module; The intensity adjustment module includes a micro-expression monitoring unit and an intensity adjustment unit; the micro-expression monitoring unit is used to monitor the micro-expression changes of the rehabilitator in real time through the visual monitoring algorithm built into the virtual reality subsystem, and score the micro-expression changes; the intensity adjustment unit is used to adjust the recommended rehabilitation training strategy and the corresponding training intensity in the training strategy in real time according to the obtained micro-expression change score and the health status level change obtained by the real-time evaluation unit.

7. A computer-readable storage medium, characterized in that Computer instructions are stored thereon, and when the computer instructions are executed, a remote rehabilitation management method based on virtual reality according to any one of claims 1 to 3 is executed.

Images