Cerebral stroke patient rehabilitation training management system and method based on Internet platform

Abstract

The invention relates to the field of cerebral apoplexy rehabilitation management, in particular to a cerebral apoplexy patient rehabilitation training management system and method based on an internet platform. The stroke patient rehabilitation training management system based on the internet platform comprises a training plan generation module, a training time dynamic adjustment module and a training intensity dynamic adjustment module. According to the method, personalized exercise priority indexes and cognitive priority indexes are calculated according to the specific exercise function recovery degree and cognitive function recovery degree of each cerebral apoplexy patient, the fields in which the patient needs more attention and training are determined, a targeted exercise function training plan and a targeted cognitive function training plan are generated according to the fields, and the training efficiency of the cerebral apoplexy patient is improved. Therefore, the allocation of rehabilitation resources is optimized; the personalized rehabilitation training plan can better meet individual differences of patients, the rehabilitation efficiency is improved, and the recovery progress is accelerated.

Description

Technical Field

[0001] The present invention relates to the field of stroke rehabilitation management, and in particular to an Internet platform-based stroke patient rehabilitation training management system and method. Background Art

[0002] With the rapid development of Internet technology and the digital transformation of the medical and health field, rehabilitation training management systems based on Internet platforms have gradually become a new trend in the rehabilitation treatment of stroke patients; stroke, as a common central nervous system disease, has a high morbidity, high mortality and high disability rate, which has brought a heavy burden to patients and their families; the purpose of rehabilitation training for stroke patients is to help patients recover or improve impaired motor and cognitive functions, improve their quality of life, and reduce their dependence on family members and society.

[0003] Traditional stroke rehabilitation training mainly relies on face-to-face guidance by professional therapists in hospitals and rehabilitation centers. This method has many limitations, such as geographical limitations, time limitations, and resource limitations. In addition, due to the lack of personalization and real-time monitoring, the effectiveness of traditional rehabilitation training is often limited, and patients are difficult to receive effective guidance and supervision when undergoing rehabilitation training at home.

[0004] In order to solve these problems, the present invention proposes a rehabilitation training management system and method for stroke patients based on an Internet platform, which provides stroke patients with personalized, real-time monitored and dynamically adjusted rehabilitation training programs through Internet technology; through this system, patients can receive professional rehabilitation training guidance through smart devices at home or other environments, and realize remote rehabilitation training. Summary of the invention

[0005] The present invention calculates personalized motor priority index and cognitive priority index according to the specific motor function recovery degree and cognitive function recovery degree of each stroke patient, determines in which areas the patient needs more attention and training, and generates targeted motor function training plan and cognitive function training plan accordingly, thereby optimizing the allocation of rehabilitation resources; this personalized rehabilitation training plan can better meet the individual differences of patients, improve rehabilitation efficiency, and accelerate recovery progress.

[0006] The rehabilitation training management method for stroke patients based on the Internet platform includes: After the patient starts rehabilitation training, a training update cycle is set. At the beginning of any training update cycle, the patient's current motor function recovery degree and cognitive function recovery degree are obtained, and then the patient's motor priority index and cognitive priority index are calculated. Based on the obtained motor priority index and cognitive priority index, the patient's motor function recovery weight and cognitive function recovery weight are calculated; the patient's basic information, including the patient's age and gender, is obtained, and the patient's basic information, current motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight are used as inputs to the training plan generation model, and the patient's motor function training plan and cognitive function training plan in the current training update cycle are output. The motor function training plan includes motor function training intensity, motor function training time and number of motor function training; the cognitive function training plan includes cognitive function training difficulty, cognitive function training time and number of cognitive function training; the patient performs rehabilitation training according to the motor function training plan and cognitive function training plan in the current training update cycle; For any motor function training of the patient within the current training update cycle, use a camera to obtain the patient's training video during the motor function training, input the patient's training video and the corresponding standard learning video into the action evaluation model, and obtain the action standard of the patient's current motor function training; set a minimum standard, and when the action standard is lower than the preset minimum standard, calculate and obtain the actual training amount and the minimum target training amount of the patient's current motor function training, and adjust the motor function training time of the patient's current motor function training based on the actual training amount and the minimum target training amount; At the same time, for any motor function training of the patient within the current training update cycle, the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained, and the patient's discomfort level in the current motor function training is calculated; if the patient's discomfort level is higher than the preset limit threshold, the motor function training intensity of the patient's next motor function training will be adjusted.

[0007] Preferably, the patient's current motor function recovery degree and cognitive function recovery degree are obtained, and then the patient's motor priority index and cognitive priority index are calculated and obtained. Based on the obtained motor priority index and cognitive priority index, the patient's motor function recovery weight and cognitive function recovery weight are calculated and obtained. The specific operations are as follows: At the beginning of any training update cycle, the Fugl-Meyer evaluation form was used to evaluate the patient's motor function, and the ratio of the motor function evaluation result to the full score of the Fugl-Meyer evaluation form (226) was calculated to obtain the patient's motor function after starting rehabilitation training. The degree of recovery of motor function in each training update cycle At the same time, the MoCA scale was used to evaluate the patient's cognitive function, and the ratio of the cognitive function evaluation result to the full score of the MoCA scale of 30 was calculated to obtain the patient's cognitive function after starting rehabilitation training. The degree of cognitive function recovery after training update cycle ; Based on the degree of motor function recovery obtained and cognitive function recovery , using the formula ; Calculate the motion priority index ;in, and are preset positive integers, and ; To control the motion priority index Constant of decay rate; cognitive priority index ; Then use the formula ; Calculate the first The weight of motor function recovery in training update cycles ;No. Cognitive function recovery weight of training update cycle .

[0008] Preferably, the actual training amount and the minimum target training amount of the patient's current motor function training are calculated and obtained, and the motor function training time of the patient's current motor function training is adjusted based on the obtained actual training amount and the minimum target training amount. The specific operations are as follows: For any sports function training in the current training update cycle, obtain the action standard degree of the current sports function training ; Set minimum standards ; The standard of the current training action Below the minimum standard When the training intensity in the sports function training plan based on the current training update cycle is and training time , using the formula Calculate and obtain the actual training volume of the current sports function training ; Using the formula Calculate the minimum target training volume ; Using the formula Calculate the additional training time required for the patient's current motor function training .

[0009] Preferably, the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained to calculate the patient's discomfort level in the current motor function training; if the patient's discomfort level is higher than the limit threshold, the motor function training intensity of the patient's next motor function training is adjusted, and the specific operations are as follows: During any exercise function training within the current training update cycle, the patient's heart rate, respiratory rate and blood pressure are monitored, and the patient's highest heart rate during the current exercise function training is obtained , maximum respiratory rate and maximum blood pressure ; Using the formula ; Calculate the heart rate excess value ,in represents the patient's heart rate at rest. Represents the expected maximum heart rate of the patient; using the formula ; Calculate the respiratory rate excess value ,in Indicates the patient's respiratory rate at rest. Represents the expected maximum respiratory rate of the patient; using the formula ; Calculate the blood pressure excess value ,in It represents the patient's blood pressure at rest. Indicates the preset maximum expected blood pressure value of the patient; uses the digital scoring method to obtain the patient's pain score , The value range is 0-10, using the formula Calculate the patient's pain threshold ; Using the formula Calculate the patient's discomfort level during current motor function rehabilitation training ; Setting the Limit Threshold , when the patient is at a level of discomfort in the current motor function rehabilitation training Above limit threshold When , use the formula Calculate the appropriate training intensity for the next motor function rehabilitation training ; To control the intensity of adaptive training The control coefficient of the decline.

[0010] Preferably, the training plan generation model is established based on a multi-task learning model, including: a first input layer, a shared layer and a task-specific layer; the first input layer is used to receive the patient's basic information, motor function recovery progress, cognitive function recovery progress, and motor function recovery weights and cognitive function recovery weights; the shared layer is used to process all input features received by the input layer, extract the relationship between the input features, and generate a comprehensive feature representation of the patient; the task-specific layer includes a motor function training plan output layer and a cognitive function training plan output layer, the motor function training plan output layer is used to receive the patient's comprehensive feature representation and motor function recovery weights, and output the patient's motor function training plan; the cognitive function training plan output layer is used to receive the patient's comprehensive feature representation and cognitive function recovery weights, and output the cognitive function training plan.

[0011] Preferably, the specific operations for training the training plan generation model are as follows: Acquire several plan generation training samples with labeled motor function training plans and cognitive function training plans, each plan generation training sample contains a patient's basic information, motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight; divide all acquired plan generation training samples into a plan generation training set and a plan generation verification set; use the plan generation training set to train the training plan generation model, and verify the training plan generation model through the plan generation verification set to obtain a first verification result; set a first training condition, and determine whether the acquired first verification result meets the first training condition, and if so, output the trained training plan generation model; if not, continue to use the plan generation training set to train the training plan generation model.

[0012] Preferably, the action evaluation model is established based on CNN and LSTM, and includes a second input layer, a CNN layer, an LSTM layer, a fully connected layer and an output layer; the second input layer is used to receive patient training videos and standard videos; the CNN layer is used to extract spatial features from the images of each frame of the patient training videos and the standard videos, and output the spatial features of each frame; the LSTM layer is used to extract the temporal features of the patient training videos and the standard videos; the fully connected layer is used to integrate the temporal features with the spatial features, and generate a scalar through the ReLU activation function; the output layer is used to output the scalar generated by the fully connected layer, which represents the action standard of the patient training video; For the training of the action evaluation model, the specific operations are as follows: Acquire several motion evaluation training samples with labeled motion standards, each of which contains a patient training video and a corresponding standard video; divide all motion evaluation training samples into a motion evaluation training set and a motion evaluation verification set, train the motion evaluation model with the motion evaluation training set, and verify the motion evaluation model with the motion evaluation verification set to obtain a second verification result; set a second training condition, and determine whether the obtained second verification result meets the second training condition; if so, output the trained motion evaluation model; if not, continue to train the motion evaluation model with the motion evaluation training set.

[0013] The rehabilitation training management system for stroke patients based on the Internet platform includes: The training plan generation module includes a priority weight calculation unit and a training plan generation unit; the priority weight calculation unit is used to obtain the patient's current motor function recovery degree and cognitive function recovery degree at the beginning of any training update cycle, and then calculate the patient's motor priority index and cognitive priority index, and calculate the patient's motor function recovery weight and cognitive function recovery weight based on the obtained motor priority index and cognitive priority index; the training plan generation unit is used to obtain the patient's basic information, the basic information includes the patient's age and gender, and use the patient's basic information, current motor function recovery progress, cognitive function recovery progress, and motor function recovery weight and cognitive function recovery weight as inputs to the training plan generation model, and output the patient's motor function training plan and cognitive function training plan in the current training update cycle; A training time dynamic adjustment module includes a motion evaluation unit and a training time adjustment unit; the motion evaluation unit is used for obtaining a patient training video using a camera during any motor function training of the patient within the current training update cycle, inputting the patient training video and the corresponding standard learning video into a motion evaluation model, and obtaining the motion standard of the patient's current motor function training; the training time adjustment unit is used for calculating and obtaining the actual training amount and the minimum target training amount of the patient's current motor function training when the motion standard is lower than a preset minimum standard, and adjusting the motor function training time of the patient's current motor function training based on the actual training amount and the minimum target training amount; The training intensity dynamic adjustment module includes a discomfort degree calculation unit and a training intensity adjustment unit; the discomfort degree calculation unit is used to obtain the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level for any motor function training of the patient within the current training update cycle, and calculate the patient's discomfort level in the current motor function training; the training intensity adjustment unit is used to adjust the motor function training intensity of the patient's next motor function training when the patient's discomfort level is higher than a preset limit threshold.

[0014] The present invention has the following advantages: The present invention calculates personalized motor priority index and cognitive priority index according to the specific motor function recovery degree and cognitive function recovery degree of each stroke patient, determines in which areas the patient needs more attention and training, and generates targeted motor function training plan and cognitive function training plan accordingly, thereby optimizing the allocation of rehabilitation resources; this personalized rehabilitation training plan can better meet the individual differences of patients, improve rehabilitation efficiency, and accelerate recovery progress.

[0015] The present invention provides instant feedback and adjustments by real-time monitoring of the patient's training video and physiological indicators, so that the training is always at the level most suitable for the patient's current condition, ensuring the quality of training and the health and safety of the patient, and avoiding insufficient or overtraining of the patient. This real-time monitoring and feedback mechanism helps to improve the patient's rehabilitation motivation and participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the structure of the Internet-based stroke patient rehabilitation training management system adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0018] Embodiment 1, a method for managing rehabilitation training for stroke patients based on an Internet platform, comprising: When the patient starts rehabilitation training, set a training update cycle. The training update cycle refers to the time interval at which the system regularly updates the patient's rehabilitation training plan. This cycle is fixed, for example, once a week, to ensure that the patient's training plan can be updated in a timely manner according to his or her latest rehabilitation status to maintain the relevance and effectiveness of the training; at the beginning of any training update cycle, obtain the patient's current motor function recovery degree and cognitive function recovery degree to provide basic data for the formulation of a personalized rehabilitation training plan to ensure that the training plan matches the patient's actual ability, then calculate and obtain the patient's motor priority index and cognitive priority index. Based on the obtained motor priority index and cognitive priority index, calculate and obtain the patient's motor function recovery weight and cognitive function recovery weight to determine which function of the patient needs more attention during the rehabilitation process, to ensure that the training plan can balance exercise and cognitive function according to the patient's specific situation. The intensity and frequency of motor and cognitive training; obtain the patient's basic information, including the patient's age and gender, and use the patient's basic information, current motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight as inputs to the training plan generation model, and output the patient's motor function training plan and cognitive function training plan in the current training update cycle. The motor function training plan includes the motor function training intensity, motor function training time and number of motor function training; the cognitive function training plan includes the cognitive function training difficulty, cognitive function training time and number of cognitive function training; the patient performs rehabilitation training according to the motor function training plan and cognitive function training plan in the current training update cycle. The purpose of the whole process is to use Internet technology and data analysis to provide stroke patients with personalized and dynamically adjusted rehabilitation training plans to improve rehabilitation efficiency and effectiveness; For any motor function training of the patient within the current training update cycle, the patient training video is obtained by using a camera during the motor function training, and the patient training video and the corresponding standard learning video are input into the action evaluation model to obtain the action standard of the patient's current motor function training. The action standard score directly reflects the patient's performance in training and is an important reference for adjusting the training plan; a minimum standard is set as the bottom line of the action quality to ensure the training effect. When the action standard is lower than the preset minimum standard, the actual training volume and the minimum target training volume of the patient's current motor function training are calculated. Based on the actual training volume and the minimum target training volume, the motor function training time of the patient's current motor function training is adjusted. Dynamic adjustment of the training time helps to optimize the patient's rehabilitation process and ensure the scientificity and effectiveness of the training plan. At the same time, it can also be flexibly adjusted according to the patient's actual situation to prevent the patient's training effect from being insufficient and improve the patient's training efficiency and rehabilitation quality. At the same time, for any motor function training of the patient within the current training update cycle, the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained, and the patient's discomfort level in the current motor function training is calculated. The discomfort level score helps to identify whether the patient feels excessive discomfort during training, which is crucial to preventing overtraining and potential injuries; if the patient's discomfort level is higher than the preset limit threshold, the motor function training intensity of the patient's next motor function training is adjusted. The limit threshold is a parameter to ensure the safety and comfort of the patient. It provides a clear standard for the system to determine whether the training plan needs to be adjusted. The role of the entire process is to ensure the safety and effectiveness of the training plan by real-time monitoring and evaluation of the patient's physiological responses; this data-driven approach can improve the level of personalization of rehabilitation training and help patients rehabilitate at the most appropriate training intensity, thereby improving rehabilitation effects and reducing potential risks.

[0019] The patient's current motor function recovery degree and cognitive function recovery degree are obtained, and then the patient's motor priority index and cognitive priority index are calculated. Based on the obtained motor priority index and cognitive priority index, the patient's motor function recovery weight and cognitive function recovery weight are calculated. The specific operations are as follows: At the beginning of any training update cycle, the Fugl-Meyer evaluation form was used to evaluate the patient's motor function, and the ratio of the motor function evaluation result to the full score of the Fugl-Meyer evaluation form (226) was calculated to obtain the patient's motor function after starting rehabilitation training. The degree of recovery of motor function in each training update cycle At the same time, the MoCA scale was used to evaluate the patient's cognitive function, and the ratio of the cognitive function evaluation result to the full score of the MoCA scale of 30 was calculated to obtain the patient's cognitive function after starting rehabilitation training. The degree of cognitive function recovery after training update cycle ; Based on the degree of motor function recovery obtained and cognitive function recovery , using the formula ; Calculate the motion priority index ;in, and are preset positive integers, and ; To control the motion priority index Constant of decay rate; cognitive priority index ; Then use the formula ; Calculate the first The weight of motor function recovery in training update cycles ;No. Cognitive function recovery weight of training update cycle .

[0020] Calculate and obtain the actual training volume and minimum target training volume of the patient's current motor function training. Based on the actual training volume and minimum target training volume, adjust the motor function training time of the patient's current motor function training. The specific operations are as follows: For any sports function training in the current training update cycle, obtain the action standard degree of the current sports function training ; Set minimum standards ; The standard of the current training action Below the minimum standard When the training intensity in the sports function training plan based on the current training update cycle is and training time , using the formula Calculate and obtain the actual training volume of the current sports function training ; Using the formula Calculate the minimum target training volume ; Using the formula Calculate the additional training time required for the patient's current motor function training .

[0021] The patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained, and the patient's discomfort level in the current motor function training is calculated; if the patient's discomfort level is higher than the limit threshold, the motor function training intensity of the patient's next motor function training is adjusted. The specific operations are as follows: During any exercise function training within the current training update cycle, the patient's heart rate, respiratory rate and blood pressure are monitored, and the patient's highest heart rate during the current exercise function training is obtained , maximum respiratory rate and maximum blood pressure ; Using the formula ; Calculate the heart rate excess value ,in represents the patient's heart rate at rest. Represents the expected maximum heart rate of the patient; using the formula ; Calculate the respiratory rate excess value ,in Indicates the patient's respiratory rate at rest. Represents the expected maximum respiratory rate of the patient; using the formula ; Calculate the blood pressure excess value ,in It represents the patient's blood pressure at rest. Indicates the preset maximum expected blood pressure value of the patient; uses the digital scoring method to obtain the patient's pain score , The value range is 0-10, using the formula Calculate the patient's pain threshold ; Using the formula Calculate the patient's discomfort level during current motor function rehabilitation training ; Setting the Limit Threshold , when the patient is at a level of discomfort in the current motor function rehabilitation training Above limit threshold When , use the formula Calculate the appropriate training intensity for the next motor function rehabilitation training ; To control the intensity of adaptive training The control coefficient of the reduction can be set and adjusted by professionals.

[0022] The training plan generation model is established based on the multi-task learning model, including: a first input layer, a shared layer and a task-specific layer; the first input layer is used to receive the patient's basic information, motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight; the shared layer is used to process all input features received by the input layer, extract the relationship between the input features, and generate a comprehensive feature representation of the patient; the task-specific layer includes a motor function training plan output layer and a cognitive function training plan output layer. The motor function training plan output layer is used to receive the patient's comprehensive feature representation and motor function recovery weight, and output the patient's motor function training plan; the cognitive function training plan output layer is used to receive the patient's comprehensive feature representation and cognitive function recovery weight, and output the cognitive function training plan.

[0023] The specific operations for training the model generated by the training plan are as follows: Acquire several plan generation training samples with labeled motor function training plans and cognitive function training plans, each plan generation training sample contains a patient's basic information, motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight; divide all acquired plan generation training samples into a plan generation training set and a plan generation verification set; use the plan generation training set to train the training plan generation model, and verify the training plan generation model through the plan generation verification set to obtain a first verification result; set a first training condition, and determine whether the acquired first verification result meets the first training condition, and if so, output the trained training plan generation model; if not, continue to use the plan generation training set to train the training plan generation model.

[0024] The action evaluation model is established based on CNN and LSTM, including the second input layer, CNN layer, LSTM layer, fully connected layer and output layer; the second input layer is used to receive patient training videos and standard videos; the CNN layer is used to extract spatial features from the images of each frame of the patient training videos and standard videos, and output the spatial features of each frame; the LSTM layer is used to extract the temporal features of the patient training videos and standard videos; the fully connected layer is used to integrate the temporal features with the spatial features, and generate a scalar through the ReLU activation function; the output layer is used to output the scalar generated by the fully connected layer, which represents the action standard of the patient training video; For the training of the action evaluation model, the specific operations are as follows: Acquire several motion evaluation training samples with labeled motion standards, each of which contains a patient training video and a corresponding standard video; divide all motion evaluation training samples into a motion evaluation training set and a motion evaluation verification set, train the motion evaluation model with the motion evaluation training set, and verify the motion evaluation model with the motion evaluation verification set to obtain a second verification result; set a second training condition, and determine whether the obtained second verification result meets the second training condition; if so, output the trained motion evaluation model; if not, continue to train the motion evaluation model with the motion evaluation training set.

[0025] Example 2, a stroke patient rehabilitation training management system based on an Internet platform, such as Figure 1 As shown, including: The training plan generation module includes a priority weight calculation unit and a training plan generation unit; the priority weight calculation unit is used to obtain the patient's current motor function recovery degree and cognitive function recovery degree at the beginning of any training update cycle, and then calculate the patient's motor priority index and cognitive priority index, and calculate the patient's motor function recovery weight and cognitive function recovery weight based on the obtained motor priority index and cognitive priority index; the training plan generation unit is used to obtain the patient's basic information, the basic information includes the patient's age and gender, and use the patient's basic information, current motor function recovery progress, cognitive function recovery progress, and motor function recovery weight and cognitive function recovery weight as inputs to the training plan generation model, and output the patient's motor function training plan and cognitive function training plan in the current training update cycle; A training time dynamic adjustment module includes a motion evaluation unit and a training time adjustment unit; the motion evaluation unit is used for obtaining a patient training video using a camera during any motor function training of the patient within the current training update cycle, inputting the patient training video and the corresponding standard learning video into a motion evaluation model, and obtaining the motion standard of the patient's current motor function training; the training time adjustment unit is used for calculating and obtaining the actual training amount and the minimum target training amount of the patient's current motor function training when the motion standard is lower than a preset minimum standard, and adjusting the motor function training time of the patient's current motor function training based on the actual training amount and the minimum target training amount; The training intensity dynamic adjustment module includes a discomfort degree calculation unit and a training intensity adjustment unit; the discomfort degree calculation unit is used to obtain the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level for any motor function training of the patient within the current training update cycle, and calculate the patient's discomfort level in the current motor function training; the training intensity adjustment unit is used to adjust the motor function training intensity of the patient's next motor function training when the patient's discomfort level is higher than a preset limit threshold.

[0026] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for the management of rehabilitation training for stroke patients based on an Internet platform, characterized in that: include: After the patient starts rehabilitation training, a training update cycle is set. At the beginning of any training update cycle, the patient's current motor function recovery degree and cognitive function recovery degree are obtained, and then the patient's motor priority index and cognitive priority index are calculated. Based on the obtained motor priority index and cognitive priority index, the patient's motor function recovery weight and cognitive function recovery weight are calculated; the patient's basic information, including the patient's age and gender, is obtained, and the patient's basic information, current motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight are used as inputs to the training plan generation model, and the patient's motor function training plan and cognitive function training plan in the current training update cycle are output. The motor function training plan includes motor function training intensity, motor function training time and motor function training times; The cognitive function training plan includes cognitive function training difficulty, cognitive function training time and cognitive function training frequency; Patients receive rehabilitation training according to the motor function training plan and cognitive function training plan during the current training update cycle; For any motor function training of the patient in the current training update cycle, use a camera to obtain the patient's training video during the motor function training, input the patient's training video and the corresponding standard learning video into the action evaluation model, and obtain the action standard degree of the patient's current motor function training; Setting a minimum standard degree, when the movement standard degree is lower than the preset minimum standard degree, calculating and obtaining the actual training amount and the minimum target training amount of the patient's current motor function training, and adjusting the motor function training time of the patient's current motor function training based on the obtained actual training amount and the minimum target training amount; At the same time, for any motor function training of the patient within the current training update cycle, the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained, and the patient's discomfort level in the current motor function training is calculated; if the patient's discomfort level is higher than the preset limit threshold, the motor function training intensity of the patient's next motor function training will be adjusted.

2. The method for managing rehabilitation training for stroke patients based on the Internet platform according to claim 1, characterized in that: The patient's current motor function recovery degree and cognitive function recovery degree are obtained, and then the patient's motor priority index and cognitive priority index are calculated. Based on the obtained motor priority index and cognitive priority index, the patient's motor function recovery weight and cognitive function recovery weight are calculated. The specific operations are as follows: At the beginning of any training update cycle, the Fugl-Meyer evaluation form was used to evaluate the patient's motor function, and the ratio of the motor function evaluation result to the full score of the Fugl-Meyer evaluation form (226) was calculated to obtain the patient's motor function after starting rehabilitation training. The degree of recovery of motor function in each training update cycle At the same time, the MoCA scale was used to evaluate the patient's cognitive function, and the ratio of the cognitive function evaluation result to the full score of the MoCA scale of 30 was calculated to obtain the patient's cognitive function after starting rehabilitation training. The degree of cognitive function recovery after training update cycles ; Based on the degree of motor function recovery obtained and cognitive function recovery , using the formula ； Calculate the motion priority index ;in, and are preset positive integers, and ; To control the motion priority index Constant of decay rate; cognitive priority index ; Then use the formula ; Calculate the first The weight of motor function recovery in training update cycles ;No. Cognitive function recovery weight of training update cycle .

3. The method for managing rehabilitation training for stroke patients based on the Internet platform according to claim 2, characterized in that: Calculate and obtain the actual training volume and minimum target training volume of the patient's current motor function training. Based on the actual training volume and minimum target training volume, adjust the motor function training time of the patient's current motor function training. The specific operations are as follows: For any sports function training in the current training update cycle, obtain the action standard degree of the current sports function training ; Set minimum standards ; The standard of the current training action Below the minimum standard When the training intensity in the sports function training plan based on the current training update cycle is and training time , using the formula Calculate and obtain the actual training volume of the current sports function training ; Using the formula Calculate the minimum target training volume ; Using the formula Calculate the additional training time required for the patient's current motor function training .

4. The Internet-based stroke patient rehabilitation training management method according to claim 3 is characterized in that: The patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level are obtained, and the patient's discomfort level in the current motor function training is calculated; if the patient's discomfort level is higher than the limit threshold, the motor function training intensity of the patient's next motor function training is adjusted. The specific operations are as follows: During any exercise function training within the current training update cycle, the patient's heart rate, respiratory rate and blood pressure are monitored, and the patient's highest heart rate during the current exercise function training is obtained , maximum respiratory rate and maximum blood pressure ; Using the formula Calculate the heart rate excess value ,in represents the patient's heart rate at rest. Indicates the preset maximum expected heart rate of the patient; Using the formula ; Calculate the respiratory rate excess value ,in Indicates the patient's respiratory rate at rest. Represents the expected maximum respiratory rate of the patient; using the formula Calculate the blood pressure excess value ,in It represents the patient's blood pressure at rest. Indicates the preset maximum expected blood pressure value of the patient; uses the digital scoring method to obtain the patient's pain score , The value range is 0-10, using the formula Calculate the patient's pain threshold ; Using the formula Calculate the patient's discomfort level during current motor function rehabilitation training ; Setting the Limit Threshold , when the patient is at a level of discomfort in the current motor function rehabilitation training Above limit threshold When , use the formula Calculate the appropriate training intensity for the next motor function rehabilitation training ; To control the intensity of adaptive training The control coefficient of the decline.

5. The Internet-based stroke patient rehabilitation training management method according to claim 4 is characterized in that: The training plan generation model is established based on the multi-task learning model, including: a first input layer, a shared layer and a task-specific layer; the first input layer is used to receive the patient's basic information, motor function recovery progress, cognitive function recovery progress, motor function recovery weight and cognitive function recovery weight; the shared layer is used to process all input features received by the input layer, extract the relationship between the input features, and generate a comprehensive feature representation of the patient; the task-specific layer includes a motor function training plan output layer and a cognitive function training plan output layer. The motor function training plan output layer is used to receive the patient's comprehensive feature representation and motor function recovery weight, and output the patient's motor function training plan; the cognitive function training plan output layer is used to receive the patient's comprehensive feature representation and cognitive function recovery weight, and output the cognitive function training plan.

6. The Internet-based stroke patient rehabilitation training management method according to claim 5, characterized in that: The specific operations for training the model generated by the training plan are as follows: Obtain a number of plan generation training samples with labeled motor function training plans and cognitive function training plans, each plan generation training sample contains a patient's basic information, motor function recovery progress, cognitive function recovery progress, motor function recovery weight, and cognitive function recovery weight; divide all the acquired plan generation training samples into a plan generation training set and a plan generation verification set; Using the plan generation training set to train the training plan generation model, and using the plan generation verification set to verify the training plan generation model, to obtain a first verification result; A first training condition is set to determine whether the first verification result obtained satisfies the first training condition. If so, the trained training plan generation model is output; if not, the training plan generation model is continuously trained by applying the plan generation training set.

7. The Internet-based rehabilitation training management method for stroke patients according to claim 6 is characterized in that: The action evaluation model is built based on CNN and LSTM, and includes a second input layer, a CNN layer, an LSTM layer, a fully connected layer, and an output layer; the second input layer is used to receive patient training videos and standard videos; the CNN layer is used to extract spatial features from the images of each frame of the patient training videos and standard videos, and output the spatial features of each frame; The LSTM layer is used to extract temporal features of patient training videos and standard videos; The fully connected layer is used to integrate the temporal features with the spatial features and generate a scalar through the ReLU activation function; the output layer is used to output the scalar generated by the fully connected layer to represent the action standard of the patient training video; For the training of the action evaluation model, the specific operations are as follows: Obtain a number of motion evaluation training samples with marked motion standards, each of which contains a patient training video and a corresponding standard video; divide all motion evaluation training samples into a motion evaluation training set and a motion evaluation verification set, use the motion evaluation training set to train the motion evaluation model, and verify the motion evaluation model through the motion evaluation verification set to obtain a second verification result; A second training condition is set to determine whether the second verification result obtained satisfies the second training condition. If so, the trained action evaluation model is output; if not, the action evaluation training set is continued to be applied to train the action evaluation model.

8. A stroke patient rehabilitation training management system based on the Internet platform, characterized in that: The system is applied to the Internet platform-based stroke patient rehabilitation training management method according to any one of claims 1 to 7, comprising: The training plan generation module includes a priority weight calculation unit and a training plan generation unit; the priority weight calculation unit is used to obtain the patient's current motor function recovery degree and cognitive function recovery degree at the beginning of any training update cycle, and then calculate the patient's motor priority index and cognitive priority index, and calculate the patient's motor function recovery weight and cognitive function recovery weight based on the obtained motor priority index and cognitive priority index; the training plan generation unit is used to obtain the patient's basic information, the basic information includes the patient's age and gender, and use the patient's basic information, current motor function recovery progress, cognitive function recovery progress, and motor function recovery weight and cognitive function recovery weight as inputs to the training plan generation model, and output the patient's motor function training plan and cognitive function training plan in the current training update cycle; A training time dynamic adjustment module includes a motion evaluation unit and a training time adjustment unit; the motion evaluation unit is used for obtaining a patient training video using a camera during any motor function training of the patient within the current training update cycle, inputting the patient training video and the corresponding standard learning video into a motion evaluation model, and obtaining the motion standard of the patient's current motor function training; the training time adjustment unit is used for calculating and obtaining the actual training amount and the minimum target training amount of the patient's current motor function training when the motion standard is lower than a preset minimum standard, and adjusting the motor function training time of the patient's current motor function training based on the actual training amount and the minimum target training amount; The training intensity dynamic adjustment module includes a discomfort degree calculation unit and a training intensity adjustment unit; the discomfort degree calculation unit is used to obtain the patient's highest heart rate, highest respiratory rate, highest blood pressure and pain level for any motor function training of the patient within the current training update cycle, and calculate the patient's discomfort level in the current motor function training; the training intensity adjustment unit is used to adjust the motor function training intensity of the patient's next motor function training when the patient's discomfort level is higher than a preset limit threshold.

Images