Intelligent auxiliary system and method for cardiovascular and cerebrovascular rehabilitation

Abstract

The invention provides an intelligent auxiliary system and method for cardiovascular and cerebrovascular rehabilitation, and relates to the technical field of medical treatment. And the controller inputs the second fused heart and brain features into a pre-trained scoring model to obtain a second scoring value of the operation state of the heart and cerebral vessels of the patient. When the second score value is smaller than or equal to the first score value, or the second score value is larger than the first score value, but the difference value between the second score value and the first score value is lower than a preset difference value threshold value, the controller indicates that the patient does a preset exercise period through the previous first rehabilitation exercise strategy. The disease degree of the heart and cerebral vessels is not improved, so that the first rehabilitation exercise strategy can be updated by utilizing a reinforcement learning model according to the second heart and cerebral operation characteristics and the second score value; therefore, understandingly, after the first rehabilitation exercise strategy obtained through the updating is moved by the patient for the preset exercise period, the disease degree of heart and cerebral vessels of the patient can be obviously improved, and the reliability is high.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular to an intelligent auxiliary system and method for cardiovascular and cerebrovascular rehabilitation. Background Art

[0002] Cardiovascular and cerebrovascular accidents are a common disease that seriously threatens the health of humans, especially the elderly over 50 years old. Even with the most advanced and complete treatment methods, many survivors of cardiovascular and cerebrovascular accidents have mobility problems, so they need to rely on persistent exercise to restore physiological muscle function.

[0003] Currently, patients only do some exercises on their own or follow simple doctor's instructions to exercise, which has poor reliability in restoring physiological muscle energy. Summary of the invention

[0004] The present application provides an intelligent auxiliary system and method for cardiovascular and cerebrovascular rehabilitation, which is used to solve the problem in the prior art that patients only perform some exercises on their own or follow simple doctor's instructions to exercise, and the reliability of physiological muscle recovery is poor.

[0005] In the first aspect, the present application provides a cardiovascular and cerebrovascular rehabilitation intelligent assistance method, which is applied to a cardiovascular and cerebrovascular rehabilitation intelligent assistance system. The cardiovascular and cerebrovascular rehabilitation intelligent assistance system includes a controller, a human-computer interaction screen, and a wearable component. The controller is electrically connected to the human-computer interaction screen and the wearable component, respectively. The method includes:

[0006] The controller receives the patient attribute characteristics entered by the patient on the human-computer interaction screen, wherein the patient attribute characteristics at least include the patient's age information, gender information, height information and weight information;

[0007] The controller receives the first heart and brain operation characteristics of the patient collected by the wearable component, wherein the first heart and brain operation characteristics at least include the patient's heart rate waveform, blood pressure waveform and brain wave;

[0008] The controller obtains a first fused heart-brain feature according to the patient's attribute feature and the first heart-brain operation feature;

[0009] The controller inputs the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operation status of the patient, wherein the scoring model is a training sample set constructed based on a plurality of historical fused cardio-cerebral features and corresponding historical scoring values, and is input into a first neural network for training;

[0010] The controller inputs the first fused heart-brain feature into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, wherein the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on a plurality of historical fused heart-brain features and corresponding historical rehabilitation exercise strategies;

[0011] After a preset exercise cycle, the controller receives the second heart and brain operation characteristics collected by the wearable component from the patient, wherein the second heart and brain operation characteristics at least include the patient's heart rate waveform, blood pressure waveform and brain wave;

[0012] The controller obtains a second fused heart-brain feature according to the patient's attribute feature and the second heart-brain operation feature;

[0013] The controller inputs the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient;

[0014] When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than a preset difference threshold, the controller updates the first rehabilitation exercise strategy according to the second heart and brain operation characteristics and the second score value using the reinforcement learning model; and returns to execute the step in which the controller receives the second heart and brain operation characteristics collected by the wearable component on the patient after the preset exercise cycle, until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than the preset difference threshold.

[0015] In a possible implementation, after the controller receives the second heart and brain operation characteristic of the patient collected by the wearable component after the preset exercise cycle, the method further includes:

[0016] The controller receives medical record information transmitted by the patient terminal, the medical record information including diagnosis results associated with cardiovascular and cerebrovascular diseases;

[0017] The controller inputs the diagnosis result associated with the cardiovascular and cerebrovascular disease into the pre-trained second rehabilitation exercise strategy determination model to obtain the second rehabilitation exercise strategy within the preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is constructed based on a training sample set constructed based on a plurality of historical diagnosis results associated with the cardiovascular and cerebrovascular disease and the corresponding historical rehabilitation exercise strategies, and is input into the second neural network for training;

[0018] The controller modifies the first rehabilitation exercise strategy according to the second rehabilitation exercise strategy.

[0019] In a possible implementation, the first rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, the second rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, and the controller modifies the first rehabilitation exercise strategy according to the second rehabilitation exercise strategy, including:

[0020] The controller adds the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy to the first rehabilitation exercise strategy;

[0021] The controller averages the exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise intensity, and uses the average exercise intensity to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy.

[0022] In a possible implementation, the exercise intensity includes the exercise frequency and the exercise duration. The controller averages the exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy and the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain the average exercise intensity. The average exercise intensity is used to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy, including:

[0023] The controller averages the movement frequency of the movement action belonging to the second rehabilitation movement strategy in the first rehabilitation movement strategy and the movement frequency of the movement action in the second rehabilitation movement strategy to obtain an average movement frequency, and updates the movement frequency of the movement action in the first rehabilitation movement strategy using the average movement frequency;

[0024] The controller averages the exercise duration of the exercise action in the first rehabilitation exercise strategy belonging to the second rehabilitation exercise strategy and the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise duration, and uses the average exercise duration to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

[0025] In a possible implementation, the cardiovascular and cerebrovascular rehabilitation intelligent assistance system includes a dedicated exercise area and a camera facing the dedicated exercise area. After the controller inputs the first fused cardiovascular and brain features into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, the method further includes:

[0026] The controller plays a first exercise video associated with a first rehabilitation exercise strategy on the human-computer interaction screen, wherein the first exercise video includes a plurality of first exercise action sub-videos in a sequential order;

[0027] The controller receives the second motion action sub-videos of the patient in the dedicated motion area captured by the camera, and determines the order of the second motion action sub-videos;

[0028] The controller determines the similarity of the ordered second movement action sub-video to the ordered first movement action sub-video in the first rehabilitation movement strategy;

[0029] When the similarity is lower than the set similarity threshold, the controller outputs motion correction prompt information to the human-computer interaction screen.

[0030] In a second aspect, the present application also provides a cardiovascular and cerebrovascular rehabilitation intelligent assistance system, which includes a controller, a human-computer interaction screen, and a wearable component. The controller is electrically connected to the human-computer interaction screen and the wearable component, respectively, and the controller is used to:

[0031] Receiving patient attribute characteristics entered by the patient on the human-computer interaction screen, wherein the patient attribute characteristics at least include patient age information, gender information, height information and weight information;

[0032] Receiving a first heart and brain operation characteristic of the patient collected by the wearable component, wherein the first heart and brain operation characteristic at least includes a heart rate waveform, a blood pressure waveform, and an electroencephalogram of the patient;

[0033] According to the patient's attribute characteristics and the first heart-brain operation characteristics, a first fusion heart-brain characteristic is obtained;

[0034] Inputting the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operation status of the patient, wherein the scoring model is a training sample set constructed based on a plurality of historical fused cardio-cerebral features and corresponding historical scoring values, and inputted into a first neural network for training;

[0035] Inputting the first fused heart-brain feature into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, wherein the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on a plurality of historical fused heart-brain features and corresponding historical rehabilitation exercise strategies;

[0036] After a preset exercise cycle, receiving a second heart and brain operation characteristic of the patient collected by the wearable component, wherein the second heart and brain operation characteristic at least includes a heart rate waveform, a blood pressure waveform and an electroencephalogram of the patient;

[0037] According to the patient's attribute characteristics and the second heart-brain operation characteristics, a second fused heart-brain characteristic is obtained;

[0038] Inputting the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient;

[0039] When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than the preset difference threshold, the reinforcement learning model is used to update the first rehabilitation exercise strategy according to the second heart and brain operation characteristics and the second score value; and the step of returning to execute after the preset exercise cycle, the controller receives the second heart and brain operation characteristics collected by the wearable component on the patient until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than the preset difference threshold.

[0040] In a possible implementation manner, the controller is further configured to:

[0041] Receiving medical record information transmitted by a patient terminal, the medical record information including diagnosis results associated with cardiovascular and cerebrovascular diseases;

[0042] Inputting the diagnosis result associated with the cardiovascular and cerebrovascular disease into the pre-trained second rehabilitation exercise strategy determination model to obtain the second rehabilitation exercise strategy within the preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is constructed based on a training sample set constructed based on a plurality of historical diagnosis results associated with the cardiovascular and cerebrovascular disease and the corresponding historical rehabilitation exercise strategies, and inputting the training sample set into the second neural network;

[0043] According to the second rehabilitation exercise strategy, the first rehabilitation exercise strategy is revised.

[0044] In a possible implementation, the first rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, the second rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, and the controller is specifically used to

[0045] Adding the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy to the first rehabilitation exercise strategy;

[0046] The exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy is averaged with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain the average exercise intensity, and the average exercise intensity is used to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy.

[0047] In a possible implementation, the exercise intensity includes the exercise frequency and the exercise duration, and the controller is specifically configured to:

[0048] averaging the movement frequency of the movement action in the first rehabilitation movement strategy that belongs to the second rehabilitation movement strategy and the movement frequency of the movement action in the second rehabilitation movement strategy to obtain an average movement frequency, and using the average movement frequency to update the movement frequency of the movement action in the first rehabilitation movement strategy;

[0049] The exercise duration of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy is averaged with the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain the average exercise duration, and the average exercise duration is used to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

[0050] In a possible implementation, the intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation further includes a dedicated motion area and a camera facing the dedicated motion area, and the controller is specifically used for:

[0051] Playing a first exercise video associated with a first rehabilitation exercise strategy on a human-computer interaction screen, the first exercise video including a plurality of first exercise action sub-videos in a sequential order;

[0052] receiving a second motion sub-video of the patient in the dedicated motion area captured by a camera, and determining an order of the second motion sub-videos;

[0053] Determining the similarity of the ordered second movement action sub-video to the ordered first movement action sub-video in the first rehabilitation movement strategy;

[0054] When the similarity is lower than the set similarity threshold, the motion correction prompt information is output to the human-computer interaction screen.

[0055] The present application provides a cardiovascular rehabilitation intelligent assistance system and method, wherein a controller obtains a first fused cardiovascular feature according to a patient's attribute characteristics and a first cardiovascular operation feature. It can be understood that the first fused cardiovascular feature can more accurately describe the cardiovascular operation of the patient before the first rehabilitation exercise strategy is executed.

[0056] The controller inputs the first fused cardio-cerebral feature into the pre-trained scoring model to obtain a first scoring value for the patient's cardiovascular and cerebrovascular operating state, wherein the first scoring value can reflect the degree of cardiovascular and cerebrovascular disease of the patient before the first rehabilitation exercise strategy is executed.

[0057] The controller inputs the first fused heart-brain feature into the pre-trained first rehabilitation exercise strategy determination model to obtain the first rehabilitation exercise strategy within the preset exercise cycle; it can be understood that the first rehabilitation exercise strategy obtained based on the first fused heart-brain feature has high reliability. At this time, the patient can exercise the preset exercise cycle based on the first rehabilitation exercise strategy.

[0058] After a preset exercise cycle, the controller receives the second cardio-cerebral operation characteristic collected by the wearable component from the patient; the controller obtains a second fused cardio-cerebral characteristic based on the patient's attribute characteristics and the second cardio-cerebral operation characteristic. It can be understood that the second fused cardio-cerebral characteristic can more accurately describe the cardiovascular and cerebrovascular operation status of the patient after executing the first rehabilitation exercise strategy.

[0059] The controller inputs the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the patient's cardiovascular and cerebrovascular operating state, wherein the second scoring value can reflect the degree of cardiovascular and cerebrovascular disease of the patient after executing the first rehabilitation exercise strategy.

[0060] When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than the preset difference threshold, it means that the patient's cardiovascular and cerebrovascular disease severity has not been improved after the preset exercise cycle of the first rehabilitation exercise strategy. Therefore, the first rehabilitation exercise strategy can be updated according to the second cardiovascular and cerebrovascular operation characteristics and the second score value using the reinforcement learning model; and the controller returns to the step of receiving the second cardiovascular and cerebrovascular operation characteristics collected by the wearable component after the preset exercise cycle, until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than the preset difference threshold. It can be understood that the first rehabilitation exercise strategy finally updated can help the patient's cardiovascular and cerebrovascular disease severity to be significantly improved after the patient exercises the preset exercise cycle, and has high reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0062] Figure 1 A schematic diagram of the structure of the intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation provided in this application;

[0063] Figure 2 This is a flow chart of the intelligent assistance method for cardiovascular and cerebrovascular rehabilitation provided in this application. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments made by ordinary technicians in this field under the enlightenment of the embodiments belong to the scope of protection of the present application.

[0065] The embodiment of the present application provides an intelligent assistance method for cardiovascular and cerebrovascular rehabilitation, which is applied to an intelligent assistance system for cardiovascular and cerebrovascular rehabilitation. Figure 1 As shown, the intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation includes a controller (not shown in the figure), a human-computer interaction screen and a wearable component (not shown in the figure), and the controller is electrically connected to the human-computer interaction screen and the wearable component respectively. Figure 2 As shown, the method provided in the embodiment of the present application includes:

[0066] S201: The controller receives the patient attribute characteristics entered by the patient on the human-computer interaction screen.

[0067] The patient attribute characteristics include at least the patient's age information, gender information, height information and weight information.

[0068] S202: The controller receives the first heart and brain operation characteristics of the patient collected by the wearable component.

[0069] The first heart and brain operation characteristic at least includes the patient's heart rate waveform, blood pressure waveform and brain wave. The wearable component may include a smart bracelet and a smart headgear.

[0070] S203: The controller obtains a first fused heart-brain feature according to the patient's attribute characteristics and the first heart-brain operation feature.

[0071] For example, the controller converts the patient attribute feature into a first feature vector, converts the first heart and brain operation feature into a second feature vector, and concatenates the first feature vector and the second feature vector to obtain a first fused heart and brain feature.

[0072] S204: The controller inputs the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operating status of the patient.

[0073] Among them, the scoring model is obtained by inputting a training sample set constructed based on multiple historical fusion heart and brain features and their corresponding historical scoring values ​​into the first neural network for training.

[0074] S205: The controller inputs the first fused heart-brain feature into the pre-trained first rehabilitation exercise strategy determination model to obtain the first rehabilitation exercise strategy within a preset exercise cycle.

[0075] Among them, the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on multiple historical fusion heart and brain characteristics and the corresponding historical rehabilitation exercise strategies.

[0076] For example, the preset exercise cycle can be one week, half a month, or one month, etc., which is not limited here. The first rehabilitation exercise strategy can be to do A action at frequency A for A, do B action at frequency B for B, do C action at frequency C for C, etc., which is not limited here.

[0077] Preferably, the first fused heart-brain feature can be obtained according to the patient attribute feature and the first heart-brain operation feature according to the following process:

[0078] S2031: Obtaining patient attribute characteristics of multiple sample patients, a first cardio-cerebral operation characteristic, and a cardio-cerebral vascular operation status value of the sample patient determined according to the first cardio-cerebral operation characteristic. For example, the cardio-cerebral vascular operation status value of the sample patient may be determined manually (e.g., by a doctor, an expert, etc.) or by a machine learning model according to the first cardio-cerebral operation characteristic;

[0079] S2032: dividing the multiple sample patients into multiple sample patient clusters according to the patient attribute characteristics of the multiple sample patients by using a K-means clustering algorithm;

[0080] S2033: for each sample patient, performing variational modal decomposition on the sample patient's heart rate waveform to obtain a plurality of heart rate waveform components corresponding to the sample patient, performing variational modal decomposition on the sample patient's blood pressure waveform to obtain a plurality of blood pressure waveform components corresponding to the sample patient, and performing variational modal decomposition on the sample patient's brain wave to obtain a plurality of brain wave components corresponding to the sample patient;

[0081] S2034: for each sample patient cluster, according to the center frequencies and limited bandwidths of the multiple heart rate waveform components of each sample patient included in the sample patient cluster, the center frequencies and limited bandwidths of the blood pressure waveform components, and the center frequencies and limited bandwidths of the brain wave components, the center frequency range of the heart rate waveform components and the limited bandwidth range of the heart rate waveform components, the center frequency range of the blood pressure waveform components and the limited bandwidth range of the blood pressure waveform components, and the center frequency range of the brain wave components and the limited bandwidth range of the brain wave components corresponding to the sample patient cluster are determined. For example, according to the center frequencies and limited bandwidths of the multiple heart rate waveform components of each sample patient included in the sample patient cluster, the maximum value and the minimum value of the center frequency of the heart rate waveform components and the maximum value and the minimum value of the limited bandwidth are determined, and then the center frequency range of the heart rate waveform components and the limited bandwidth range of the heart rate waveform components are determined. The center frequency range of the blood pressure waveform components and the limited bandwidth range of the blood pressure waveform components, and the center frequency range of the brain wave components and the limited bandwidth range of the brain wave components are determined in a manner similar to the manner of determining the center frequency range of the heart rate waveform components and the limited bandwidth range of the heart rate waveform components, and are not described in detail here.

[0082] S2035: For each sample patient cluster, divide the center frequency range of the heart rate waveform component corresponding to the sample patient cluster to determine multiple center frequency sub-ranges of the heart rate waveform component corresponding to the sample patient cluster, divide the limited bandwidth range of the heart rate waveform component corresponding to the sample patient cluster to determine multiple limited bandwidth sub-ranges of the heart rate waveform component corresponding to the sample patient cluster, determine the correlation coefficient between the center frequency sub-range of each heart rate waveform component and the sample patient cluster and the correlation coefficient between the limited bandwidth sub-range of each heart rate waveform component and the sample patient cluster according to the cardiovascular and cerebrovascular operation status value of each sample patient included in the sample patient cluster, determine the target center frequency sub-range of the heart rate waveform corresponding to the sample patient cluster from the multiple center frequency sub-ranges of the corresponding heart rate waveform components, and determine the target center frequency sub-range of the heart rate waveform corresponding to the sample patient cluster from the multiple center frequency sub-ranges of the corresponding heart rate waveform components;

[0083] S2036: For each sample patient cluster, referring to the method of S2035, determine the target center frequency sub-range of the blood pressure waveform and the corresponding target center frequency sub-range of the blood pressure waveform corresponding to the sample patient cluster, and determine the target center frequency sub-range of the brain wave and the corresponding target center frequency sub-range of the brain wave corresponding to the sample patient cluster;

[0084] S2037: Determine a similar sample patient cluster from multiple sample patient clusters according to the patient attribute characteristics entered by the patient on the human-computer interaction screen, for example, calculate the similarity between the patient attribute characteristics and the patient attribute characteristics of the sample patient corresponding to the cluster center of each sample patient cluster, and take the sample patient cluster to which the cluster center with the greatest similarity belongs as the similar sample patient cluster;

[0085] S2038: performing variational modal decomposition on the patient's heart rate waveform to obtain multiple heart rate waveform components corresponding to the patient, and selecting heart rate waveform components whose center frequencies are within the target center frequency subrange and whose center frequencies are within the target center frequency subrange from the multiple heart rate waveform components corresponding to the patient as effective heart rate waveform components according to the target center frequency subrange and the target center frequency subrange of the heart rate waveform corresponding to the similar sample patient cluster, performing variational modal decomposition on the patient's blood pressure waveform to obtain multiple blood pressure waveform components corresponding to the patient, and selecting blood pressure waveform components whose center frequencies are within the target center frequency subrange and whose center frequencies are within the target center frequency subrange from the multiple blood pressure waveform components corresponding to the patient as effective blood pressure waveform components according to the target center frequency subrange and the target center frequency subrange of the blood pressure waveform corresponding to the similar sample patient cluster, performing variational modal decomposition on the patient's brain wave to obtain multiple brain wave components corresponding to the patient, and selecting brain wave components whose center frequencies are within the target center frequency subrange and whose center frequencies are within the target center frequency subrange from the multiple brain wave components corresponding to the patient as effective brain wave components according to the target center frequency subrange and the target center frequency subrange of the brain wave corresponding to the similar sample patient cluster;

[0086] S2039: Convert the patient attribute characteristics entered by the patient on the human-computer interaction screen into a first characteristic vector, convert the effective heart rate waveform component, the effective blood pressure waveform component and the effective brain wave component into a second characteristic vector, and concatenate the first characteristic vector and the second characteristic vector to obtain a first fused heart-brain characteristic.

[0087] Specifically, the correlation coefficient between the central frequency sub-range of the heart rate waveform component and the sample patient cluster can be calculated according to the following formula:

[0088]

[0089] Among them, r ((i,j),frequency) is the correlation coefficient between the ith central frequency sub-range corresponding to the ith sample patient cluster and the ith sample patient cluster, N is the total number of valid sample patients in the ith central frequency sub-range corresponding to the ith sample patient cluster, and the valid sample patients are sample patients in the ith sample patient cluster whose central frequency of at least one heart rate waveform component is located in the ith central frequency sub-range corresponding to the ith sample patient cluster, F n is the component characteristics (e.g., amplitude, energy, mean, etc.) of the heart rate waveform component of the nth valid sample patient whose center frequency is located in the ith center frequency subrange corresponding to the ith sample patient cluster, V n is the cardiovascular and cerebrovascular operating status value of the nth valid sample patient.

[0090] The similarity between the patient attribute characteristics of the patient and the patient attribute characteristics of the sample patient corresponding to the cluster center of the sample patient cluster can be calculated according to the following formula:

[0091]

[0092] Among them, S (present,k) is the similarity between the patient attribute characteristics of the patient and the patient attribute characteristics of the sample patient corresponding to the cluster center of the kth sample patient cluster, P1 is a preset parameter, P1 is greater than 0, M is the number of types of information included in the patient attribute characteristics, F (present,m) is the mth information included in the patient attribute characteristics of the patient (for example, one of the patient's age information, gender information, height information, and weight information, etc.), F (k,m) The mth type of information included in the patient attribute characteristics of the sample patient corresponding to the cluster center of the kth sample patient cluster.

[0093] Optionally, after S205, the method in the embodiment of the present application further includes:

[0094] Step 1: The controller receives medical record information transmitted by the patient terminal, where the medical record information includes diagnosis results associated with cardiovascular and cerebrovascular diseases.

[0095] Step 2: The controller inputs the diagnosis results associated with cardiovascular and cerebrovascular diseases into the pre-trained second rehabilitation exercise strategy determination model to obtain a second rehabilitation exercise strategy within a preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is constructed based on a training sample set constructed based on multiple historical diagnosis results associated with cardiovascular and cerebrovascular diseases and the corresponding historical rehabilitation exercise strategies, and the training sample set is input into the second neural network for training.

[0096] It can be understood that the second rehabilitation exercise strategy determined within the preset exercise cycle based on the diagnosis result associated with cardiovascular and cerebrovascular diseases has high reliability.

[0097] Step 3: The controller modifies the first rehabilitation exercise strategy according to the second rehabilitation exercise strategy.

[0098] Based on the above steps 1 to 3, the first rehabilitation exercise strategy obtained has higher reliability.

[0099] S206: After a preset exercise cycle, the controller receives the second heart and brain operation characteristics of the patient collected by the wearable component.

[0100] The second heart and brain operation characteristic includes at least the patient's heart rate waveform, blood pressure waveform and brain wave.

[0101] S207: The controller obtains a second fused heart-brain feature according to the patient's attribute feature and the second heart-brain operation feature.

[0102] S208: The controller inputs the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient.

[0103] S209: The controller determines whether the second score value is less than or equal to the first score value, or whether the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than a preset difference threshold. If yes, execute S210; if not, execute S211.

[0104] S210: using the reinforcement learning model to update the first rehabilitation exercise strategy according to the second heart and brain operation characteristics and the second score value; and returning to execute S206.

[0105] S211: End

[0106] The second heart and brain operation characteristic is the state of the reinforcement learning model, updating the first rehabilitation exercise strategy is the action of the reinforcement learning model, and the second scoring value is the reward of the reinforcement learning model.

[0107] In summary, in the embodiment of the present application, a cardiovascular rehabilitation intelligent assistance method is provided, wherein the controller obtains a first fused cardiovascular feature according to the patient's attribute characteristics and the first cardiovascular operation characteristic. It can be understood that the first fused cardiovascular feature can more accurately describe the cardiovascular operation of the patient before the first rehabilitation exercise strategy is executed.

[0108] The controller inputs the first fused cardio-cerebral feature into the pre-trained scoring model to obtain a first scoring value for the patient's cardiovascular and cerebrovascular operating state, wherein the first scoring value can reflect the degree of cardiovascular and cerebrovascular disease of the patient before the first rehabilitation exercise strategy is executed.

[0109] The controller inputs the first fused heart-brain feature into the pre-trained first rehabilitation exercise strategy determination model to obtain the first rehabilitation exercise strategy within the preset exercise cycle; it can be understood that the first rehabilitation exercise strategy obtained based on the first fused heart-brain feature has high reliability. At this time, the patient can exercise the preset exercise cycle based on the first rehabilitation exercise strategy.

[0110] After a preset exercise cycle, the controller receives the second cardio-cerebral operation characteristic collected by the wearable component from the patient; the controller obtains a second fused cardio-cerebral characteristic based on the patient's attribute characteristics and the second cardio-cerebral operation characteristic. It can be understood that the second fused cardio-cerebral characteristic can more accurately describe the cardiovascular and cerebrovascular operation status of the patient after executing the first rehabilitation exercise strategy.

[0111] The controller inputs the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the patient's cardiovascular and cerebrovascular operating state, wherein the second scoring value can reflect the degree of cardiovascular and cerebrovascular disease of the patient after executing the first rehabilitation exercise strategy.

[0112] When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than the preset difference threshold, it means that the patient's cardiovascular and cerebrovascular disease severity has not been improved after the preset exercise cycle of the first rehabilitation exercise strategy. Therefore, the first rehabilitation exercise strategy can be updated according to the second cardiovascular and cerebrovascular operation characteristics and the second score value using the reinforcement learning model; and the controller returns to the step of receiving the second cardiovascular and cerebrovascular operation characteristics collected by the wearable component after the preset exercise cycle, until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than the preset difference threshold. It can be understood that the first rehabilitation exercise strategy finally updated can help the patient's cardiovascular and cerebrovascular disease severity to be significantly improved after the patient exercises the preset exercise cycle, and has high reliability.

[0113] In some embodiments, the first rehabilitation exercise strategy includes multiple different types of exercise movements and exercise intensities corresponding to each exercise movement, and the second rehabilitation exercise strategy includes multiple different types of exercise movements and exercise intensities corresponding to each exercise movement. The above step 3 can be specifically implemented as follows:

[0114] Step A: The controller adds the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy into the first rehabilitation exercise strategy.

[0115] Step B: The controller averages the exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise intensity, and uses the average exercise intensity to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy.

[0116] Specifically, the exercise intensity includes exercise frequency and exercise duration. Step B can be specifically implemented as follows: the controller averages the exercise frequency of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy with the exercise frequency of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise frequency, and uses the average exercise frequency to update the exercise frequency of the exercise action in the first rehabilitation exercise strategy; the controller averages the exercise duration of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy with the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise duration, and uses the average exercise duration to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

[0117] In addition, the intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation includes a dedicated exercise area and a camera facing the dedicated exercise area. After S205, the method provided in the embodiment of the present application further includes:

[0118] The controller plays a first exercise video associated with a first rehabilitation exercise strategy on the human-computer interaction screen, where the first exercise video includes a plurality of first exercise action sub-videos in a sequential order.

[0119] The first exercise video is used to guide the patient to exercise according to the exercise movements in the first exercise video. The plurality of first exercise action sub-videos in a sequential order may include exercise action sub-video 1, exercise action sub-video 2, exercise action sub-video 3, ..., exercise action sub-video N.

[0120] The controller receives the second motion action sub-videos of the patient in the dedicated motion area captured by the camera, and determines the order of the second motion action sub-videos.

[0121] For example, the camera captures multiple second motion action sub-videos, and the current second motion action sub-video is the second one, then the order of the currently captured second motion action sub-video is determined to be 2.

[0122] The controller determines the similarity between the second movement action sub-video of the sequence and the first movement action sub-video of the sequence in the first rehabilitation movement strategy.

[0123] For example, the similarity between the second motion action sub-video with the order of 2 and the first motion action sub-video with the order of 2 is determined.

[0124] When the similarity is lower than the set similarity threshold, the controller outputs motion correction prompt information to the human-computer interaction screen.

[0125] When the similarity of the controller is lower than the set similarity threshold, it indicates that the patient's movement is not standardized. Therefore, the human-computer interaction screen outputs movement correction prompt information to remind the patient to perform more standardized movements.

[0126] Specifically, the similarity between the second movement action sub-video of the order and the first movement action sub-video of the order in the first rehabilitation movement strategy may be determined according to the following process:

[0127] S401: Determine multiple candidate key actions corresponding to the first motion action sub-video of the sequence. For example, the multiple candidate key actions corresponding to the first motion action sub-video of the sequence may be determined manually (e.g., by a doctor, an expert, etc.) or by a machine learning model. The candidate key actions are actions that are prone to deformation.

[0128] S402: For each sample patient cluster, obtain test videos of a plurality of sample patients included in the sample patient cluster performing the first motion action sub-video of the sequence and action standard scores corresponding to the test videos. For example, the action standard scores corresponding to the test videos may be determined manually (e.g., by a doctor, an expert, etc.);

[0129] S403: For each test video, test action features (eg, relative positional relationships of multiple joint points, etc.) corresponding to multiple candidate key actions of the test video may be determined;

[0130] S404: according to the first motion action sub-video in the order, standard action features of a plurality of candidate key actions may be determined;

[0131] S405: For each sample patient cluster, a correlation coefficient between each candidate key action and the sample patient cluster may be calculated based on the test action features and standard action features of the test video corresponding to the plurality of candidate key actions, and a target key action corresponding to the sample patient cluster may be determined according to the correlation coefficient between each candidate key action and the sample patient cluster;

[0132] S406: According to the target key action corresponding to the patient cluster of similar sample patients, the action features of the target key action are extracted from the second motion action sub-video of the order, and the standard action features of the target key action are extracted from the first motion action sub-video of the order; based on the action features of the target key action and the standard action features of the target key action, the similarity between the second motion action sub-video of the order and the first motion action sub-video of the order in the first rehabilitation exercise strategy is calculated.

[0133] Specifically, the correlation coefficient between the candidate key action and the sample patient cluster can be calculated according to the following formula:

[0134]

[0135] Among them, r ((i,q),action) is the correlation coefficient between the qth candidate key action corresponding to the ith sample patient cluster and the ith sample patient cluster, F (g,test) is the test action feature of the qth candidate key action corresponding to the test video of the gth sample patient included in the i-th sample patient cluster, F (g,stamdard) is the standard action feature of the qth candidate key action corresponding to the i-th sample patient cluster, G is the total number of sample patients included in the i-th sample patient cluster, S g is the action standard score corresponding to the test video of the g-th sample patient included in the i-th sample patient cluster.

[0136] The similarity between the second movement action sub-video of the order and the first movement action sub-video of the order in the first rehabilitation movement strategy can be calculated according to the following formula:

[0137]

[0138] Among them, S action is the similarity between the second motion action sub-video of the sequence and the first motion action sub-video of the sequence in the first rehabilitation exercise strategy, P2 is a preset parameter, P2 is greater than 0, r ((similar,u),action) is the correlation coefficient between the u-th target key action corresponding to the patient's similar sample patient cluster and the patient's similar sample patient cluster, U is the total number of target key actions corresponding to the patient's similar sample patient cluster, and F (u,present) is the action feature of the u-th target key action corresponding to the patient cluster of similar sample patients, F (u,standard) is the standard action feature of the u-th target key action corresponding to the patient's similar sample patient cluster.

[0139] Still see Figure 1 , the embodiment of the present application also provides an intelligent assistance system for cardiovascular and cerebrovascular rehabilitation. It should be noted that the basic principle and technical effects of the intelligent assistance system for cardiovascular and cerebrovascular rehabilitation provided by the embodiment of the present application are the same as those of the above-mentioned embodiment. For the sake of brief description, for matters not mentioned in the embodiment of the present application, reference can be made to the corresponding contents in the above-mentioned embodiment. The intelligent assistance system for cardiovascular and cerebrovascular rehabilitation includes a controller, a human-computer interaction screen, and a wearable component. The controller is electrically connected to the human-computer interaction screen and the wearable component, respectively. The controller is used to:

[0140] Receiving patient attribute characteristics entered by the patient on the human-computer interaction screen, wherein the patient attribute characteristics at least include patient age information, gender information, height information and weight information;

[0141] Receiving a first heart and brain operation characteristic of the patient collected by the wearable component, wherein the first heart and brain operation characteristic at least includes a heart rate waveform, a blood pressure waveform, and an electroencephalogram of the patient;

[0142] According to the patient's attribute characteristics and the first heart-brain operation characteristics, a first fusion heart-brain characteristic is obtained;

[0143] Inputting the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operation status of the patient, wherein the scoring model is a training sample set constructed based on a plurality of historical fused cardio-cerebral features and corresponding historical scoring values, and inputted into a first neural network for training;

[0144] Inputting the first fused heart-brain feature into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, wherein the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on a plurality of historical fused heart-brain features and corresponding historical rehabilitation exercise strategies;

[0145] After a preset exercise cycle, receiving a second heart and brain operation characteristic of the patient collected by the wearable component, wherein the second heart and brain operation characteristic at least includes a heart rate waveform, a blood pressure waveform and an electroencephalogram of the patient;

[0146] According to the patient's attribute characteristics and the second heart-brain operation characteristics, a second fused heart-brain characteristic is obtained;

[0147] Inputting the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient;

[0148] When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than the preset difference threshold, the reinforcement learning model is used to update the first rehabilitation exercise strategy according to the second heart and brain operation characteristics and the second score value; and the step of returning to execute after the preset exercise cycle, the controller receives the second heart and brain operation characteristics collected by the wearable component on the patient until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than the preset difference threshold.

[0149] In a possible implementation manner, the controller is further configured to:

[0150] Receiving medical record information transmitted by a patient terminal, the medical record information including diagnosis results associated with cardiovascular and cerebrovascular diseases;

[0151] Inputting the diagnosis result associated with the cardiovascular and cerebrovascular disease into the pre-trained second rehabilitation exercise strategy determination model to obtain the second rehabilitation exercise strategy within the preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is constructed based on a training sample set constructed based on a plurality of historical diagnosis results associated with the cardiovascular and cerebrovascular disease and the corresponding historical rehabilitation exercise strategies, and inputting the training sample set into the second neural network;

[0152] According to the second rehabilitation exercise strategy, the first rehabilitation exercise strategy is revised.

[0153] In a possible implementation, the first rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, the second rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each exercise movement, and the controller is specifically used to

[0154] Adding the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy to the first rehabilitation exercise strategy;

[0155] The exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy is averaged with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain the average exercise intensity, and the average exercise intensity is used to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy.

[0156] In a possible implementation, the exercise intensity includes the exercise frequency and the exercise duration, and the controller is specifically configured to:

[0157] averaging the movement frequency of the movement action in the first rehabilitation movement strategy that belongs to the second rehabilitation movement strategy and the movement frequency of the movement action in the second rehabilitation movement strategy to obtain an average movement frequency, and using the average movement frequency to update the movement frequency of the movement action in the first rehabilitation movement strategy;

[0158] The exercise duration of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy is averaged with the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain the average exercise duration, and the average exercise duration is used to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

[0159] In a possible implementation, the intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation further includes a dedicated motion area and a camera facing the dedicated motion area, and the controller is specifically used for:

[0160] Playing a first exercise video associated with a first rehabilitation exercise strategy on a human-computer interaction screen, the first exercise video including a plurality of first exercise action sub-videos in a sequential order;

[0161] receiving a second motion sub-video of the patient in the dedicated motion area captured by a camera, and determining an order of the second motion sub-videos;

[0162] Determining the similarity of the ordered second movement action sub-video to the ordered first movement action sub-video in the first rehabilitation movement strategy;

[0163] When the similarity is lower than the set similarity threshold, the motion correction prompt information is output to the human-computer interaction screen.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent auxiliary method for cardiovascular and cerebrovascular rehabilitation, characterized in that: Applied to a cardiovascular and cerebrovascular rehabilitation intelligent assistance system, the cardiovascular and cerebrovascular rehabilitation intelligent assistance system includes a controller, a human-computer interaction screen and a wearable component, the controller is electrically connected to the human-computer interaction screen and the wearable component respectively, and the method includes: The controller receives the patient attribute characteristics entered by the patient on the human-computer interaction screen, wherein the patient attribute characteristics at least include the patient's age information, gender information, height information and weight information; The controller receives the first cardio-cerebral operation characteristics of the patient collected by the wearable component, wherein the first cardio-cerebral operation characteristics at least include the patient's heart rate waveform, blood pressure waveform and brain wave; The controller obtains a first fused heart-brain feature according to the patient attribute feature and the first heart-brain operation feature; The controller inputs the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operation status of the patient, wherein the scoring model is a training sample set constructed based on a plurality of historical fused cardio-cerebral features and corresponding historical scoring values, and is input into a first neural network for training; The controller inputs the first fused heart-brain feature into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, wherein the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on a plurality of historical fused heart-brain features and respectively corresponding historical rehabilitation exercise strategies; The controller receives the second cardio-cerebral operation characteristics collected by the wearable component from the patient after the preset exercise cycle, wherein the second cardio-cerebral operation characteristics at least include the patient's heart rate waveform, blood pressure waveform and brain wave; The controller obtains a second fused heart-brain feature according to the patient attribute feature and the second heart-brain operation feature; The controller inputs the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient; The controller updates the first rehabilitation exercise strategy according to the second cardio-cerebral operation characteristic and the second score value using a reinforcement learning model when the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value but the difference between the second score value and the first score value is lower than a preset difference threshold; and returns to the step in which the controller receives the second cardio-cerebral operation characteristic collected by the wearable component from the patient after the preset exercise cycle, until the second score value is greater than the first score value and the difference between the second score value and the first score value is higher than a preset difference threshold.

2. The method according to claim 1, characterized in that After the controller receives the second heart and brain operation characteristic of the patient collected by the wearable component after the preset exercise cycle, the method further includes: The controller receives medical record information transmitted by the patient terminal, wherein the medical record information includes diagnosis results associated with cardiovascular and cerebrovascular diseases; The controller inputs the diagnosis result associated with the cardiovascular and cerebrovascular disease into a pre-trained second rehabilitation exercise strategy determination model to obtain a second rehabilitation exercise strategy within a preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is a training sample set constructed based on a plurality of historical diagnosis results associated with the cardiovascular and cerebrovascular disease and the corresponding historical rehabilitation exercise strategies, and is input into a second neural network for training; The controller modifies the first rehabilitation exercise strategy according to the second rehabilitation exercise strategy.

3. The method according to claim 2, characterized in that The first rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each of the exercise movements, the second rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each of the exercise movements, and the controller modifies the first rehabilitation exercise strategy according to the second rehabilitation exercise strategy, including: The controller adds the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy into the first rehabilitation exercise strategy; The controller averages the exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise intensity, and uses the average exercise intensity to update the exercise intensity of the exercise action in the first rehabilitation exercise strategy.

4. The method according to claim 3, characterized in that The exercise intensity includes exercise frequency and exercise duration, and the controller averages the exercise intensity of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy and the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise intensity, and updates the exercise intensity of the exercise action in the first rehabilitation exercise strategy using the average exercise intensity, including: The controller averages the movement frequency of the movement action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy and the movement frequency of the movement action in the second rehabilitation exercise strategy to obtain an average movement frequency, and updates the movement frequency of the movement action in the first rehabilitation exercise strategy using the average movement frequency; The controller averages the exercise duration of the exercise action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy and the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise duration, and uses the average exercise duration to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

5. The method according to claim 1, characterized in that The cardiovascular and cerebrovascular rehabilitation intelligent assistance system includes a dedicated exercise area and a camera facing the dedicated exercise area. After the controller inputs the first fused cardiovascular and cerebrovascular features into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, the method further includes: The controller plays a first exercise video associated with the first rehabilitation exercise strategy on the human-computer interaction screen, wherein the first exercise video includes a plurality of first exercise action sub-videos in a sequential order; The controller receives the second motion sub-videos of the patient in the dedicated motion area captured by the camera, and determines the order of the second motion sub-videos; The controller determines the similarity between the second movement action sub-video in the order and the first movement action sub-video in the order in the first rehabilitation movement strategy; When the similarity is lower than a set similarity threshold, the controller outputs motion correction prompt information to the human-computer interaction screen.

6. An intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation, characterized in that: The intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation includes a controller, a human-computer interaction screen and a wearable component. The controller is electrically connected to the human-computer interaction screen and the wearable component respectively, and the controller is used to: Receiving patient attribute characteristics entered by the patient on the human-computer interaction screen, wherein the patient attribute characteristics at least include patient age information, gender information, height information, and weight information; Receiving a first cardio-cerebral operation characteristic of a patient collected by the wearable component, wherein the first cardio-cerebral operation characteristic at least includes a heart rate waveform, a blood pressure waveform, and an electroencephalogram of the patient; Obtaining a first fused heart-brain feature according to the patient attribute feature and the first heart-brain operation feature; Inputting the first fused cardio-cerebral feature into a pre-trained scoring model to obtain a first scoring value for the cardiovascular and cerebrovascular operation status of the patient, wherein the scoring model is a training sample set constructed based on a plurality of historical fused cardio-cerebral features and their corresponding historical scoring values, and is input into a first neural network for training; Inputting the first fused heart-brain feature into a pre-trained first rehabilitation exercise strategy determination model to obtain a first rehabilitation exercise strategy within a preset exercise cycle, wherein the first rehabilitation exercise strategy determination model is trained based on a training sample set constructed based on a plurality of historical fused heart-brain features and corresponding historical rehabilitation exercise strategies; After the preset exercise cycle, receiving the second cardio-cerebral operation characteristics of the patient collected by the wearable component, wherein the second cardio-cerebral operation characteristics at least include the patient's heart rate waveform, blood pressure waveform and brain wave; Obtaining a second fused heart-brain feature according to the patient attribute feature and the second heart-brain operation feature; Inputting the second fused cardio-cerebral feature into the pre-trained scoring model to obtain a second scoring value for the cardiovascular and cerebrovascular operating status of the patient; When the second score value is less than or equal to the first score value, or when the second score value is greater than the first score value, but the difference between the second score value and the first score value is lower than a preset difference threshold, the first rehabilitation exercise strategy is updated according to the second heart and brain operation characteristic and the second score value using a reinforcement learning model; and the step of receiving the second heart and brain operation characteristic collected by the wearable component from the patient after the preset exercise cycle is returned to execution until the second score value is greater than the first score value, and the difference between the second score value and the first score value is higher than a preset difference threshold.

7. The system according to claim 6, characterized in that The controller is also used for: Receiving medical record information transmitted by a patient terminal, wherein the medical record information includes diagnosis results associated with cardiovascular and cerebrovascular diseases; Inputting the diagnosis result associated with the cardiovascular and cerebrovascular disease into a pre-trained second rehabilitation exercise strategy determination model to obtain a second rehabilitation exercise strategy within a preset exercise cycle, wherein the second rehabilitation exercise strategy determination model is a training sample set constructed based on a plurality of historical diagnosis results associated with the cardiovascular and cerebrovascular disease and the corresponding historical rehabilitation exercise strategies, and inputting the training sample set into the second neural network; The first rehabilitation exercise strategy is modified according to the second rehabilitation exercise strategy.

8. The system according to claim 7, characterized in that The first rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each of the exercise movements, the second rehabilitation exercise strategy includes a plurality of different types of exercise movements and an exercise intensity corresponding to each of the exercise movements, and the controller is specifically used to Adding the exercise movements and corresponding exercise intensities in the second rehabilitation exercise strategy that do not belong to the first rehabilitation exercise strategy to the first rehabilitation exercise strategy; The exercise intensity of the exercise action belonging to the second rehabilitation exercise strategy in the first rehabilitation exercise strategy is averaged with the exercise intensity of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise intensity, and the exercise intensity of the exercise action in the first rehabilitation exercise strategy is updated using the average exercise intensity.

9. The system according to claim 8, characterized in that The exercise intensity includes exercise frequency and exercise duration, and the controller is specifically used for: averaging the movement frequency of the movement action in the first rehabilitation exercise strategy that belongs to the second rehabilitation exercise strategy and the movement frequency of the movement action in the second rehabilitation exercise strategy to obtain an average movement frequency, and using the average movement frequency to update the movement frequency of the movement action in the first rehabilitation exercise strategy; The exercise duration of the exercise action belonging to the second rehabilitation exercise strategy in the first rehabilitation exercise strategy is averaged with the exercise duration of the exercise action in the second rehabilitation exercise strategy to obtain an average exercise duration, and the average exercise duration is used to update the exercise duration of the exercise action in the first rehabilitation exercise strategy.

10. The system according to claim 6, characterized in that The intelligent auxiliary system for cardiovascular and cerebrovascular rehabilitation further includes a dedicated motion area and a camera facing the dedicated motion area, and the controller is specifically used for: Playing a first exercise video associated with the first rehabilitation exercise strategy on the human-computer interaction screen, wherein the first exercise video includes a plurality of first exercise action sub-videos in a sequential order; receiving a second motion sub-video of the patient in the dedicated motion area captured by the camera, and determining an order of the second motion sub-videos; Determine the similarity between the second movement action sub-video in the sequence and the first movement action sub-video in the sequence in the first rehabilitation exercise strategy; When the similarity is lower than a set similarity threshold, motion correction prompt information is output to the human-computer interaction screen.

Images