Internet rehabilitation training method and system based on rehabilitation medical digitization

Abstract

The invention discloses an internet rehabilitation training method and system based on rehabilitation medical digitization, and relates to the technical field of rehabilitation training, and the method comprises the steps: collecting and preprocessing the initial motion, emotion and heart rate information of a patient; capturing the motion and emotion of the patient in real time, collecting the real-time heart rate of the patient, and calculating the standardized motion quality score, emotion deviation and heart rate index in combination with the baseline data of the patient; performing multi-modal fusion on the standardized action quality score, the emotion deviation and the heart rate index to obtain a scoring model, making a decision according to the scoring model, and adjusting a current training scheme; in combination with active feedback of the patient, secondary decision making is carried out, and an optimal training scheme is given. On the basis of a multi-modal fusion method, standardized action quality scores, emotion deviations and heart rate indexes are combined to form a scoring model; the current training scheme is dynamically adjusted according to the model so as to adapt to the real-time state of the patient, and the effectiveness and safety of training are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rehabilitation training, and in particular to an Internet rehabilitation training method and system based on digital rehabilitation medicine. Background Art

[0002] As the global population ages, problems such as degenerative bone and joint diseases, muscle atrophy, and balance disorders are becoming increasingly prominent among the elderly. At the same time, various sports injuries caused by factors such as traffic accidents, sports, and industrial production are becoming more frequent. Coupled with the high incidence of neurological diseases such as stroke, brain trauma, and spinal cord injury, many patients face a rigid demand for long-term, systematic rehabilitation training. At the same time, there are approximately 15 million new cases of stroke worldwide each year, of which approximately one-third experience varying degrees of motor dysfunction. Furthermore, patients with spinal cord injuries, due to the lengthy and complex recovery process, also place higher demands on rehabilitation services.

[0003] Although digital and internet-based rehabilitation training systems are currently being applied, they still have the following shortcomings: On the one hand, traditional remote rehabilitation programs often rely on preset thresholds or physician experience to adjust intensity, lacking the ability to integrate multimodal information such as movement quality, emotional bias, and heart rate indicators in real time, making it difficult to achieve dynamic program optimization based on the patient's current comprehensive status. On the other hand, although some rehabilitation training platforms support user ratings or feedback, these are usually only used as supplementary information after training and are not deeply integrated with objective data. This makes it difficult to make secondary adjustments to the training rhythm and difficulty, resulting in limited program personalization and patient satisfaction. Therefore, there is an urgent need for an internet-based rehabilitation training method and system based on the digitalization of rehabilitation medicine to achieve comprehensive assessment of the patient's all-round status and intelligent training program optimization. Summary of the Invention

[0004] The present invention is proposed in view of the lack of ability to perform real-time weighted fusion of multimodal information such as movement quality, emotional deviation and heart rate indicators during rehabilitation training, which makes it difficult to achieve dynamic program optimization based on the patient's current comprehensive state; and the inability to deeply integrate with objective data makes it difficult to make secondary fine-tuning of training rhythm and difficulty, resulting in limited program personalization and patient satisfaction.

[0005] Therefore, the problem to be solved by the present invention is how to provide an Internet rehabilitation training method based on digital rehabilitation medicine.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides an Internet rehabilitation training method based on digital rehabilitation medicine, including: collecting the patient's initial movements, emotions and heart rate information and performing preprocessing to obtain the patient's baseline data; capturing the patient's movements and emotions in real time, collecting the patient's real-time heart rate, combining the patient's baseline data, calculating the movement quality score, emotion deviation measurement and relative heart rate variability index, and performing standardization processing to obtain the standardized movement quality score, emotion deviation and heart rate index; multimodally fusing the standardized movement quality score, emotion deviation and heart rate index in a weighted manner to obtain a scoring model, and making decisions based on the output results of the scoring model to adjust the current training plan; making secondary decisions based on the patient's active feedback to provide the best training plan.

[0008] As a preferred solution of the Internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the patient's movements, emotions and heart rate information are obtained through a motion capture camera, an expression camera and a heart rate belt respectively.

[0009] As a preferred solution of the internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the patient's movement characteristics are obtained and standardized, and the formula is:

[0010]

[0011] in, is the normalized action quality score, σ mov is the standard deviation of the historical action quality score, μ mov is the mean of the historical action quality scores, Q mov (t) Score the patient's real-time movement quality.

[0012] As a preferred solution of the internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the patient's emotional characteristics are obtained and standardized, and the formula is:

[0013]

[0014] in, is the standardized emotional deviation, σ emo is the standard deviation of historical sentiment deviation, μ emo is the mean of historical sentiment deviations, D emo (t) is the patient's real-time emotional deviation measurement.

[0015] As a preferred solution of the Internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the patient's heart rate characteristics are obtained and standardized, and the formula is:

[0016]

[0017] in, is the normalized heart rate index, σ hrv is the standard deviation of the heart rate ratio relative to 1, Q hrv (t) is the relative heart rate variability index.

[0018] As a preferred solution of the internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the motion features, emotion features and heart rate features are multimodally fused in a weighted manner to obtain a scoring model, the formula of which is:

[0019]

[0020] Where S(t) is the fusion score; w1, w2, and w3 are the weight coefficients of the standardized action quality score, the standardized emotional deviation, and the standardized heart rate index, respectively;

[0021] The specific situations in which decisions are made based on the scoring model are:

[0022] If S(t) is less than or equal to the first threshold, it means that the patient's state is "relaxed" and the intensity of the current training should be increased;

[0023] If S(t) is greater than the first threshold and less than or equal to the second threshold, it means that the patient's condition is "normal" and the current training intensity is maintained;

[0024] If S(t) is greater than the second threshold, it means that the patient is in an "overload" state and the intensity of the current training should be reduced.

[0025] As a preferred solution of the internet rehabilitation training method based on digital rehabilitation medicine described in the present invention, the patient's active feedback is incorporated to make a secondary decision, specifically:

[0026] F subj,1 (t)∈[-1,1]: patient’s subjective “too slow / too fast” feedback;

[0027] Specifically, when receiving patient feedback of "too slow", a secondary decision is made to shorten the rest time between the current movements based on the decision of the scoring model S(t); when receiving patient feedback of "too fast", the rest time between the current movements is increased based on the decision of the scoring model S(t);

[0028] F subj,2 (t)∈[-1,1]: patient’s subjective feedback of “too difficult / too easy”;

[0029] Specifically, when the patient feedback "too difficult" is received, a secondary decision is made, and the number of repetitions of the training action is reduced based on the decision of the scoring model S(t); when the patient feedback "too easy" is received, the number of repetitions of the training action is increased based on the decision of the scoring model S(t).

[0030] On the second aspect, in order to further solve the problems existing in rehabilitation training, the embodiment of the present invention provides an Internet rehabilitation training system based on digital rehabilitation medicine, which includes: a data acquisition module for collecting the patient's initial movement, emotion and heart rate information to obtain the patient's baseline data; a feature extraction module for extracting the patient's movement characteristics, emotion characteristics, and heart rate characteristics to obtain the patient's movement quality score, emotion deviation and heart rate index; a multimodal fusion and decision-making module for multimodally fusing the movement quality score, emotion deviation and heart rate index to obtain a scoring model and make a decision; a subjective interaction module for integrating the patient's active feedback into the scoring model to make a secondary decision.

[0031] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of an Internet rehabilitation training method based on digital rehabilitation medicine as described in the first aspect of the present invention is implemented.

[0032] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein: when the computer program is executed by a processor, it implements any step of an Internet rehabilitation training method based on digital rehabilitation medicine as described in the first aspect of the present invention.

[0033] The beneficial effects of the present invention are:

[0034] 1. This invention collects the patient's initial movements, emotions, and heart rate information and captures this data in real time to comprehensively assess the patient's condition. Based on a multimodal fusion approach, it combines standardized movement quality scores, emotional deviations, and heart rate indicators to form a scoring model. Based on the output of this model, the current training plan can be dynamically adjusted to adapt to the patient's immediate condition, improving the effectiveness and safety of the training.

[0035] 2. By incorporating active patient feedback, such as subjective feelings of "too slow / too fast" and "too difficult / too easy," the present invention can make secondary decisions based on the scoring model, further fine-tune the training plan, and implement a human-machine collaborative rehabilitation control mechanism, thereby enhancing patient participation and satisfaction.

[0036] 3. The present invention introduces facial expression feature recognition and emotional deviation assessment, which can identify abnormal psychological states of patients during training and adjust the training rhythm and difficulty accordingly. This allows the present invention to not only consider the quality of action execution, but also incorporate changes in emotions and physiological states (such as heart rate), which helps to reduce rehabilitation anxiety, improve training compliance, and take into account both psychological and physiological rehabilitation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0038] Figure 1 This is a flow chart for implementing the present invention in Example 1.

[0039] Figure 2 This is a flow chart of the secondary decision-making of the present invention in Example 1. DETAILED DESCRIPTION

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0041] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0042] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0043] Example 1

[0044] Reference Figures 1 to 2 , which is the first embodiment of the present invention, provides an Internet rehabilitation training method based on digital rehabilitation medicine, comprising the following steps:

[0045] S1: Collect the patient's initial movements, emotions, and heart rate information and pre-process them to obtain the patient's baseline data.

[0046] Specifically:

[0047] S11, capture N times of the patient's standard demonstration movements through the motion capture camera, and calculate the average angle vector of each key joint:

[0048]

[0049] in, is the average joint angle vector, which serves as the baseline for subsequent deviation comparison; M is the number of key joints; θ j,i is the angle of the jth joint in the i-th demonstration, obtained by the motion capture camera;

[0050] S12, collect the patient's facial expressions in the resting state through the expression camera and calculate the average vector:

[0051]

[0052] in, is the average facial expression vector, which serves as the resting (neutral) expression baseline for subsequent identification of emotional deviation; N is the number of resting samples; F i The feature vector extracted from a resting-state facial expression is captured by the camera and combined with the output of the CNN model;

[0053] It should be noted that the CNN model is a deep learning model designed for processing grid-structured data (such as images, audio, and text), and is an existing mature technology. Its specific implementation method will not be described in detail. In the present invention, it is used to extract facial expression features.

[0054] S13, collect the patient's heart rate data through a heart rate monitor and calculate the emotional deviation:

[0055]

[0056] Among them, SDNN is the patient's heart rate variability, which measures the standard deviation of heart rate fluctuations in the resting state and serves as the baseline of heart rate variability; RR k is the RR interval of the kth heartbeat, obtained from the heart rate belt; is the mean of the RR interval; K is the number of RR interval samples.

[0057] S2: Capture the patient's movements and emotions in real time, collect the patient's real-time heart rate, combine the patient's baseline data, calculate the movement quality score, emotion deviation measurement and relative heart rate variability index, and perform standardization to obtain the standardized movement quality score, emotion deviation and heart rate index.

[0058] Specifically:

[0059] S21, real-time capture of the patient's movements, combined with the patient's average joint angle vector Calculate the action quality score:

[0060]

[0061] Among them, Q mov (t) is the action quality score, ranging from [0, 1], the closer to 1, the more standard the action; θ j (t) is the real-time angle of joint j at time t; is the baseline average angle vector of joint j; max|Δθ j | is the maximum allowable deviation of joint j, set by experts; M is the total number of joints;

[0062] It should be noted that when Q mov (t) When the threshold is lower than the standard action threshold, the patient should be reminded by voice or text, such as "too high elbow";

[0063] When Q mov (t) When the threshold of the action is greater than the standard threshold, the patient can be encouraged by voice or text, such as "You are great";

[0064] It should be noted that the action standard threshold is established by expert experience;

[0065] Furthermore, the action quality score is standardized to obtain the standardized action quality score:

[0066]

[0067] in, is the normalized action quality score, σ mov is the standard deviation of the historical action quality score, μ mov is the mean of the historical action quality scores, Q mov (t) Score the patient's real-time movement quality;

[0068] S22, real-time capture of the patient's facial expression, combined with the patient's facial expression average vector Calculate the sentiment bias metric:

[0069]

[0070] Among them, D emo (t) is the emotion deviation measure, the larger the value, the more obvious the emotion deviates from neutrality; F(t) is the facial expression feature vector extracted at time t, is the average facial expression vector;

[0071] Furthermore, the sentiment deviation measure is standardized to obtain the standardized sentiment deviation:

[0072]

[0073] in, is the standardized emotional deviation, σ emo is the standard deviation of historical sentiment deviation, μ emo is the mean of historical sentiment deviations, D emo (t) is the patient's real-time emotional deviation measurement;

[0074] S23, collect the patient's heart rate in real time, combine it with the patient's heart rate variability SDNN, and calculate the heart rate variability index:

[0075]

[0076] Among them, Q hrv (t) is the relative heart rate variability index, if Q hrv (t)>1, indicating that heart rate variability increases. If Q hrv (t) < 1, indicating decreased heart rate variability; SDNN win (t) is the SDNN of the RR interval in the sliding window at time t; SDNN baseline is the baseline SDNN of heart rate variability in the resting state;

[0077] It should be noted that when Q hrv (t) When the heart rate exceeds the threshold, the patient can be reminded by voice or text, such as "pay attention to take a deep breath";

[0078] When Q hrv (t) When the heart rate is below the threshold, the patient can be encouraged by voice or text, such as "hold on";

[0079] It should be noted that the heart rate thresholds were established by experts;

[0080] Furthermore, the heart rate variability index is standardized to obtain the standardized heart rate index:

[0081]

[0082] in, is the normalized heart rate index, σ hrv is the standard deviation of the heart rate ratio relative to 1, Q hrv (t) is the relative heart rate variability index.

[0083] S3: The standardized movement quality score, emotional deviation and heart rate index are multimodally fused in a weighted manner to obtain a scoring model. Decisions are made based on the output of the scoring model to adjust the current training plan.

[0084] Specifically:

[0085] S31, multimodally fuse the motion features, emotion features, and heart rate features in a weighted manner to obtain a scoring model, the formula of which is:

[0086]

[0087] Where S(t) is the fusion score that comprehensively reflects the action, emotion, and physiological state; w1, w2, and w3 are the weight coefficients of the standardized action quality score, standardized emotion deviation, and standardized heart rate index, respectively, which are calibrated by experts, and w1+w2+w3=1;

[0088] S32, the specific circumstances of decision making based on the scoring model are:

[0089] If S(t) is less than or equal to the first threshold, it means that the patient is in a “relaxed” state and the intensity of the current training should be increased, such as increasing the number of repetitions of the training action and shortening the rest time between the current actions;

[0090] If S(t) is greater than the first threshold and less than or equal to the second threshold, it means that the patient's condition is "normal" and the current training intensity is maintained;

[0091] If S(t) is greater than the second threshold, it means that the patient is in an "overload" state and the intensity of the current training should be reduced, such as reducing the number of repetitions of the training action and extending the rest time between the current actions.

[0092] S4: Reference Figure 2 , based on the patient's active feedback, make secondary decisions and provide the best training plan;

[0093] The details are as follows:

[0094] F subj,1 (t)∈[-1,1]: patient’s subjective “too slow / too fast” feedback;

[0095] Specifically, when receiving patient feedback of "too slow", a secondary decision is made to shorten the rest time between the current movements based on the decision of the scoring model S(t); when receiving patient feedback of "too fast", the rest time between the current movements is increased based on the decision of the scoring model S(t);

[0096] F subj,2 (t)∈[-1,1]: patient’s subjective feedback of “too difficult / too easy”;

[0097] Specifically, when the patient feedback "too difficult" is received, a secondary decision is made, and the number of repetitions of the training action is reduced based on the decision of the scoring model S(t); when the patient feedback "too easy" is received, the number of repetitions of the training action is increased based on the decision of the scoring model S(t).

[0098] This embodiment also provides an Internet rehabilitation training method and system based on digital rehabilitation medicine, including: a data acquisition module, used to collect the patient's initial movement, emotion and heart rate information to obtain the patient's baseline data; a feature extraction module, used to extract the patient's movement characteristics, emotion characteristics, and heart rate characteristics to obtain the patient's movement quality score, emotion deviation and heart rate index; a multimodal fusion and decision-making module, used to multimodally fuse the movement quality score, emotion deviation and heart rate index to obtain a scoring model and make a decision; a subjective interaction module, used to integrate the patient's active feedback into the scoring model and make a secondary decision.

[0099] This embodiment also provides a computer device, which is suitable for an Internet rehabilitation training method based on digital rehabilitation medicine, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement an Internet rehabilitation training method based on digital rehabilitation medicine as proposed in the above embodiment.

[0100] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0101] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements an Internet rehabilitation training method based on digital rehabilitation medicine as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk.

[0102] In summary, the present invention can comprehensively evaluate the patient's condition by collecting the patient's initial movements, emotions, and heart rate information and capturing these data in real time; and based on the multimodal fusion method, the standardized movement quality score, emotional deviation and heart rate index are combined to form a scoring model; according to the results output by this model, the current training plan can be dynamically adjusted to adapt to the patient's immediate condition and improve the effectiveness and safety of the training.

[0103] Example 2

[0104] This is the second embodiment of the present invention. This embodiment is different from the first embodiment in that, in order to verify its beneficial effects, operating data and related instructions of the present invention in actual scenarios are provided.

[0105] Step S1: collecting and preprocessing the patient's initial movements, emotions, and heart rate information;

[0106] S11, use the motion capture camera to capture the patient's 5 standard demonstration movements (such as "elbow flexion and extension") and calculate the average joint angle vector:

[0107]

[0108] The calculation results are as follows: baseline joint angle: 90° (elbow flexion angle); maximum allowable deviation set by experts: ±5°;

[0109] S12, collect the facial expression feature vector of the patient in the resting state through the expression camera:

[0110]

[0111] The result is: the baseline emotion feature vector is [0.2, 0.3, 0.1];

[0112] S13, collect heart rate variability at rest using a heart rate monitor:

[0113]

[0114] Conclusion: The baseline of heart rate variability is 10ms.

[0115] Step S2: The patient performs elbow flexion and extension movements, and the system captures data in real time.

[0116] At S21, the real-time joint angle was captured as 88° (2° away from the baseline);

[0117] Calculate the action quality score:

[0118]

[0119] Standardization:

[0120]

[0121] Among them, the historical mean μ mov =0.5, standard deviation σ mov =0.1;

[0122] S22, obtain the real-time facial expression feature vector as [0.25, 0.3, 0.1];

[0123] Calculating sentiment bias:

[0124]

[0125] Standardization:

[0126]

[0127] Among them, the historical mean μ emo =0.05, standard deviation σ emo =0.04.

[0128] S23, obtain real-time heart rate variability of 12ms;

[0129] Calculate relative heart rate variability:

[0130]

[0131] Standardization:

[0132]

[0133] Where, the standard deviation σ relative to 1hrv =0.2.

[0134] Step S3: multimodal fusion and decision making;

[0135] Parameter setting: w1=0.5, w2=0.3, w3=0.2; threshold setting: the first threshold is -0.5,

[0136] The second threshold is 0.5;

[0137] Fusion score calculation:

[0138] S=(1.0*0.5)+(0.0*0.3)+(1.0*0.2)=0.7;

[0139] Decision result: Fusion score S = 0.7, which is greater than the second threshold of 0.5, and the patient's status is determined to be "overload";

[0140] Adjustment plan: Reduce training intensity, such as reducing the number of repetitions of training movements (from 10 times → 8 times) and extending the rest time between movements (from 10 seconds → 12 seconds).

[0141] Step S4: Active feedback and secondary decision-making;

[0142] Patient feedback: Patients provided feedback through the system interface: “Too fast / too difficult”;

[0143] Secondary decision: Based on the original decision of the fusion scoring model (reducing the training intensity) and patient feedback (too fast / too difficult), further adjustments were made: number of repetitions (from 8 to 6) and rest time between exercises (from 12 seconds to 14 seconds).

[0144] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An Internet rehabilitation training method based on digital rehabilitation medicine, characterized by: include: S1, collects the patient's initial movements, emotions, and heart rate information and preprocesses them to obtain the patient's baseline data; S2 captures the patient's movements and emotions in real time, collects the patient's real-time heart rate, and combines the patient's baseline data to calculate the movement quality score, emotion deviation measure, and relative heart rate variability index. These are then normalized to obtain the standardized movement quality score, emotion deviation, and heart rate index. S3, through weighted multimodal fusion of the standardized movement quality score, emotional bias, and heart rate index, to obtain a scoring model. Based on the output of the scoring model, decisions are made to adjust the current training plan. S4, combined with the patient's active feedback, makes a secondary decision and gives the best training plan.

2. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 1, characterized in that: The patient's movements, emotions and heart rate information are obtained through motion capture cameras, expression cameras and heart rate belts respectively.

3. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 2, characterized in that: The patient's motion characteristics are obtained and standardized, and the formula is: in, is the normalized action quality score, σ mov is the standard deviation of the historical action quality score, μ mov is the mean of the historical action quality scores, Q mov (t) Score the patient's real-time movement quality.

4. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 3, characterized in that: The patient's emotional characteristics are obtained and standardized. The formula is: in, is the standardized emotional deviation, σ emo is the standard deviation of historical sentiment deviation, μ emo is the mean of historical sentiment deviations, D emo (t) is the patient's real-time emotional deviation measurement.

5. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 4, characterized in that: The patient's heart rate characteristics are obtained and standardized. The formula is: in, is the normalized heart rate index, σ hrv is the standard deviation of the heart rate ratio relative to 1, Q hrv (t) is the relative heart rate variability index.

6. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 5, characterized in that: The motion features, emotion features and heart rate features are multimodally fused in a weighted manner to obtain a scoring model, whose formula is: Where S(t) is the fusion score; w1, w2, and w3 are the weight coefficients of the standardized action quality score, the standardized emotional deviation, and the standardized heart rate index, respectively; The specific situations in which decisions are made based on the scoring model are: If S(t) is less than or equal to the first threshold, it means that the patient's condition is "relaxed" and the intensity of the current training should be increased; if S(t) is greater than the first threshold and less than or equal to the second threshold, it means that the patient's condition is "normal" and the intensity of the current training should be maintained; If S(t) is greater than the second threshold, it means that the patient is in an "overload" state and the intensity of the current training should be reduced.

7. The Internet rehabilitation training method based on digital rehabilitation medicine according to claim 6, characterized in that: Incorporate active patient feedback for secondary decision-making, specifically: F subj,1 (t)∈[-1,1]: patient’s subjective “too slow / too fast” feedback; Specifically, when receiving patient feedback of "too slow", a secondary decision is made to shorten the rest time between the current movements based on the decision of the scoring model S(t); when receiving patient feedback of "too fast", the rest time between the current movements is increased based on the decision of the scoring model S(t); F subj,2 (t)∈[-1,1]: patient’s subjective feedback of “too difficult / too easy”; Specifically, when the patient feedback "too difficult" is received, a secondary decision is made, and the number of repetitions of the training action is reduced based on the decision of the scoring model S(t); when the patient feedback "too easy" is received, the number of repetitions of the training action is increased based on the decision of the scoring model S(t).

8. An Internet rehabilitation training system based on digital rehabilitation medicine, based on the Internet rehabilitation training method based on digital rehabilitation medicine according to any one of claims 1 to 7, characterized in that: include: The data collection module is used to collect the patient's initial movements, emotions, and heart rate information to obtain the patient's baseline data; The feature extraction module is used to extract the patient's movement features, emotional features, and heart rate features, and obtain the patient's movement quality score, emotional deviation, and heart rate index; The multimodal fusion and decision-making module is used to perform multimodal fusion of action quality scores, emotional deviations, and heart rate indicators to obtain a scoring model and make decisions; The subjective interaction module is used to incorporate patients' active feedback into the scoring model for secondary decision-making.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of an Internet rehabilitation training method based on digital rehabilitation medicine according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of an Internet rehabilitation training method based on digital rehabilitation medicine according to any one of claims 1 to 7 are implemented.

Images