A method for classifying and scoring completion of training movements in a rehabilitation training system

Abstract

The application discloses a training action classification and completion degree scoring method in a rehabilitation training system, and is applied to a rehabilitation training doctor-patient interactive system.The method comprises the following steps: in step one, a patient completes a rehabilitation training action according to a guide instruction and a standard training action guide video preset in the system, and collects action data information; in step two, the training action collected in step one is classified according to a preset training action classification criterion, and different completion degree scoring methods are adopted for different classifications; and in step three, suitable action features are extracted according to the training action collected in step one, and the completion degree score of the current training action is calculated in combination with the corresponding completion degree scoring method in step two.The training score obtained by the application is more in line with the rehabilitation training condition, the score can help a rehabilitation doctor more intuitively understand the rehabilitation training condition of the patient, and meanwhile, it is convenient for the patient to know whether the training action of the patient is standard or not, so that timely correction can be made.

Description

Technical Field

[0001] This invention relates to the field of smart healthcare, and in particular to a method for classifying training movements and scoring completion in a rehabilitation training system. Background Technology

[0002] With the increasing aging of the population, the prevalence of cardiovascular and cerebrovascular diseases is gradually rising, while the frequent occurrence of traffic accidents is also leading to an increase in the number of people suffering accidental disabilities. Rehabilitation training is particularly important in the treatment of paralyzed and disabled patients, requiring a significant investment of medical resources. In light of this situation, there is an urgent need for a system that can automatically guide patients through rehabilitation training.

[0003] Traditional rehabilitation training programs include a set of movements such as arm extension, elbow rotation, limb stretching, horse stance squats, and lunge stretches. These movements are designed by rehabilitation physicians who guide patients to complete them daily in a set quantity. The achievement of these rehabilitation exercises requires evaluation by professional rehabilitation physicians. Because of the numerous types of exercises and the highly specialized and subjective nature of evaluating their completion, a large number of professional medical staff are needed to guide patients and evaluate their training effectiveness. Given the current shortage of medical personnel, there is an urgent need for a method that can automatically guide patients' rehabilitation training and score their movements.

[0004] Because rehabilitation training exercises are diverse and have varying evaluation criteria, and because rehabilitation training scoring is highly subjective and specialized, accurate scoring requires extensive experience from rehabilitation training experts. Existing methods for calculating rehabilitation exercise completion scores do not consider the complexities of practical applications by categorizing rehabilitation exercises. They use the same or similar methods to calculate scores for different categories of exercises, thus failing to obtain reasonable and accurate completion scores. Chinese invention patent CN108346457A discloses a rehabilitation training assessment method, device, and system. This solution compares the patient's limb movements with preset standard limb movements to obtain the degree of conformity between the current limb movement and the standard limb movement, and then generates a rehabilitation training score for the current limb based on this conformity. While this solution can obtain the patient's rehabilitation training score, it is only suitable for calculating scores for single joint or limb rehabilitation exercises. It is not suitable for calculating scores for rehabilitation exercises involving multiple joints or combined limb movements, nor for calculating scores for continuous state training exercises. Chinese invention patent CN113707268A discloses a rehabilitation training assessment method and system. This method analyzes video footage to compare the patient's training movements with preset movement trajectories, determining the standardization of the patient's movements and thus obtaining a rehabilitation training score. While this method can solve the problem of calculating rehabilitation training scores for single-joint and multi-joint combination training movements, it still cannot solve the problem of calculating rehabilitation training scores for continuous state training movements. Therefore, a systematic analysis of rehabilitation training movements and a reasonable and accurate assessment of the completion rate of rehabilitation training movements are currently lacking.

[0005] Therefore, we propose a method that uses computer technology to classify existing rehabilitation training movements and employs a more targeted approach to score the completion of rehabilitation movements for different movement categories. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for classifying and scoring training movements in a rehabilitation training system. This method collects the patient's training movement data through a training movement data acquisition module, calculates the training score through a classification and scoring module, and determines the degree of rehabilitation in conjunction with the rehabilitation training plan.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows:

[0008] A method for classifying training movements and scoring completion in a rehabilitation training system includes the following steps:

[0009] Step 1: Collect motion data information of patients when they perform rehabilitation training movements according to the pre-set guidance instructions and standard training movement guidance videos in the system;

[0010] Step 2: Based on the preset training movement classification criteria, classify the movement data information collected in Step 1, and select different completion scoring methods for different classifications;

[0011] Step 3: Based on the motion data information collected in Step 1, extract motion features and, in conjunction with the completion scoring method corresponding to Step 2, calculate the completion score of the current training motion.

[0012] Preferably, the specific method for collecting motion data in step one is as follows: collect videos of the patient's rehabilitation training movements, and use a skeletal point recognition algorithm to obtain the coordinate information of the patient's skeletal points during the training process, which is used as motion data.

[0013] Preferably, the preset training movement classification criteria specifically include: single-joint training movements, multi-joint combination training movements, and continuous state training movements. Among them, single-joint training movements refer to the process of moving a single joint or limb from a starting position to an ending position; multi-joint combination training movements refer to the process of moving multiple joints or limbs from a starting position to an ending position; and continuous state training movements refer to the state of multiple joints or limbs of the body moving to a certain state and maintaining it for a certain period of time.

[0014] Preferably, the extracted motion features include: angles between joints, normalized distances between skeletal points, completion time of the training motion, and duration of the motion state.

[0015] Preferably, the range of motion and time of a single joint during single-joint training movements are calculated and scored. The specific scoring method is as follows: the minimum and maximum range of motion of a single joint are detected throughout the entire movement, and the difference between the two is used to obtain the patient's training movement range, which is then combined with the time taken to complete the movement for scoring. The specific scoring formula is as follows:

[0016] ;

[0017] Where F single Scoring of single-joint training movements; This refers to the amplitude of movement of a single joint or limb during rehabilitation training exercises. The standard value for the range of motion of a single joint or limb during rehabilitation training exercises; The time it takes for a patient to complete a single joint or limb training exercise. It is the standard time for completing the actions in the single-joint or limb training process; A1 is the influence weight value of the movement amplitude on the rehabilitation training score; A2 is the influence weight value of the action completion time on the rehabilitation training score.

[0018] Preferably, for the multi-joint combined training actions, the coordinated cooperation of multiple joints is adopted. When calculating the comprehensive score according to the movement amplitudes of each joint, the movement amplitudes of each joint should be comprehensively considered. The specific calculation method is: detect the minimum and maximum values of the movement amplitudes of each joint during the entire movement process, and obtain the movement amplitude of each joint in the patient's training action based on the difference between the two. Different weight values are assigned to the movement amplitudes of each joint, and the score is calculated in combination with the time to complete the action. The specific scoring formula is:

[0019] ;

[0020] Among them, F multi is the score of the multi-joint combined training action; F i single is the individual score of the i-th joint or limb action during the rehabilitation training action, and 1 < i <= n, where n is the number of joints or limb actions involved in the current multi-joint combined training action; is the amplitude value of the i-th joint or limb action during the rehabilitation training action; is the standard amplitude value of the i-th joint or limb action during the rehabilitation training action; T i is the time for the patient to complete the i-th joint or limb action; T si is the standard time for completing the i-th joint or limb action; Ai 1 is the influence weight value of the i-th joint or limb action amplitude on the rehabilitation training score; Ai 2 is the influence weight value of the i-th joint or limb action completion time on the rehabilitation training score; B i is the influence weight value of the i-th joint or limb action score on the overall score.

[0021] Preferably, for the continuous state training action, it should be considered qualified when reaching the standard action state and maintaining it for a certain time. The specific calculation method is: detect the movement amplitude value of a single or multiple joints during the entire movement process, and select a stable peak value as the movement state amplitude of a single or multiple joints in the patient's training action; and score this value in combination with the duration of the action state. The specific scoring formula is:

[0022] ;

[0023] Among them, F state is the score of the continuous state training action; k is the action state matching coefficient of the continuous state training action; T is the duration of the action state of the continuous state training action; T sThe standard action state duration for training continuous state movements; sim(Q,Q) s The similarity between the action state of a continuous state training movement and the standard action state is calculated. When the similarity exceeds a similarity threshold, the duration of the action state is accumulated. TH is the action state similarity threshold for the continuous state training movement. m is the number of action state features involved in the current continuous state training movement based on expert experience. j δ represents the weight value of the influence of the j-th action state feature on the overall action state similarity in the continuous state training action, and 1 <= j <= m; Q is the state feature vector composed of all j action state features involved in the continuous state training action, where δ j To train the j-th action state feature; Q s Let δs be the standard state feature vector composed of all j standard action state features involved in the continuous state training action, where δs j Let j be the state features of the j-th standard action in the training action.

[0024] The beneficial effects of this invention are:

[0025] This invention pre-classifies different rehabilitation training movements and uses more targeted scoring methods to score different types of movements. At the same time, different movement characteristics can be selected for scoring calculation, so that the calculated scores are more accurate and more consistent with the subjective judgment of medical staff.

[0026] When applied to the calculation of rehabilitation training movement scores, this invention can obtain a more accurate, intuitive and reasonable score result. Medical staff can use this score to analyze the patient's rehabilitation training status more accurately. At the same time, patients can also understand the completion status of each rehabilitation training movement more intuitively, rather than a general score, which makes it easier for patients to improve their rehabilitation training movements. Attached Figure Description

[0027] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0028] Figure 1 This is an overall system block diagram of a specific embodiment of the present invention;

[0029] Figure 2 This is a flowchart illustrating the use of the patient's terminal in a specific embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of virtual human skeleton points in a specific embodiment of the present invention;

[0031] Figure 4 This is a diagram illustrating the right arm raising motion in a specific embodiment of the present invention;

[0032] Figure 5 This is a diagram illustrating the right leg lunge and leg press motion in a specific embodiment of the present invention;

[0033] Figure 6 This is a diagram illustrating the horse stance squatting motion in a specific embodiment of the present invention.

[0034] The numbers in the diagram are as follows: 1 represents the nasal bone point, 2 represents the right shoulder bone point, 3 represents the left shoulder bone point, 4 represents the right elbow bone point, 5 represents the left elbow bone point, 6 represents the right wrist bone point, 7 represents the left wrist bone point, 8 represents the right hip bone point, 9 represents the left hip bone point, 10 represents the right knee bone point, 11 represents the left knee bone point, 12 represents the right ankle bone point, and 13 represents the left ankle bone point. Detailed Implementation

[0035] The following description and accompanying drawings fully illustrate specific embodiments of the invention to enable those skilled in the art to practice them.

[0036] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0037] In the following description, when referring to the accompanying drawings, the same numbers in different drawings denote the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of systems and methods consistent with some aspects of the invention as detailed in the appended claims.

[0038] In the description of this invention, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of these terms in this invention based on the specific circumstances. Furthermore, in the description of this invention, unless otherwise stated, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0039] Please see Figure 1As shown in the figure, a specific embodiment of the present invention provides an interactive system for rehabilitation exercise training, including a training plan setting module, a training result query module, a training plan acquisition module, a training exercise data acquisition module, a training exercise classification and scoring module, a training result query and upload module, a medical staff terminal, a patient terminal, a system administrator terminal, and a server system.

[0040] This terminal is for use by healthcare professionals such as rehabilitation physicians. Rehabilitation physicians can use this terminal to select rehabilitation exercises for patients, create rehabilitation training plans, and view patient rehabilitation training results data. The terminal includes a training plan setting module and a training result query module.

[0041] The training plan setting module is used to create suitable rehabilitation training plans for patients. Specifically, the rehabilitation physician first selects appropriate rehabilitation training exercises based on the patient's injury, illness, or disability; then, based on the patient's physical condition, determines the daily number of repetitions and training time for each exercise; finally, the entire training plan is sent to the patient through the server system.

[0042] The training results query module allows rehabilitation physicians to easily view the results of patients' rehabilitation exercise training. This module allows physicians to view the scores of patients completing different rehabilitation exercises and, based on the results, provide training suggestions or modify training plans for patients.

[0043] The patient terminal is for use by patients undergoing rehabilitation training or their families. The specific usage procedure is as follows: Figure 2 As shown. Patients and their families can view the patient's rehabilitation training plan on this terminal and complete the rehabilitation training exercises according to the guidance. The terminal's built-in scoring system scores the completion of the rehabilitation training exercises and displays the scores on the patient's terminal for viewing. Simultaneously, the rehabilitation training exercise scores are saved and uploaded to the medical staff's terminal for rehabilitation physicians to review the patient's training progress. This terminal includes a training plan acquisition module, a training exercise classification and scoring module, a training result query module, and a reporting module.

[0044] The training plan acquisition module is used to access and view rehabilitation training plans developed by rehabilitation physicians. Each rehabilitation training plan contains several exercises. Patients can use this module to view these exercises, along with the required number of repetitions, the standard for completion, and the rehabilitation physician's training guidance.

[0045] The training movement data acquisition module is used to collect rehabilitation training movement data information of patients. In this embodiment, a camera device is used to capture video of the patient's movements, and then human skeletal point recognition algorithms such as BlazePose are used to obtain the coordinate information of each skeletal point of the patient during the rehabilitation movement. The patient's rehabilitation training movement data information is obtained by processing the coordinate information of the skeletal points.

[0046] It should be noted that the BlazePose human skeleton point recognition algorithm is used in this embodiment. This algorithm can recognize a total of 33 skeleton points, but in practical applications, some useless skeleton point coordinate information can be discarded as needed. The skeleton point information used in this invention is as follows: Figure 3 As shown. However, in practical applications, the above-mentioned human skeleton point recognition algorithm can also be OpenPose, etc. The human skeleton point recognition algorithm is not limited in this embodiment.

[0047] The training movement classification and scoring module is used to classify and score patients' rehabilitation training movements. Because rehabilitation training movements are numerous and complex, a single scoring method cannot accurately and effectively evaluate the completion and accuracy of patients' training movements. Therefore, it is necessary to pre-classify different rehabilitation training movements and then use different scoring methods for different movement categories to obtain a more accurate training completion score.

[0048] Rehabilitation training movements can be broadly categorized into three main types: single-joint training movements, multi-joint combination training movements, and continuous state training movements. Using targeted calculation methods for different types of training movements can effectively evaluate the accuracy and completion of the training movements.

[0049] Single-joint training movements include flexion, extension, and rotation of trunk joints. These movements have no fixed initial or final state. The scoring is based on the range of motion of a single joint and the time taken to complete the movement during the rehabilitation training.

[0050] The scoring formula is as follows:

[0051] ;

[0052] Where F single Scoring of single-joint training movements; This refers to the amplitude of movement of a single joint or limb during rehabilitation training exercises. The standard value for the range of motion of a single joint or limb during rehabilitation training exercises; The time it takes for a patient to complete a single joint or limb training exercise. The standard time for completing a single joint or limb training process; A1 is the weight value of the influence of movement range on rehabilitation training score; A2 is the weight value of the influence of movement completion time on rehabilitation training score; the judgment criteria for joint or limb movement range should adopt different judgment characteristics according to different joint or limb movements.

[0053] The scoring calculation process is explained in detail below using the raising motion of the right arm joint as an example. The raising motion of the right arm joint is as follows: Figure 4 As shown, from (a) the initial state through (b) the intermediate state to (c) the final state, the right arm gradually rises throughout the process, and the right arm joint is trained.

[0054] Guided by the patient terminal, the patient raises their right arm with their right side facing the camera. The skeletal coordinates of the patient during the movement are obtained through the training motion data acquisition module.

[0055] To perform the lifting motion of the right arm joint, the coordinate information of three skeletal points needs to be collected, as follows:

[0056] The x-coordinate and y-coordinate of the right gluteal bone point are referred to as RHx and RHy below;

[0057] The x-coordinate and y-coordinate of the right scapular bone point are referred to as RSx and RSy below;

[0058] The x-coordinate and y-coordinate of the right elbow bone point are referred to as REx and REy, respectively, below.

[0059] The formula for calculating the right arm joint range of motion W is:

[0060] ;

[0061] The angle of the right arm joint during the entire lifting motion is calculated using the above formula, and then processed to obtain the maximum value W of the right arm joint angle during the motion. max With minimum value W min The time T for completing the movement, and the amplitude W of the change in the right arm joint angle during the preset standard movement. s and standard motion completion time T s The scores are obtained through comparison and calculation. The influence of the range of motion on the rehabilitation training score can be changed by a preset weight value A1; the influence of the time taken to complete the motion on the rehabilitation training score can be changed by a preset weight value A2.

[0062] The specific scoring formula is as follows:

[0063] ;

[0064] The scoring calculation process of the above-mentioned right elbow joint rotation action is only an embodiment of the action scoring calculation in the single-joint training process. The actions in the single-joint training process not only include the right elbow joint rotation action, and the influencing factors of the scoring not only include the angles of the joints. For some actions, the normalized distance of the skeletal joints can also be used for scoring.

[0065] The multi-joint combined training actions include actions such as lunges and stretching of both arms. Such actions have no fixed initial state or fixed end state, and require the coordinated cooperation of multiple joints during the rehabilitation training actions. Therefore, when calculating the score for such actions, the movement amplitudes of each joint should be comprehensively considered.

[0066] The scoring calculation formula is as follows:

[0067] ;

[0068] where F multi is the score of the multi-joint combined training action; F i single is the individual score of the i-th joint or limb action during the rehabilitation training action, and 1 < i <= n, where n is the number of joints or limb actions involved in the current multi-joint combined training action; is the amplitude value of the i-th joint or limb action during the rehabilitation training action; is the standard amplitude value of the i-th joint or limb action during the rehabilitation training action; T i is the time for the patient to complete the i-th joint or limb action; T si is the standard time to complete the i-th joint or limb action; Ai 1 is the influence weight value of the amplitude of the i-th joint or limb action on the rehabilitation training score; Ai 2 is the influence weight value of the completion time of the i-th joint or limb action on the rehabilitation training score; B i is the influence weight value of the score of the i-th joint or limb action on the overall score; The judgment criteria for the amplitude of the joint or limb action should adopt different judgment features according to different joint or limb actions.

[0069] The following takes the right-leg lunge as an example to elaborate on the scoring calculation process in detail. The completion score of the right-leg lunge action needs to be judged using two joint or limb action features according to expert experience, namely the bending angle of the right knee joint and the normalized distance between the left and right ankles. Therefore, the value of n in the above formula is 2. The process of the right-leg lunge action is as Figure 5 shown, starting from the initial state (a), going through the intermediate state (b) and moving to the end state (c). During the whole process, both hands press on the right knee joint and squat continuously, while the left foot and the right foot are separated by a certain distance, and the right knee joint and the hip joint are trained.

[0070] Guided by the patient terminal, the patient faces the camera with their right side and performs a right-leg lunge. The training motion data acquisition module obtains the skeletal coordinates of the patient during the motion.

[0071] For the right leg lunge stretch, the coordinate information of four skeletal points needs to be collected, as follows:

[0072] The x-coordinate and y-coordinate of the right gluteal bone point are referred to as RHx and RHy below;

[0073] The x-coordinate and y-coordinate of the right knee bone point are referred to as RKx and RKy, respectively, below.

[0074] The x-coordinate and y-coordinate of the right ankle bone point are referred to as RAx and RAY, respectively, below.

[0075] The x-coordinate and y-coordinate of the left ankle bone point are referred to as LAx and LAy, respectively, below.

[0076] The formula for calculating the right knee joint range of motion W is:

[0077] ;

[0078] The formula for calculating the normalized distance D between the left and right ankle joints is:

[0079] ;

[0080] The angle of the right knee joint during the entire lunge stretch is calculated using the above formula, and then processed to obtain the maximum value W of the right knee joint angle during the movement. max With minimum value W min The time T1 for completing the right knee joint movement, and the amplitude W of the right knee joint angle change during the preset standard movement. s The standard right knee joint movement completion time T s1 Scores are obtained through comparison and calculation. The influence of the right knee joint angle on the rehabilitation training score can be changed by a preset weight value A11, and the influence of the right knee joint movement completion time on the rehabilitation training score can be changed by a preset weight value A12. Simultaneously, for the lunge stretching exercise, the normalized distance D of the left and right ankle joints during the training movement and the normalized distance D of the left and right ankle joints in the standard movement are compared. sThe impact on rehabilitation training effectiveness is also significant. The influence of the normalized distance between the left and right ankle joints on the rehabilitation training score can be altered by a preset weight value A2 1; the influence of the movement completion time of the left and right feet on the rehabilitation training score can be altered by a preset weight value A2 2; the influence of the right knee joint movement completion score on the overall score can be altered by a preset weight value B1; and the influence of the distance score between the left and right feet on the overall score can be altered by a preset weight value B2. The specific scoring calculation formula is as follows:

[0081] ;

[0082] The scoring calculation process for the lunge stretching exercise described above is only one example of scoring calculation for multi-joint combination training processes. Multi-joint combination training processes do not only include lunge stretching exercises, and the factors influencing the combination score are not limited to joint angles and normalized distances.

[0083] Continuous state training exercises include movements such as squatting in a horse stance and standing with arms raised. Patients are considered qualified if they can achieve a certain state of motion and maintain it for a certain period of time.

[0084] The scoring formula is as follows:

[0085] ;

[0086] Where F state The rating of the continuous state training action; k is the action state matching coefficient of the continuous state training action; T is the duration of the action state of the continuous state training action; T s The standard action state duration for training continuous state movements; sim(Q,Q) s The similarity between the action state of a continuous state training movement and the standard action state is calculated. When the similarity exceeds a similarity threshold, the duration of the action state is accumulated. TH is the action state similarity threshold for the continuous state training movement. m is the number of action state features involved in the current continuous state training movement based on expert experience. j δ represents the weight value of the influence of the j-th action state feature on the overall action state similarity in the continuous state training action, and 1 <= j <= m; Q is the state feature vector composed of all j action state features involved in the continuous state training action, where δ j To train the j-th action state feature; Q s Let δs be the standard state feature vector composed of all j standard action state features involved in the continuous state training action, where δs jThe j-th standard movement state characteristic is used in the training movement; different movement state characteristics need to be selected based on expert experience for different continuous state training movements.

[0087] The following uses the horse stance squat as an example to explain the scoring calculation process in detail. The completion score of the horse stance squat requires the use of three joint or limb movement characteristics based on expert experience: the bending angle of the right knee joint, the bending angle of the left knee joint, and the vertical angle of the upper limb. Therefore, the value of m in the above formula is 3. The horse stance squat process is as follows: Figure 6 As shown, the movement progresses from (a) the initial state through (b) the intermediate state to (c) the final state. Throughout the process, the knees are constantly bent while the upper body remains relatively perpendicular to the ground, thus training the knee joints.

[0088] Guided by the patient's terminal, the patient faces the camera with their right side and performs a horse stance squat. The motion data acquisition module then captures the skeletal coordinates of the patient during this movement. For the horse stance squat, the coordinates of three skeletal points need to be collected, as follows:

[0089] The x-coordinate and y-coordinate of the right gluteal bone point are referred to as RHx and RHy below;

[0090] The x-coordinate and y-coordinate of the right knee bone point are referred to as RKx and RKy, respectively, below.

[0091] The x-coordinate and y-coordinate of the right ankle bone point are referred to as RAx and RAY, respectively, below.

[0092] The x-coordinate and y-coordinate of the left gluteal bone point are referred to as LHx and LHy below;

[0093] The x-coordinate and y-coordinate of the left knee bone point are referred to as LKx and LKy, respectively, below.

[0094] The x-coordinate and y-coordinate of the left ankle bone point are referred to as LAx and LAy, respectively, below.

[0095] The x-coordinate and y-coordinate of the right scapular bone point are referred to as RSx and RSy below;

[0096] For the horse stance squat, the knees must be bent at a certain angle while the upper limbs remain upright. If this state is met and maintained for a period of time, it is considered a perfect score. There is no need to judge the process in between.

[0097] The formula for calculating the right knee joint range of motion W1 is:

[0098] ;

[0099] The formula for calculating the left knee joint range of motion W2 is:

[0100] ;

[0101] The formula for calculating the vertical angle W3 of the upper limb is:

[0102] ;

[0103] The specific scoring formula is as follows:

[0104] ;

[0105] Where F3 is the score for the horse stance squat; k is the motion state matching coefficient for the horse stance squat; and T is the duration of the motion state for the horse stance squat. s The standard duration of the horse stance squat is sim(Q,Q). s ) represents the similarity between the horse stance squatting motion and the standard motion state. When the similarity of the motion states is greater than the similarity threshold, the duration of the motion state begins to accumulate; TH is the motion state similarity threshold for the horse stance squatting motion; Q is the state feature vector composed of the motion state features of the right knee joint, left knee joint, and upper limb in the horse stance squatting motion, where W1, W2, and W3 are the motion state features of the right knee joint, left knee joint, and upper limb in the horse stance squatting motion, respectively; Q s W is a standard state feature vector composed of the motion state features of the right knee joint, left knee joint, and upper limb during the horse stance squatting action, where W s1 W s2 W s3 A1, A2, and A3 represent the standard movement state characteristics of the right knee joint, left knee joint, and upper limb in the horse stance squatting action; A1, A2, and A3 are the influence weight values ​​of the movement state of the right knee joint, left knee joint, and upper limb in the horse stance squatting action on the overall movement state similarity.

[0106] The above-described scoring calculation process for the horse stance squat is only one example of the scoring calculation for state training movements. State training movements include more than just the horse stance squat.

[0107] The training result query and reporting module allows patients to check their rehabilitation training scores. If the score is too low, patients can choose to redo the rehabilitation training movements to make the movements more standard and obtain better rehabilitation training results. If patients are satisfied with the score, they can choose to save the rehabilitation training movement scores and rehabilitation training movement data through the server system and upload them to the medical staff's terminal so that rehabilitation physicians can provide rehabilitation training opinions or modify the rehabilitation training plan.

[0108] The server system is used to transmit and store interactive data between terminals of medical staff, patients, and system administrators.

[0109] The system administrator terminal is used by administrators of the rehabilitation training interactive system. Administrators can use this terminal to modify and delete rehabilitation training plans set by rehabilitation physicians and rehabilitation training exercise scores uploaded by patients. At the same time, administrators can use this terminal to register and deregister users of medical staff terminals and patient terminals.

[0110] The technical features in the above examples can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for classifying training movements and scoring completion in a rehabilitation training system, characterized in that, Includes the following steps: Step 1: Collect motion data information of patients when they perform rehabilitation training movements according to the pre-set guidance instructions and standard training movement guidance videos in the system; Step 2: According to the preset training action classification criteria, classify the data information collected in Step 1, and select different completion scoring methods for different classifications; Step 3: Based on the motion data information collected in Step 1, extract motion features and, in conjunction with the completion scoring method corresponding to Step 2, calculate the completion score of the current training motion. The pre-defined training movement classification criteria specifically include: single-joint training movements, multi-joint combination training movements, and continuous state training movements. Single-joint training movements refer to the process of moving a single joint or limb from a starting position to an ending position. Multi-joint combination training movements refer to the process of moving multiple joints or limbs from a starting position to an ending position. Continuous state training movements refer to the state of multiple joints or limbs of the body moving to a certain state and maintaining it for a certain period of time. For continuous state training movements, the score F state It is calculated using the following formula: ； Where F state The rating of the continuous state training action; k is the action state matching coefficient of the continuous state training action; T is the duration of the action state of the continuous state training action; T s The standard action state duration for training continuous state actions; sim(Q, Q s The similarity between the action state of a continuous state training movement and the standard action state is calculated. When the similarity exceeds a similarity threshold, the duration of the action state is accumulated. TH is the action state similarity threshold for the continuous state training movement. m is the number of action state features involved in the current continuous state training movement based on expert experience. j δ represents the weight value of the influence of the j-th action state feature on the overall action state similarity in the continuous state training action, and 1 <= j <= m; Q is the state feature vector composed of all j action state features involved in the continuous state training action, where δ j To train the j-th action state feature; Q s Let δ be the standard state feature vector composed of all j standard action state features involved in the continuous state training action, where δ sj Let j be the state features of the j-th standard action in the training action.

2. The method for classifying training movements and scoring completion in the rehabilitation training system as described in claim 1, characterized in that, The specific method for collecting motion data in step one is as follows: collect videos of the patient's rehabilitation training movements, and use a skeletal point recognition algorithm to obtain the coordinate information of the patient's skeletal points during the training process, which is used as motion data.

3. The method for classifying training movements and scoring completion in the rehabilitation training system as described in claim 1, characterized in that... The extracted motion features include: angles between joints, normalized distances between skeletal points, completion time of the training movement, and duration of the motion state.

4. The method for classifying training movements and scoring completion in the rehabilitation training system according to claim 1, characterized in that, The specific scoring formula for single-joint training movements is as follows: ； Where F single Scoring of single-joint training movements; This refers to the amplitude of movement of a single joint or limb during rehabilitation training exercises. The standard value for the range of motion of a single joint or limb during rehabilitation training exercises; The time it takes for a patient to complete a single joint or limb training exercise. A1 represents the standard time for completing a single joint or limb training process; A2 represents the weighting of the range of motion on the rehabilitation training score; and A3 represents the weighting of the time taken to complete the movement on the rehabilitation training score.

5. The method for classifying training movements and scoring completion in the rehabilitation training system according to claim 1, characterized in that, The specific scoring formula for multi-joint combination training movements is as follows: ； Where F multi is the score of the multi-joint combined training action; F i single is the individual score of the i-th joint or limb movement during the rehabilitation training action, and 1 < i <= n, where n is the number of joints or limb movements involved in the current multi-joint combined training action; is the amplitude value of the i-th joint or limb movement during the rehabilitation training action; is the standard amplitude value of the i-th joint or limb movement during the rehabilitation training action; T i T represents the time it takes for the patient to complete the movement of the i-th joint or limb; si The standard time to complete the movement of the i-th joint or limb; Ai1 is the weight value of the influence of the range of motion of the i-th joint or limb on the rehabilitation training score; Ai2 is the weight value of the influence of the completion time of the movement of the i-th joint or limb on the rehabilitation training score; B i The weight value for the influence of the i-th joint or limb movement score on the overall score.

Images