Female pelvic floor rehabilitation training supervision intelligent adjusting system based on applet and cloud platform

Abstract

The invention discloses a female pelvic floor rehabilitation training supervision intelligent adjusting system based on an applet and a cloud platform, and belongs to the technical field of female pelvic floor rehabilitation training supervision. The user side comprises an information acquisition module, a training guidance module and a training feedback module; the information acquisition is used for acquiring pelvic floor information of a user; the training guidance module is used for guiding the user to perform rehabilitation training; the training feedback module is used for carrying out training effect analysis and determining a training evaluation result according to the pelvic floor scoring graph; when the training evaluation result is abnormal training, the rehabilitation training of the user is supervised; when the training evaluation result is always normal training in the evaluation stage, determining a training adjustment result, and generating training feedback data according to the training adjustment result; the platform end comprises a pelvic floor analysis module; and the pelvic floor analysis module is used for analyzing the pelvic floor information of the user, determining training guidance data and sending the training guidance data to the corresponding user side.

Description

Technical Field

[0001] The present invention belongs to the technical field of female pelvic floor rehabilitation training supervision, and specifically is an intelligent adjustment system for female pelvic floor rehabilitation training supervision based on a mini-program and a cloud platform. Background Art

[0002] The female pelvic floor muscles, as an important structure for supporting pelvic visceral organs, play an important role in maintaining female physiological functions and improving the quality of life. However, due to various factors such as childbirth, aging, and chronic diseases, the female pelvic floor muscles often suffer from varying degrees of damage and relaxation, leading to the occurrence of pelvic floor dysfunction diseases.

[0003] In existing pelvic floor rehabilitation training methods, although there are various active training methods (such as Kegel training) and passive training and adjuvant therapies (such as electrostimulation therapy, biofeedback therapy, etc.), patients often face some challenges when performing these trainings. Specifically, patients may be unable to perform the training on time for various reasons, such as improper time management, lack of motivation, or being interfered by other life events. In addition, as the training progresses, the condition of the patient's pelvic floor muscles may change, but existing training methods often cannot adjust the training content in a timely or accurate manner according to these changes; moreover, the short-term training effect may not be obvious, resulting in patients having no intuitive training effect to motivate them to train. For the problem of not training on time, most existing solutions rely on the patient's self-management and the doctor's oral reminder, but these methods often lack effectiveness and sustainability. For the problem of untimely adjustment or not knowing how to adjust the training content, existing training methods usually lack a personalized feedback mechanism and cannot be dynamically adjusted according to the specific situation and needs of the patient.

[0004] Based on this, the present invention provides an intelligent adjustment system for female pelvic floor rehabilitation training supervision based on a mini-program and a cloud platform. Summary of the Invention

[0005] In order to solve the problems existing in the above solution, the present invention provides an intelligent adjustment system for female pelvic floor rehabilitation training supervision based on a mini-program and a cloud platform.

[0006] The object of the present invention can be achieved by the following technical solutions:

[0007] An intelligent adjustment system for female pelvic floor rehabilitation training supervision based on a mini-program and a cloud platform, comprising a platform end and a user end;

[0008] The user end includes an information collection module, a training guidance module, and a training feedback module;

[0009] The information collection is used to collect the pelvic floor information of the user, and after tagging the collected pelvic floor information with the corresponding user tags, send it to the platform end.

[0010] The training guidance module is used to guide the user to perform rehabilitation training, identify the training guidance data sent by the platform end, collect the corresponding training tutorials according to the training tutorial information in the training guidance data; guide the user to perform rehabilitation training according to the training tutorials, and generate the user's rehabilitation training records.

[0011] The training feedback module is used to analyze the training effect, set a pelvic floor score graph, where the horizontal axis of the pelvic floor score graph is time and the vertical axis is the pelvic floor score. The pelvic floor score graph includes a pelvic floor score curve, a pelvic floor prediction curve, and a training prediction curve;

[0012] Fit the pelvic floor score curve, the pelvic floor prediction curve, and the training prediction curve respectively to obtain a pelvic floor score function, a pelvic floor prediction function, and a training prediction function; mark the pelvic floor score function, the pelvic floor prediction function, and the training prediction function as PT(t), PY(t), and XY(t) respectively, where t is time;

[0013] Real-time identify the evaluation stage corresponding to the current time, real-time identify the time interval corresponding to the evaluation stage, and mark it as [t a ,t c ;

[0014] Analyze the pelvic floor prediction function according to the training prediction function to obtain a training evaluation result, and the training evaluation result includes normal training and abnormal training;

[0015] When the training evaluation result is abnormal training, supervise the user's rehabilitation training;

[0016] When the training evaluation result has been normal training within the evaluation stage, analyze the pelvic floor score function according to the training prediction function to obtain a training adjustment result, and the training adjustment result includes adjusting the training tutorial and not adjusting the training tutorial; generate training feedback data according to the training adjustment result, and send the training feedback data to the pelvic floor analysis module of the platform end.

[0017] Furthermore, a posture training unit is set in the training guidance module, and the posture training unit is used to guide the user to learn the corresponding training tutorials, identify the training tutorials, and set the corresponding guided training postures according to the training tutorials;

[0018] When the user needs to learn the training tutorial, correct the user's training posture according to the guided training posture.

[0019] Furthermore, the method for setting the guided training posture according to the training tutorial includes:

[0020] Identify the rehabilitation postures in the training tutorial and count the completion degree of the rehabilitation postures;

[0021] Build a posture evaluation model, and the expression of the posture evaluation model is:

[0022]

[0023] In the formula: K i is the input data, representing the completion degree of the corresponding rehabilitation posture. i is the subscript, i = 1, 2,..., n, and n is the number of rehabilitation postures; the posture guidance requirements are preset by the platform side; the output data is the posture evaluation value ZS(K i ), and the posture evaluation value is 1 or 0;

[0024] Analyze the completion degree of each rehabilitation posture through the posture evaluation model to obtain the posture evaluation value of the corresponding rehabilitation posture;

[0025] Mark the rehabilitation postures with a posture evaluation value of 1 as the guided training postures.

[0026] Furthermore, the method for correcting the user's training posture according to the guided training posture includes:

[0027] Obtain the camera permission of the user and display the guided training posture to the user; the user trains according to the guided training posture, and the training posture of the user is recognized in real time; integrate the guided training posture into the training posture;

[0028] Determine the adjustment position according to the training posture, mark the adjustment position in the training posture; generate a posture guidance suggestion according to the adjustment position, and correct the user's training posture according to the posture guidance suggestion.

[0029] Furthermore, the method for setting the pelvic floor score chart includes:

[0030] Identify the corresponding pelvic floor score according to the training guidance data, set the pelvic floor score curve according to the pelvic floor score, the horizontal axis of the pelvic floor score curve is time, and the vertical axis is the pelvic floor score; set the initial pelvic floor score chart according to the pelvic floor score curve; update the pelvic floor score curve in real time according to the pelvic floor score obtained in real time;

[0031] Define a reference score, and the reference score is the pelvic floor score corresponding to the training guidance data; mark the corresponding reference score in the initial pelvic floor score chart;

[0032] Mark the corresponding training stage in the initial pelvic floor score chart in real time, associate the corresponding training tutorial with the training stage, perform real-time rehabilitation effect prediction according to the training tutorial, and obtain the pelvic floor prediction score at the corresponding time; generate a pelvic floor prediction curve for the corresponding training stage in the initial pelvic floor score chart according to the pelvic floor prediction score;

[0033] Obtain the user's rehabilitation training records in real time, estimate the real-time rehabilitation effect according to the rehabilitation training records, obtain the training estimation score at the corresponding time, and generate the training estimation curve of the corresponding training stage in the initial pelvic floor score chart according to the training estimation score;

[0034] Mark the current initial pelvic floor score chart as the pelvic floor score chart.

[0035] Furthermore, analyze the pelvic floor estimation function according to the training estimation function, including:

[0036] Analyze the training estimation function and the pelvic floor estimation function according to the training evaluation formula to obtain the corresponding training evaluation value. The training evaluation formula is:

[0037]

[0038] In the formula: WA is the training evaluation value;

[0039] Determine the training evaluation result according to the training evaluation value.

[0040] Furthermore, analyze the pelvic floor score function according to the training estimation function, including:

[0041] Analyze the training estimation function for the pelvic floor score function according to the adjustment evaluation formula to obtain the corresponding adjustment evaluation value. The training evaluation formula is:

[0042]

[0043] In the formula: WB is the adjustment evaluation value; [t v , t s is the adjustment evaluation interval;

[0044] Determine the training adjustment result according to the adjustment evaluation value.

[0045] Furthermore, the user side further includes a detection recommendation module, and the detection recommendation module is used to recommend that the user perform pelvic floor detection at the corresponding time.

[0046] The platform side includes a pelvic floor analysis module;

[0047] The pelvic floor analysis module is used to analyze the user's pelvic floor information, determine the corresponding training tutorial information and pelvic floor score, integrate the training tutorial information and the pelvic floor score into training guidance data, and send the training guidance data to the corresponding user side.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] The present invention effectively promotes patients to perform rehabilitation training on time by real-time monitoring of the training progress and status of patients, and through means such as intelligent reminders and incentive mechanisms, thereby significantly improving the training enthusiasm and compliance of patients. It realizes the personalization and dynamic adjustment of training content; aiming at the problems of untimely adjustment or lack of knowledge on how to adjust training content, the present invention proposes an intelligent evaluation algorithm based on patients' training data and rehabilitation progress, which can automatically analyze the training effect of patients and dynamically adjust the training content according to the evaluation results to ensure the personalization and pertinence of training content and accelerate the rehabilitation process of patients. Through intelligent training management and content adjustment, the present invention can ensure that patients perform rehabilitation training at the most appropriate training intensity and difficulty, avoiding over-training or under-training, thereby improving the overall effect of rehabilitation training and the satisfaction of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a block diagram of the principle of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0053] As Figure 1 shown, a smart regulation system for female pelvic floor rehabilitation training based on a mini-program and a cloud platform includes a platform end and a user end;

[0054] The user end includes an information collection module, a training guidance module, a training feedback module, and a detection recommendation module.

[0055] The information collection is used to collect the pelvic floor information of the user, and the user uploads the information. For example, after the user performs a pelvic floor muscle detection, the corresponding detection information is uploaded. Specifically, various methods such as personal assessment, ultrasound assessment, physical electrical stimulation assessment, and doctor examination assessment can be used for detection. Generally, for users without sufficient detection experience, professional assessment methods can be adopted. For example, the pelvic floor information includes data such as the strength, endurance, and coordination of the pelvic floor muscles. After the collected pelvic floor information is labeled with the corresponding user tags, it is sent to the platform side.

[0056] The training guidance module is used to guide the user to perform rehabilitation training, identify the training guidance data sent by the platform side, and collect the corresponding training tutorials according to the training tutorial information in the training guidance data.

[0057] According to the training tutorial, guide the user to perform rehabilitation training and generate the user's rehabilitation training record. The rehabilitation training record includes relevant information such as training content, training progress, training time, and training duration.

[0058] In one embodiment, in order to ensure the accuracy of the user's training posture, a posture training unit is set. The posture training unit is used to guide the user to learn the corresponding training tutorial, identify the corresponding training tutorial, and set the corresponding guided training posture according to the training tutorial.

[0059] When the user needs to learn the training tutorial, correct the user's training posture according to the guided training posture.

[0060] In one embodiment, the method for setting the guided training posture according to the training tutorial includes:

[0061] Identify each training posture in the training tutorial. For the sake of distinction, mark the training postures in the training tutorial as rehabilitation postures; obtain the completion degree of each rehabilitation posture. The completion degree is the degree of completion of a normal patient directly training according to the rehabilitation posture to complete the rehabilitation posture, which is obtained by statistically analyzing the training data of each historical user or patient.

[0062] Establish a posture evaluation model. The expression of the posture evaluation model is:

[0063]

[0064] In the formula: K i is the input data, representing the completion degree of the corresponding rehabilitation posture. i is the subscript, i = 1, 2,..., n, and n is the number of rehabilitation postures; the posture guidance requirement is set by the platform side, specifically as the corresponding completion degree. For example, the posture guidance requirement is that the completion degree is lower than 70%, that is, the completion degree is too low and guidance practice is required; the output data is the posture evaluation value ZS(K i ), and the posture evaluation value is 1 or 0.

[0065] Analyze the completion of each rehabilitation posture through a posture evaluation model to obtain the posture evaluation value of the corresponding rehabilitation posture;

[0066] Mark the rehabilitation posture with a posture evaluation value of 1 as the guided training posture.

[0067] In other embodiments, the guided training posture can also be identified based on other methods, such as presetting the behavioral characteristics corresponding to the guided training posture, and then performing feature recognition through the behavioral characteristics to determine the guided training posture.

[0068] In one embodiment, the method for correcting the user's training posture according to the guided training posture includes:

[0069] Obtain the user's camera permission, that is, allow the collection of images of the user during training; display the guided training posture to the user; the user trains according to the displayed guided training posture, and the user's training posture is recognized in real time; integrate the guided training posture into the recognized training posture, that is, integrate it into the same image; determine the positions where the posture does not meet the requirements according to the training posture, which can be directly compared with the images, mark the corresponding positions as the adjustment positions, and mark the adjustment positions in the training posture; generate posture guidance suggestions according to the adjustment positions, that is, determine according to the differences between the adjustment positions and the guided training postures, such as raising the abdomen; correct the user's training posture according to the posture guidance suggestions.

[0070] In other embodiments, the user's training posture can also be corrected based on other methods according to the guided training posture.

[0071] The training feedback module is used to analyze the training effect, set a pelvic floor score graph, the horizontal axis of the pelvic floor score graph is time, the vertical axis is the pelvic floor score, and the pelvic floor score graph includes a pelvic floor score curve, a pelvic floor prediction curve, and a training prediction curve;

[0072] Fit the pelvic floor score curve, the pelvic floor prediction curve, and the training prediction curve respectively to obtain the pelvic floor score function, the pelvic floor prediction function, and the training prediction function; mark the pelvic floor score function, the pelvic floor prediction function, and the training prediction function as PT(t), PY(t), and XY(t) respectively, where t is time;

[0073] Real-time identify the training stage to which the current time belongs, mark it as the evaluation stage, and real-time identify the time interval corresponding to the evaluation stage, mark it as [t a , t c ;

[0074] Analyze the pelvic floor prediction function according to the training prediction function to obtain the training evaluation result, and the training evaluation result includes normal training and abnormal training;

[0075] When the training evaluation result is abnormal training, supervise the user's rehabilitation training, such as prompting or urging the user to perform rehabilitation training;

[0076] When the training evaluation result remains normal training until the end of this evaluation stage, analyze the pelvic floor scoring function according to the training prediction function to obtain the training adjustment result. The training adjustment result includes adjusting the training tutorial and not adjusting the training tutorial; that is, analyze whether the training content corresponding to the training tutorial is suitable for this user, and then determine the training adjustment result; generate training feedback data according to the training adjustment result, and send the training feedback data to the pelvic floor analysis module on the platform side; that is, when the training adjustment result is not to adjust the training tutorial, the training feedback data is none and no feedback is made; when the training adjustment result is to adjust the training tutorial, integrate data such as the pelvic floor scoring chart, rehabilitation training records, the patient's self-assessment, discomfort or improvement suggestions during the training into the training feedback data.

[0077] In one embodiment, the setting method of the pelvic floor scoring chart includes:

[0078] Identify the corresponding pelvic floor score according to the training guidance data, set the pelvic floor score curve according to the pelvic floor score. The horizontal axis of the pelvic floor score curve is time, and the vertical axis is the pelvic floor score; set the initial pelvic floor score chart according to the pelvic floor score curve; update the pelvic floor score curve in real time according to the real-time obtained pelvic floor score; mark the corresponding reference score in the initial pelvic floor score chart according to the received pelvic floor score, that is, the accurate pelvic floor score corresponding in the training guidance data;

[0079] Regard the interval between adjacent reference scores as a training stage, and also regard the stage where the current time is located as a training stage. For example, if the training stages are 1-2, 2-3, and the current is 3.1, then the current training stage is 3-3.1 until reaching 4; mark the corresponding training stage in the initial pelvic floor score chart in real time, associate the corresponding training tutorial with the training stage, and perform real-time rehabilitation effect prediction according to the training tutorial to obtain the pelvic floor prediction score at the corresponding time, that is, predict the pelvic floor score of the user when performing rehabilitation training according to the training tutorial until the current time, and mark it as the pelvic floor prediction score; generate the pelvic floor prediction curve of the corresponding training stage in the initial pelvic floor score chart according to the obtained pelvic floor prediction score;

[0080] Obtain the user's rehabilitation training record in real time, perform real-time rehabilitation effect prediction according to the obtained rehabilitation training record to obtain the training prediction score at the corresponding time, and generate the training prediction curve of the corresponding training stage in the initial pelvic floor score chart according to the obtained training prediction score;

[0081] Mark the current initial pelvic floor score chart as the pelvic floor score chart.

[0082] Among them, for the rehabilitation training prediction based on the training tutorial or rehabilitation training record, it can be evaluated based on the existing rehabilitation effect evaluation technology, such as establishing an intelligent model based on the current intelligent technology and making predictions through the intelligent model.

[0083] Exemplarily, collect data such as the user's age, gender, physical condition, pelvic floor medical history, etc. Collect the pelvic floor score before the user undergoes training as the input feature of the model. Collect the pelvic floor score after the user undergoes training as the output target or label of the model. Screen out the features that have a significant impact on the prediction through methods such as correlation analysis and chi-square test. Use methods such as PCA (Principal Component Analysis) and LDA (Linear Discriminant Analysis) for feature dimensionality reduction to improve the computational efficiency of the model. Standardize the data so that it is on the same scale for subsequent analysis and modeling. Select a suitable model according to the type of data and the prediction target, such as linear regression, decision tree, random forest, support vector machine (SVM), neural network, etc. Considering the complexity of the pelvic floor score, a model that can handle non-linear relationships may need to be selected, such as a neural network or a support vector machine.

[0084] Use the training dataset to train the selected model and adjust the model parameters by optimizing the loss function. During the training process, methods such as cross-validation can be used to evaluate the stability and generalization ability of the model. Evaluate the performance of the model through indicators such as confusion matrix, ROC curve, accuracy, recall rate, F1 score, etc. Select appropriate evaluation indicators and make trade-offs and selections according to specific business requirements. Improve the performance of the model by adjusting hyperparameters (such as learning rate, regularization parameter, etc.). Use ensemble learning methods (such as Bagging, Boosting, etc.) to improve the stability and accuracy of the model. Iteratively optimize the model according to the evaluation results until a satisfactory performance level is achieved.

[0085] In one embodiment, analyzing the pelvic floor prediction function according to the training prediction function includes:

[0086] Analyze the training prediction function and the pelvic floor prediction function according to the training evaluation formula to obtain the corresponding training evaluation value. The training evaluation formula is:

[0087]

[0088] In the formula: WA is the training evaluation value;

[0089] Determine the training evaluation result based on the training evaluation value, that is, set the corresponding limit value by the platform or the user. When the training evaluation value is greater than the limit value (the limit value is positive), it indicates that the user's training lag progress exceeds the preset standard, belonging to abnormal training; or when the training evaluation value is less than a certain limit value (the limit value is negative), the training intensity exceeds the preset standard, also belonging to abnormal training; between the two limit values is normal training.

[0090] In one embodiment, analyze the pelvic floor scoring function according to the training prediction function, including:

[0091] Analyze the pelvic floor scoring function according to the training prediction function according to the adjustment evaluation formula to obtain the corresponding adjustment evaluation value. The training evaluation formula is:

[0092]

[0093] In the formula: WB is the adjustment evaluation value; [t v , t s is the adjustment evaluation interval, and the adjustment evaluation interval is the evaluation stage corresponding to the current complete training stage, that is, the training stage corresponding to 3 - 4, and can also be regarded as the nearest complete training stage to the current time;

[0094] Determine the training adjustment result according to the adjustment evaluation value, that is, when the adjustment evaluation is too large, it indicates that under the normal training background, the actual training effect cannot reach the expected training effect, and subsequent training adjustments are required. The specific allowable error is set by the platform or the user.

[0095] The detection and recommendation module is used to intelligently recommend users to perform professional pelvic floor detection to obtain accurate pelvic floor information.

[0096] In one embodiment, the time threshold method can be used for intelligent recommendation, that is, when the accumulated undetected duration reaches the preset value, recommend users to perform detection.

[0097] In one embodiment, intelligent recommendation can also be performed through the pelvic floor scoring chart, and recommendation is made when the curve trend is abnormal and the training effect cannot reach the expected; specifically, multiple methods can be used for intelligent recommendation.

[0098] The platform side includes a pelvic floor analysis module;

[0099] The pelvic floor analysis module is used to analyze the user's pelvic floor information, determine the corresponding training tutorial information and pelvic floor score, integrate the training tutorial information and pelvic floor score into training guidance data, and send the training guidance data to the user side.

[0100] In one embodiment, analyzing the user's pelvic floor information includes:

[0101] Identify user information, and identify the training feedback data of the user based on the user information. If the user is in a situation such as the first evaluation, the training feedback data is none. The training feedback data is sent by the training feedback module in the user terminal and is used to indicate whether the effect of the training tutorial meets the requirements, serving as one of the bases for adjusting the training tutorial. It can also include the completion status of the training tutorial, the patient's self-evaluation, discomfort during the training process, or improvement suggestions, etc.

[0102] Analyze the training feedback data and pelvic floor information to determine the training tutorial, and then generate the training tutorial information of the training tutorial. The training tutorial information can be the access link of the corresponding training tutorial, etc., which is convenient for data transmission. Specifically, analyze the training feedback data and pelvic floor information based on the existing methods, such as analyzed by professional personnel of the platform side; establish a corresponding intelligent model using intelligent technology for analysis, etc.

[0103] Evaluate the corresponding pelvic floor score according to the pelvic floor information. The pelvic floor score is set according to the percentage system. Set the pelvic floor score in the normal state to 100, set the pelvic floor score of the most severe possible pelvic floor condition to 0, and set the corresponding reference pelvic floor information for each pelvic floor score according to the actual severity difference. Then match according to the pelvic floor information to determine the pelvic floor score; it can also be evaluated using the current evaluation method.

[0104] Exemplarily, collect pelvic floor information and training feedback results from the patient, and clean the collected data to handle missing values and outliers. For example, for missing pelvic floor information, it can be filled with the mean value, interpolation method, or deletion; for abnormal training feedback results, they can be detected and processed through statistical methods. Convert the data into a format suitable for analysis, such as unifying the date format, encoding categorical variables (such as One-Hot Encoding), etc. Standardize or normalize the data to make it on the same scale for subsequent analysis and modeling.

[0105] Screen out the features that have a significant impact on the model through methods such as correlation analysis and chi-square test. For example, in the pelvic floor information, indicators such as muscle strength and endurance that are closely related to the training effect may be screened out. Reduce the number of features to improve the generalization ability of the model. Convert high-dimensional data into low-dimensional features through methods such as PCA (Principal Component Analysis) and LDA (Linear Discriminant Analysis), retaining the main information of the data. This helps to reduce the complexity of the model and improve the calculation efficiency.

[0106] Select a suitable model according to the type of data and the analysis goal. For example, for classification problems (such as determining whether a training tutorial meets the requirements), models such as decision trees, random forests, support vector machines (SVMs), etc. can be selected; for regression problems (such as predicting the training effect of patients), models such as linear regression, SVR, etc. can be selected.

[0107] Use the training dataset to train the selected model. Adjust the model parameters by optimizing the loss function (such as mean squared error, cross-entropy, etc.) to make the fitting effect of the model on the training data the best. During the training process, methods such as cross-validation can be used to evaluate the stability and generalization ability of the model.

[0108] Evaluate the accuracy and generalization ability of the model through methods such as cross-validation, confusion matrix, ROC curve, etc. Select appropriate evaluation metrics, such as accuracy, recall, F1 score, AUC-ROC, etc., and make trade-offs and selections according to specific business requirements.

[0109] Improve the performance of the model by adjusting hyperparameters (such as learning rate, regularization parameter, etc.) and using ensemble learning methods (such as Bagging, Boosting, etc.). Iteratively optimize the model according to the evaluation results until a satisfactory performance level is achieved.

[0110] Use interpretive models (such as LIME, SHAP, etc.) to explain the decision-making process of the model, helping to understand the internal mechanism of the model and the importance of features. This helps doctors or therapists better understand the training situation of patients and formulate personalized training plans accordingly.

[0111] Use data visualization tools (such as Matplotlib, Seaborn, Tableau, etc.) to display and analyze the results. Draw various charts (such as bar charts, line charts, heatmaps, etc.) to display the data and analysis results, helping to understand the distribution and trend of the data.

[0112] Deploy the trained model to the actual application scenario.

[0113] The above formulas are all calculated by removing the dimension and taking their numerical values. The formula is obtained by collecting a large amount of data and performing software simulation to get a formula closest to the actual situation. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through a large amount of data simulation.

[0114] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The intelligent adjustment system for female pelvic floor rehabilitation training supervision based on mini-programs and cloud platforms is characterized by: Including platform side and user side; The user terminal includes an information collection module, a training guidance module and a training feedback module; The information collection is used to collect the user's pelvic floor information, and the collected pelvic floor information is marked with a corresponding user tag and sent to the platform end; The training guidance module is used to guide the user to perform rehabilitation training, identify the training guidance data sent by the platform end, collect the corresponding training tutorial according to the training tutorial information in the training guidance data; guide the user to perform rehabilitation training according to the training tutorial, and generate the user's rehabilitation training record; The training feedback module is used to analyze the training effect, and to set a pelvic floor scoring diagram, wherein the horizontal axis of the pelvic floor scoring diagram is time, and the vertical axis is the pelvic floor score, and the pelvic floor scoring diagram includes a pelvic floor scoring curve, a pelvic floor prediction curve, and a training prediction curve; The pelvic floor scoring curve, the pelvic floor prediction curve and the training prediction curve are fitted respectively to obtain the pelvic floor scoring function, the pelvic floor prediction function and the training prediction function; the pelvic floor scoring function, the pelvic floor prediction function and the training prediction function are marked as PT(t), PY(t) and XY(t), respectively, where t is time; Real-time identification of the evaluation phase corresponding to the current time, marked as [t a , t c ]; Analyzing the pelvic floor prediction function according to the training prediction function to obtain a training evaluation result, wherein the training evaluation result includes normal training and abnormal training; When the training evaluation result is abnormal training, the user's rehabilitation training is supervised; When the training evaluation result is normal training throughout the evaluation phase, the pelvic floor scoring function is analyzed according to the training estimation function to obtain a training adjustment result, training feedback data is generated according to the training adjustment result, and the training feedback data is sent to the pelvic floor analysis module on the platform end; The platform end includes a pelvic floor analysis module; The pelvic floor analysis module is used to analyze the user's pelvic floor information, determine the corresponding training course information and pelvic floor score, integrate the training course information and the pelvic floor score into training guidance data, and send the training guidance data to the corresponding user terminal.

2. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 1 is characterized in that: The training guidance module is provided with a posture training unit, which is used to guide the user to learn the corresponding training tutorial, identify the training tutorial, and set the corresponding guiding training posture according to the training tutorial; When the user needs to learn the training tutorial, the user's training posture is corrected according to the guiding training posture.

3. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 2 is characterized in that: The methods for setting guided training postures according to the training tutorial include: Identify the rehabilitation postures in the training tutorial and count the completion degree of the rehabilitation postures; A posture evaluation model is established, and the expression of the posture evaluation model is: Where: K i is the input data, indicating the completion degree of the corresponding rehabilitation posture, i is the subscript, i = 1, 2, ..., n, n is the number of rehabilitation postures; the posture guidance requirements are pre-set by the platform; the output data is the posture evaluation value ZS (K i ), the posture evaluation value is 1 or 0; The completion degree of each rehabilitation posture is analyzed through the posture assessment model to obtain the posture assessment value of the corresponding rehabilitation posture; The rehabilitation posture with a posture evaluation value of 1 is marked as a guided training posture.

4. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 3 is characterized in that: The method of correcting the user's training posture according to the guided training posture includes: Obtain the user's camera permission and display the guided training posture to the user; the user trains according to the guided training posture, and the user's training posture is recognized in real time; the guided training posture is integrated into the training posture; An adjustment position is determined according to the training posture, and the adjustment position is marked in the training posture; a posture guidance suggestion is generated according to the adjustment position, and the user's training posture is corrected according to the posture guidance suggestion.

5. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 1 is characterized in that: The method of setting up the pelvic floor score chart includes: Identify the corresponding pelvic floor score according to the training guidance data, set a pelvic floor score curve according to the pelvic floor score, wherein the horizontal axis of the pelvic floor score curve is time and the vertical axis is the pelvic floor score; set an initial pelvic floor score graph according to the pelvic floor score curve; and update the pelvic floor score curve in real time according to the pelvic floor score obtained in real time; Define a benchmark score, where the benchmark score is the pelvic floor score corresponding to the training guidance data; and mark the corresponding benchmark score in the pelvic floor score initial graph; Marking the corresponding training stage in real time on the pelvic floor score initial graph, associating the corresponding training tutorial for the training stage, performing real-time rehabilitation effect estimation according to the training tutorial, and obtaining the pelvic floor estimation score at the corresponding time; generating a pelvic floor estimation curve for the corresponding training stage in the pelvic floor score initial graph according to the pelvic floor estimation score; Acquire the user's rehabilitation training records in real time, estimate the rehabilitation effect in real time according to the rehabilitation training records, obtain the training estimation score of the corresponding time, and generate the training estimation curve of the corresponding training stage in the pelvic floor scoring initial graph according to the training estimation score; The current pelvic floor scoring initial map is marked as the pelvic floor scoring map.

6. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 1 is characterized in that: The pelvic floor prediction function is analyzed based on the training prediction function, including: According to the training evaluation formula, the training estimation function and the pelvic floor estimation function are analyzed to obtain the corresponding training evaluation value. The training evaluation formula is: Where: WA is the training evaluation value; A training evaluation result is determined according to the training evaluation value.

7. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 1 is characterized in that: The pelvic floor scoring function is analyzed based on the training estimation function, including: According to the adjustment evaluation formula, the training estimation function is analyzed for the pelvic floor scoring function to obtain the corresponding adjustment evaluation value. The training evaluation formula is: Where: WB is the adjusted evaluation value; [t v , t s ] is to adjust the evaluation interval; The training adjustment result is determined according to the adjustment evaluation value.

8. The female pelvic floor rehabilitation training supervision intelligent adjustment system based on mini program and cloud platform according to claim 1 is characterized in that: The user end also includes a detection recommendation module, which is used to recommend the user to perform pelvic floor detection at a corresponding time.

Images