Diabetic complication meditation behavior management method, system, equipment and medium

Abstract

The invention provides a diabetic complication meditation behavior management method and device, equipment and a medium. The meditation behavior management method comprises the following steps: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained meditation posture real-time monitoring model, and obtaining key point position information; determining whether the meditation behavior is standard or not according to the key point position information and preset standard posture library data; when it is determined that the meditation behavior is not standard, the meditation behavior is adjusted and corrected; wherein the meditation posture real-time monitoring model is obtained by training a basic posture detection model based on meditation behavior training data. The method is used for monitoring and improving the meditation behavior of the patient with the diabetic peripheral neuropathy, intelligent recognition and management of the meditation behavior are achieved through the trained model, meditation duration is dynamically adjusted, adjustment suggestions are provided for the meditation behavior of the target user in real time, and the purpose of personalized health guidance is achieved.

Description

Technical Field

[0001] This invention relates to the field of medical and health recommendations, and in particular to a method, system, device, and medium for managing meditation behavior in patients with diabetic complications. Background Technology

[0002] Diabetic peripheral neuropathy (DPN) is a common chronic disease with a high incidence and disability rate, requiring long-term healthy lifestyle management for patients. Studies have shown that meditation can help DPN patients improve symptoms such as numbness, coldness, and pain in the limbs, thus improving their quality of life. However, many patients find it difficult to maintain a consistent meditation practice, or they engage in improper meditation techniques, affecting the effectiveness of their meditation.

[0003] Currently, many patient management systems for diabetic peripheral neuropathy (DPN) are available on the market, generally relying on wearable devices or smartphone apps. These systems typically include a smart bracelet or smartwatch that detects physiological signals, along with a dedicated application. The smart bracelet can monitor the user's heart rate, activity level, and other physiological data, and display the data in sync with the mobile application via Bluetooth. The application can be used to set reminders, record the patient's daily activities, and analyze the patient's behavioral data to provide health advice and guidance.

[0004] However, the current management methods still have the following problems: Insufficient personalization: Most existing health advice is based on fixed algorithms and general health recommendations, which cannot provide highly personalized management and guidance based on each DPN patient's specific condition, personal health data and lifestyle habits; Inaccurate data: Wearable devices may produce biases when measuring physiological signals, especially when patients are engaged in static activities such as meditation, which can lead to inaccurate data and affect the credibility of health advice; Limited data processing capabilities: It generally only provides basic data recording and tracking, and cannot achieve complex pattern recognition and personalized analysis.

[0005] Therefore, a new method for managing meditation behavior is urgently needed to solve the problems existing in the current technology. Summary of the Invention

[0006] This invention provides a method, system, device, and medium for managing meditation behavior in diabetic complication patients. It addresses the technical problems of biased and insufficiently personalized measurement data when using wearable devices to monitor the meditation behavior of DNP patients in the prior art. This invention utilizes a trained real-time meditation posture monitoring model to monitor the patient's meditation behavior in real time and make adjustments accordingly. Based on the adjusted data, the model is optimized for real-time personalized management while ensuring monitoring accuracy, ultimately aiming to improve the patient's health.

[0007] In a first aspect, the present invention provides a method for managing meditation behavior in patients with diabetic complications, used to monitor and improve the meditation behavior of patients with diabetic peripheral neuropathy, comprising: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point location information; determining whether the meditation behavior is standard based on the key point location information and a preset standard posture library; and adjusting and correcting the meditation behavior if it is determined to be non-standard; wherein the real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

[0008] In one embodiment of the present invention, the key point location information includes at least first joint point location information, second joint point location information, and third joint point location information. The step of determining whether the meditation behavior is standard based on the key point location information and preset standard posture database data includes: Based on the position information of the first joint point and the position information of the first target joint point in the standard posture library, the deviation value of the first joint point is obtained; using the law of cosines, based on the position information of the first joint point, the position information of the second joint point, and the position information of the third joint point, the angle of the first joint point is obtained; based on the deviation value of the first joint point and the angle of the first joint point, it is determined whether the meditation behavior is standard.

[0009] In another embodiment of the present invention, determining whether the meditation behavior is standard based on the first joint point deviation value and the first joint point angle includes: obtaining a first joint point deviation score based on the first joint point deviation value and a first mapping table; obtaining a first joint point angle deviation score based on the first joint point angle and a second mapping table; obtaining a first joint point comprehensive score based on the first joint point deviation score, a first deviation weight value, the first joint point angle deviation score, and a first angle weight value; determining that the meditation behavior is non-standard when the first joint point comprehensive score is less than the standard score; and determining that the meditation behavior is standard when the first joint point comprehensive score is greater than or equal to the standard score. The first mapping table is a table showing the correspondence between joint deviation values ​​and position deviation scores, and the second mapping table is a table showing the correspondence between joint angles and angle deviation scores.

[0010] In another embodiment of the present invention, adjusting and correcting the meditation behavior when it is determined that the meditation behavior is not standard includes: If the meditation behavior is determined to be non-standard, information is provided to the target user through visual or audio signals, indicating the parts of the meditation behavior that need to be adjusted and the adjustment method.

[0011] In another embodiment of the present invention, after adjusting and correcting the meditation behavior when it is determined that the meditation behavior is not standard, the method further includes: recording the key point location information and the characteristic parameters of the target user in real time; and optimizing the real-time monitoring model of the meditation posture based on the key point location information and the characteristic parameters of the target user.

[0012] In another embodiment of the present invention, the method further includes: The analysis results are obtained by analyzing the meditation behavior data of the target users. Based on the analysis results, corresponding solutions are formulated for the problems existing in meditation behavior.

[0013] In another embodiment of the present invention, the target users are multiple, and the real-time acquisition of meditation behavior data of the target users includes: Real-time acquisition of meditation behavior data from multiple target users.

[0014] Secondly, the present invention also provides a meditation behavior management system for diabetic complications applied to any of the above-described methods for managing meditation behavior in diabetic complications, comprising a real-time meditation posture monitoring module and a real-time feedback correction module. The real-time meditation posture monitoring module is connected to the real-time feedback correction module. The real-time meditation posture monitoring module is used to monitor the meditation behavior of a target user in real time and obtain the analysis results and adjustment plan of the target user's meditation behavior. The real-time feedback correction module is used to provide real-time feedback of the analysis results and adjustment plan of the meditation behavior to the target user.

[0015] Thirdly, the present invention provides an electronic device comprising: Processor, memory, and bus, among which, The processor and the memory communicate with each other via the bus; The memory stores program instructions that can be executed by the processor, which can invoke the program instructions to perform the steps of the meditation behavior management method for diabetic complications as described in any of the preceding items.

[0016] Fourthly, the present invention provides a non-transitory computer-readable storage medium storing computer instructions that cause a computer to perform the steps of the meditation behavior management method for diabetic complications described above.

[0017] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the meditation behavior management method for diabetic complications as described in any of the preceding claims.

[0018] This invention provides a method, device, equipment, and medium for managing meditation behavior in patients with diabetic complications. The meditation behavior management method includes: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point location information; determining whether the meditation behavior is standard based on the key point location information and a preset standard posture database; and adjusting and correcting the meditation behavior if it is determined to be non-standard. The real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data. This invention is used to monitor and improve the meditation behavior of patients with diabetic peripheral neuropathy. Through the trained model, it achieves intelligent recognition and management of meditation behavior, dynamically adjusts meditation duration, and provides real-time adjustment suggestions for the target user's meditation behavior, thereby achieving personalized health guidance. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the overall process of a method for managing meditative behavior in diabetic complications provided by the present invention; Figure 2 This is a schematic diagram of the specific process of the meditation behavior management method for diabetic complications provided by the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] In one embodiment of the present invention, such as Figure 1 The diagram shown is a schematic of the meditation behavior management method for diabetic complications provided by the present invention. Figure 1 As shown, this invention provides a method for managing meditation behavior in patients with diabetic complications, used to monitor and improve meditation behavior in patients with diabetic peripheral neuropathy, including: Step S101: Acquire the meditation behavior data of the target user in real time; In this embodiment, in order to achieve intelligent and high-precision real-time monitoring of the target user's meditation behavior, the target user's meditation behavior data is captured in real time by a camera. This data can be video stream data, camera image data, or other data. Different data formats are obtained when different acquisition devices are used, and no specific limitation is made here.

[0023] Step S102: Transmit the meditation behavior data to the pre-trained real-time meditation posture monitoring model to obtain key point location information; Specifically, the acquired meditation behavior data is transmitted to a real-time meditation posture monitoring model. This model analyzes and processes the key point position information and estimates the initial posture using a pre-trained deep learning model (such as OpenPose).

[0024] Step S103: Determine whether the meditation behavior is standard based on the key point location information and the preset standard posture library data; Specifically, it is necessary to compare the key point location information of the target user with the preset standard pose library data in real time. The key point location information includes the coordinates of each joint of the target user, and the joint angles (e.g., the angle between the elbow and knee) are calculated using the coordinates of each joint. In this embodiment, when calculating the joint angle, three adjacent key points are selected to form a virtual triangle, and then the corresponding angle is calculated using the inverse cosine function, for example, using θ = arccos((b 2 + c 2 - a 2 ) / (2bc)), where a, b, and c are the lengths of the three sides that form the triangle.

[0025] For example, to detect whether a target user's arms are flat when meditating, three points can be selected: "left shoulder," "left elbow," and "left wrist." Taking "left elbow" as vertex a, connect left shoulder b and left wrist c, and calculate the angle at the elbow using the cosine theorem.

[0026] The accuracy of meditation is determined by calculating the angle. Based on the precision requirements of different movements, a 15-degree error threshold is set. This angle setting can be preset with a corresponding score; when the obtained angle value exceeds this error threshold, the corresponding score will decrease. It should be noted that this embodiment requires calculating a comprehensive score by combining the angle difference and the deviation of key point positions. This comprehensive score is used to determine whether the meditation behavior is standard. If the score is lower than a certain standard value (e.g., 70 points), it indicates that the posture deviation is too large, and the meditation behavior needs to be adjusted. Specific calculation and analysis methods are described in the following embodiments and are not specifically limited here.

[0027] Step S104: If the meditation behavior is determined to be non-standard, the meditation behavior is adjusted and corrected; wherein, the real-time meditation posture monitoring model is obtained by training the basic posture detection model based on the meditation behavior training data.

[0028] Specifically, when the system detects that the target user's posture deviates significantly, it will immediately provide feedback to the target user through visual or audio signals to adjust and correct the meditation behavior, such as pointing out the specific parts that need adjustment, in order to help them return to the correct posture.

[0029] It should be noted that this invention uses the OpenPose pose estimation model, which can monitor human keypoints in real time. The entire model pre-training process mainly includes data preparation, model training, and pose accuracy detection.

[0030] Data preparation: A dataset was constructed by selecting 100 sets of correct meditation posture video data. The data in the dataset were labeled, and key points of the human body (such as head, shoulders, elbows, wrists, etc.) were labeled. A convolutional neural network was used to extract spatial features from the images. To balance stability and speed, a batch size of 32 was adopted, and the Adam optimizer was used to dynamically adjust the learning rate. The PAF loss function was selected to measure the accuracy of predicting limb connections.

[0031] Model training: The model undergoes forward propagation, loss calculation, backpropagation, and iterative training, iterating multiple times until the loss converges. Based on the evaluation results, hyperparameters such as the learning rate and batch size are adjusted. For learning rate adjustment, a cosine annealing learning rate scheduler is used, setting an initial learning rate, as well as T_max and eta_min, gradually reducing it to a relatively optimized result. A weight decay parameter is added to the optimizer, and a Dropout layer is added to the model with a Dropout probability of 0.5 and weight decay (L2 regularization) of 1e-4.

[0032] Posture Accuracy Detection: Based on standard meditation posture parameters, a custom algorithm is developed to detect posture accuracy. It calculates the angles and distances of joint points to determine if the target user's actual posture conforms to the standard. Key points include the knees, ankles, wrists, and arms. For each specific posture, the angles and distances between the monitored key points are determined. The corresponding angles are calculated based on the joint coordinates, using the law of cosines. The deviation between the actual and ideal values ​​is calculated by comparing the results with the posture parameter set for each movement. An overall posture score (0-100) is calculated based on the magnitude of the deviation at each joint to quantify posture accuracy.

[0033] In this embodiment, a custom algorithm is used to detect the accuracy of the posture. The core logic of this custom algorithm is based on the spatial geometric relationship of key points on the human body, and quantifies the standardization of the posture by calculating angle and distance deviations. The specific process is as follows: 1. Selection and calculation basis of key points In this embodiment, key points refer to the joints of the human body. The custom algorithm uses the OpenPose model to extract 17 joints of the human body in real time (such as nose, shoulder, elbow, wrist, knee, ankle, etc.).

[0034] Calculating the included angle: Select three adjacent joint points to form a virtual triangle. Calculate the angle value of a specific joint using the law of cosines. Let the lengths of the three sides of the triangle be a, b, and c, where a is the side opposite the joint vertex. Then the formula for calculating the joint angle is: θ = arccos((b...c ... 2 + c 2 - a 2 ) / (2bc)).

[0035] 2. Multi-dimensional scoring This custom algorithm generates a comprehensive score by comparing actual measurements with ideal values ​​from a standard pose library. Position deviation score: The corresponding score is obtained by referring to the first mapping table based on the "Joint Point Position Deviation Value".

[0036] Angle deviation score: Based on the difference between the "joint angle" and the standard value, the corresponding score is obtained by referring to the second mapping table.

[0037] Weighted Composite Score: This custom algorithm uses weighted logic to calculate the final score. The formula logic is as follows: Overall score = w_1 * Joint position deviation score + w_2 * Joint angle deviation score Among them, w_1 (deviation weight value) and w_2 (joint angle weight value) are usually set to 0.5 each.

[0038] Application examples: 1. Arm-flat posture detection To detect whether a target user's arms are level while meditating, the algorithm selects three joint points: the left shoulder, the left elbow, and the left wrist. Using the left elbow as the vertex, the system connects the shoulder and wrist to form two sides and calculates the elbow angle in real time using the law of cosines. If the calculated angle deviates significantly from the standard posture, the system immediately alerts the user visually or audio-visually.

[0039] 2. Long-term fatigue and postural deformity detection By recording data for a week, the system discovered that if a target user's "left shoulder key point" frequently sags during the later stages of meditation (resulting in a consistently low posture score), the algorithm will determine that the target user's left shoulder is prone to fatigue or has an involuntary tilt. Based on this analysis, the system will specifically push a video of "relaxation exercises for the left shoulder muscles" in the report and suggest shortening the duration of meditation next time to prevent injury.

[0040] In the implementation process, this embodiment mainly involves the following three steps: Collect data on standard meditation postures: Extract key points from videos of professional meditation practitioners to serve as standard postures.

[0041] Target user action capture: The user meditates at home, and the action is recorded through a camera.

[0042] Real-time evaluation: The application extracts key points from the user in real time, compares them with standard actions, and provides feedback.

[0043] Breathing and heart rate monitoring: Users wear wearable devices that monitor respiratory rate and heart rate data via sensors. The system uses this data to determine whether the user is in a safe and relaxed state.

[0044] According to the method for managing meditation behavior in patients with diabetic complications provided by the present invention, the method includes: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point location information; determining whether the meditation behavior is standard based on the key point location information and a preset standard posture database; and adjusting and correcting the meditation behavior if it is determined to be non-standard. The real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data. This invention is used to monitor and improve the meditation behavior of patients with diabetic peripheral neuropathy. Through the trained model, it achieves intelligent recognition and management of meditation behavior, dynamically adjusts meditation duration, and provides real-time adjustment suggestions for the meditation behavior of target users, thereby achieving personalized health guidance.

[0045] In another embodiment of the present invention, the key point location information includes at least first joint point location information, second joint point location information, and third joint point location information. The step of determining whether the meditation behavior is standard based on the key point location information and preset standard posture library data includes: obtaining a first joint point deviation value based on the first joint point location information and the first target joint point location information in the standard posture library; obtaining the first joint point angle using the law of cosines based on the first joint point location information, the second joint point location information, and the third joint point location information; and determining whether the meditation behavior is standard based on the first joint point deviation value and the first joint point angle.

[0046] Specifically, determining whether the meditation behavior is standard based on the first joint point deviation value and the first joint point angle includes: obtaining the first joint point deviation score based on the first joint point deviation value and a first mapping table; obtaining the first joint point angle deviation score based on the first joint point angle and a second mapping table; obtaining the first joint point comprehensive score based on the first joint point deviation score, a first deviation weight value, the first joint point angle deviation score, and a first angle weight value; determining that the meditation behavior is non-standard when the first joint point comprehensive score is less than the standard score; and determining that the meditation behavior is standard when the first joint point comprehensive score is greater than or equal to the standard score. The first mapping table is a table showing the correspondence between joint deviation values ​​and position deviation scores, and the second mapping table is a table showing the correspondence between joint angles and angle deviation scores.

[0047] Specifically, in this embodiment, whether the meditation behavior is standard is determined by comparing the overall score with the standard score. In this embodiment, the overall score for each key point is calculated using a weighted method, and the formula logic is roughly as follows: Overall Score = w1 * Joint Position Deviation Score + w2 * Joint Angle Deviation Score. Here, w1 is the deviation weight value, and w2 is the angle weight value, each being 0.5. If the angle deviation of a certain joint point exceeds 15 degrees, the corresponding sub-item deviation score will decrease, resulting in an overall score lower than the standard score, thus determining that the meditation behavior is not standard.

[0048] It should be noted that this embodiment uses a table mapping method, such as pre-setting a table to represent the correspondence between joint deviation values ​​and deviation scores. For example, the score corresponding to deviation values ​​between a and b is 90 points, and the score corresponding to deviation values ​​between b and c is 70 points, etc. The angle setting is similar, and the corresponding score is set according to the range of angle difference.

[0049] According to the method for managing meditation behavior in cases of diabetic complications provided in this embodiment, the comprehensive scoring can accurately analyze whether the meditation behavior is standard, thus ensuring the accuracy of meditation behavior management.

[0050] In another embodiment of the invention, such as Figure 2 As shown, adjusting and correcting the meditation behavior when it is determined to be non-standard includes: If the meditation behavior is determined to be non-standard, information is provided to the target user through visual or audio signals, indicating the parts of the meditation behavior that need to be adjusted and the adjustment method.

[0051] If the meditation behavior is determined to be non-standard, after adjusting and correcting the meditation behavior, the method further includes: recording the key point location information and the target user's characteristic parameters in real time; optimizing the real-time meditation posture monitoring model based on the key point location information and the target user's characteristic parameters; analyzing the target user's meditation behavior data to obtain analysis results; and developing corresponding solutions to the problems existing in the meditation behavior based on the analysis results.

[0052] In this embodiment, the meditation management system continuously collects key point location and posture correctness data for each meditation session of the target user, records the two-dimensional coordinates (x, y) and confidence score of each key point, and collects feature parameters of the target user's body shape (such as height, arm length, and leg length). These can be used for adaptive adjustment of the meditation posture real-time monitoring model, construct a target user-specific dataset, and provide data basis for subsequent model optimization.

[0053] Specifically, if the target user's head center point (5,9) and elbow (2,5) are detected, the corresponding confidence levels are 0.2 and 0.5, respectively. If the confidence level of a key point (such as the elbow) is extremely low (e.g., 0.2), it indicates that the area may be obstructed or the lighting is poor. The system will filter out this abnormal data to prevent false alarms, or prompt the user to "adjust the camera angle to avoid limb obstruction," thereby ensuring the accuracy of the monitoring. Simultaneously, after the target user enters a resting state, the system performs small-batch fine-tuning or incremental learning on the collected data, using an optimization algorithm to update the model parameters, gradually adapting it to the individual characteristics of the target user.

[0054] In this embodiment, by analyzing target user feedback data and pose data, a reinforcement learning strategy is used to optimize the detection and feedback mechanism. A variant of Q-learning, such as a Deep Q-Network (DQN), is used to learn the optimal policy among different poses and feedback. As more feedback data becomes available, the model is periodically trained online, and the model parameters are fine-tuned to improve robustness.

[0055] In this embodiment, based on the analysis results of the current model, the suggested duration and difficulty level for the next meditation session are adjusted. Simultaneously, targeted video content is recommended to compensate for the user's weak posture areas. For example, if the system analyzes the target user's data from the past week and finds that the "left shoulder key point" frequently sinks during the later stages of meditation (resulting in a low posture score), the system will determine that the target user's left shoulder is prone to fatigue or involuntary tilting. Therefore, the report will provide a video of "relaxation exercises for the left shoulder muscles" and suggest shortening the suggested duration for the next meditation session to avoid injury caused by posture distortion.

[0056] In this embodiment, the system periodically evaluates the improvement effect of the real-time meditation posture monitoring model, generates a usage report, and allows target users to provide feedback to help further adjust personalized suggestions and overall system performance.

[0057] In this embodiment, intelligent reminders and check-in are also included. Regular reminders: The system sends reminder notifications via mobile application, email, or SMS based on the meditation plan set by the target user, ensuring the target user meditates on time. Automatic check-in: After meditation, the system automatically records data such as meditation time, duration, and posture quality. The target user can also manually confirm the completion of meditation.

[0058] In this embodiment, data analysis and feedback are also included. Data Collection: The system collects detailed data for each meditation session, including time, posture, breathing rate, and heart rate. Data Analysis: Using data analysis and machine learning techniques, the system processes the collected data to identify the target user's meditation patterns and health trends. Health Report: The system generates a detailed health report, including meditation frequency, posture improvement suggestions, and changes in breathing and heart rate. Personalized Suggestions: Based on the analysis results, the system provides personalized health management suggestions, such as adjusting meditation time, improving posture, and increasing meditation frequency.

[0059] The method for managing meditation behavior in cases of diabetes complications provided in this embodiment enables real-time adjustment and optimization of the meditation behavior of target users. At the same time, it adjusts and optimizes the parameters of the real-time meditation posture monitoring model based on the correct data of the target users, thereby achieving personalized management and ensuring the health management of the meditation behavior of target users.

[0060] In another embodiment of the present invention, the target users are multiple, and the real-time acquisition of meditation behavior data of the target users includes: Real-time acquisition of meditation behavior data from multiple target users.

[0061] Specifically, this embodiment can implement facial recognition and voice recognition. Facial recognition: Before starting meditation, the target user's face is recognized via a camera. The system uses deep learning algorithms (such as convolutional neural networks, CNN) to match the target user's facial features, ensuring that the meditator is the registered user. Voice recognition: The target user can submit a meditation request via voice, and the system uses natural language processing (NLP) technology to recognize and verify the user's voice.

[0062] Furthermore, the meditation behavior management method for diabetic complications provided in this embodiment can simultaneously acquire meditation behavior data from multiple target users and optimize and adjust the meditation behavior of each target user in parallel. This achieves diversification of meditation behavior management.

[0063] Secondly, the present invention also provides a meditation behavior management system for diabetic complications applied to any of the above-described methods for managing meditation behavior in diabetic complications, comprising a real-time meditation posture monitoring module and a real-time feedback correction module. The real-time meditation posture monitoring module is connected to the real-time feedback correction module. The real-time meditation posture monitoring module is used to monitor the meditation behavior of a target user in real time and obtain the analysis results and adjustment plan of the target user's meditation behavior. The real-time feedback correction module is used to provide real-time feedback of the analysis results and adjustment plan of the meditation behavior to the target user.

[0064] In this embodiment, a meditation behavior management system for diabetic complications is provided to implement the meditation behavior management method for diabetic complications described in the above embodiment. The system includes a real-time meditation posture monitoring module and a real-time feedback correction module. The following is the collaborative process of each module: The real-time meditation posture monitoring module uses cameras and sensors to monitor the meditation posture of target users in real time. It analyzes and identifies the posture using a posture evaluation algorithm to determine if the user is meditating correctly and dynamically adjusts the user's meditation time based on the analysis results. The module mainly consists of three processes: model configuration, data preparation, and training optimization. First, in model configuration, the real-time meditation posture monitoring model P is OpenPose. OpenPose is suitable for multi-person posture estimation, striking a balance between real-time performance and accuracy. The input image size is typically set to (1, 3, 256, 256), adjusted according to computational resources and accuracy requirements. The model generates heatmaps and offset maps to represent the probability distribution of keypoint locations. A typical output consists of 17 keypoints (e.g., in COCO format), covering various important parts of the human body. During detection, mean squared error (MSE) loss is used to minimize the difference between predicted locations and ground truth annotations, ensuring high-quality keypoint detection. Training data comes from the COCO keypoint dataset or the MPII dataset, which provides comprehensive annotations. For meditation posture-specific recognition, proprietary data for that posture is supplemented. Data augmentation is a key step in improving the generalization ability of a model. It diversifies the input samples through random rotation, scaling, and flipping. This enhances adaptability to different perspectives and environmental changes, and also improves the robustness of detection.

[0065] During pose detection, the system identifies the user's "left knee" at pixel (200, 450) and "left ankle" at (205, 580). These specific values ​​constitute the positional information, used for subsequent calculations of joint angles or comparison with standard poses. The COCO keypoint dataset is typically used, with 17 keypoints including: nose, left / right eye, left / right ear, left / right shoulder, left / right elbow, left / right wrist, left / right hip, left / right knee, and left / right ankle.

[0066] In this embodiment, to accelerate the training process, especially in deep network architectures, weights pre-trained from ImageNet are used to initialize the model, which is highly effective in improving initial performance and accelerating convergence. During training, Adam is selected as the optimizer, and backpropagation is used to update the model weights. A learning rate decay strategy is adopted to improve efficiency and accuracy during training. The model training lasts 50-100 epochs. During this process, metrics such as PCK (Percentage of Correct Keypoints) or AP (Average Precision) are used to measure the model's performance, ensuring that the model can accurately identify keypoints when detecting different poses. Once the model requires fine-tuning for specific action recognition, further optimization relies on specific task metrics to ensure high-precision detection.

[0067] It should be noted that this embodiment uses weights, which are applied during the model inference (pre-training) phase. When the camera captures a real-time video stream and inputs it into the system, the system loads the pre-trained weight parameters for forward propagation calculation, thereby predicting the location of key human body points in the current image in real time.

[0068] Real-time feedback correction module: The camera continuously captures the target user's posture video stream, uses a configured model to estimate the target user's key point positions in real time, and compares the target user's posture with the standard posture in the database. If the deviation of the user's posture from the standard exceeds a preset threshold, the posture is considered incorrect. When a deviation occurs, the system prompts the user to adjust their posture through audio, video, or interface. After several consecutive detection cycles (e.g., 30 seconds), if the target user's posture remains incorrect, the system dynamically adjusts the meditation time. If the initial meditation time is 15 minutes, it can be shortened to 10 minutes to reduce the potential impact of improper posture. At the same time, it switches to a more detailed meditation video to guide the target user on how to correctly adjust their posture. Personalized feedback is generated based on historical posture data. For example, if the user's shoulder position is frequently incorrect, relevant correction suggestions or practice videos can be provided. It should be noted that this real-time feedback correction module is usually based on logical judgment rather than deep learning training. It receives key point position information output by the real-time meditation posture monitoring module, determines whether the posture is acceptable based on preset thresholds (e.g., angle deviation > 15 degrees), and triggers the corresponding feedback mechanism (e.g., playing guidance videos).

[0069] The meditation behavior management system for diabetic complications provided by this invention is used to monitor and improve the meditation behavior of patients with diabetic peripheral neuropathy. It achieves intelligent recognition and management of meditation behavior through a trained model, dynamically adjusts the meditation duration, and provides real-time adjustment suggestions for the meditation behavior of target users, thereby achieving the purpose of personalized health guidance.

[0070] Thirdly, the present invention provides a device for managing meditation behavior in patients with diabetic complications, the device comprising: The acquisition module is used to acquire meditation behavior data of target users in real time; The meditation posture real-time monitoring module is used to transmit the meditation behavior data to the pre-trained meditation posture real-time monitoring model to obtain key point location information. The determination module is used to determine whether the meditation behavior is standard based on the key point location information and the preset standard posture library data. The real-time adjustment and correction module is used to adjust and correct the meditation behavior when it is determined that the meditation behavior is not standard; wherein, the real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

[0071] The meditation behavior management device for diabetic complications provided by the present invention is used to monitor and improve the meditation behavior of patients with diabetic peripheral neuropathy. It achieves intelligent recognition and management of meditation behavior through a trained model, dynamically adjusts the meditation duration, and provides real-time adjustment suggestions for the meditation behavior of target users, thereby achieving the purpose of personalized health guidance.

[0072] The communication device in this embodiment is implemented on the same principle as the method described above, and will not be described in detail here.

[0073] Figure 3 This is a schematic diagram of the physical structure of the electronic device provided in the embodiments of the present invention, such as... Figure 3 As shown, the present invention provides an electronic device, including: a processor 701, a memory 702, and a bus 703; The processor 701 and the memory 702 communicate with each other via the bus 703. The processor 701 is used to call program instructions in the memory 702 to execute the methods provided in the above-described method embodiments, such as: the method is applied to a master node and multiple slave nodes, the method includes: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point position information; determining whether the meditation behavior is standard based on the key point position information and a preset standard posture library; and adjusting and correcting the meditation behavior if it is determined that the meditation behavior is not standard; wherein, the real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

[0074] This invention provides a non-transitory computer-readable storage medium storing computer instructions that cause the computer to execute the methods provided in the above-described method embodiments. These instructions include, for example: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point location information; determining whether the meditation behavior is standard based on the key point location information and a preset standard posture library; and adjusting and correcting the meditation behavior if it is determined to be non-standard. The real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

[0075] This invention also provides a computer program product, comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, enable the computer to perform the methods provided in the above embodiments. The method includes: acquiring meditation behavior data of a target user in real time; transmitting the meditation behavior data to a pre-trained real-time meditation posture monitoring model to obtain key point location information; determining whether the meditation behavior is standard based on the key point location information and a preset standard posture library; and adjusting and correcting the meditation behavior if it is determined to be non-standard. The real-time meditation posture monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

[0076] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for managing meditation behavior in patients with diabetic complications, used to monitor and improve meditation behavior in patients with diabetic peripheral neuropathy, characterized in that, include: Real-time acquisition of meditation behavior data from target users; The meditation behavior data is transmitted to a pre-trained real-time meditation posture monitoring model to obtain key point location information; Based on the key point location information and the preset standard posture library data, it is determined whether the meditation behavior is standard; If the meditation behavior is determined to be non-standard, the meditation behavior shall be adjusted and corrected. The meditation posture real-time monitoring model is obtained by training a basic posture detection model based on meditation behavior training data.

2. The method for managing meditative behavior in diabetic complications according to claim 1, characterized in that, The key point location information includes at least first joint point location information, second joint point location information, and third joint point location information. Determining whether the meditation behavior is standard based on the key point location information and preset standard posture database data includes: Based on the first joint position information and the first target joint position information in the standard knowledge base, the first joint deviation value is obtained; Using the law of cosines, the angle of the first joint is obtained based on the position information of the first joint, the position information of the second joint, and the position information of the third joint. Based on the deviation value of the first joint point and the angle of the first joint point, it is determined whether the meditation behavior is standard.

3. The method for managing meditative behavior in diabetic complications according to claim 2, characterized in that, The step of determining whether the meditation behavior is standard based on the first joint point deviation value and the first joint point angle includes: The first joint deviation score is obtained based on the first joint deviation value and the first mapping table; Based on the first joint point angle and the second mapping table, the deviation score of the first joint point angle is obtained; The comprehensive score of the first joint is obtained based on the first joint deviation score, the first joint angle deviation score, and the first angle weight value. When the comprehensive score of the first joint point is less than the standard score, the meditation behavior is determined to be non-standard. When the comprehensive score of the first key point is greater than or equal to the standard score, the standard for the meditation behavior is determined. The first mapping table is a table showing the correspondence between joint deviation values ​​and position deviation scores, and the second mapping table is a table showing the correspondence between joint angles and angle deviation scores.

4. The method for managing meditative behavior in diabetic complications according to claim 1, characterized in that, The step of adjusting and correcting the meditation behavior when it is determined to be non-standard includes: If the meditation behavior is determined to be non-standard, information is provided to the target user through visual or audio signals, indicating the parts of the meditation behavior that need to be adjusted and the adjustment method.

5. The method for managing meditative behavior in diabetic complications according to claim 1, characterized in that, If the meditation behavior is determined to be non-standard, and the meditation behavior is adjusted and corrected, the method further includes: Real-time recording of the key point location information and the target user's characteristic parameters; The meditation posture real-time monitoring model is optimized based on the key point location information and the target user's feature parameters.

6. The method for managing meditative behavior in diabetic complications according to claim 1, characterized in that, The method also includes: The analysis results are obtained by analyzing the meditation behavior data of the target users. Based on the analysis results, corresponding solutions are formulated for the problems existing in meditation behavior.

7. The method for managing meditative behavior in diabetic complications according to claim 1, characterized in that, The target users are multiple, and the real-time acquisition of meditation behavior data of the target users includes: Real-time acquisition of meditation behavior data from multiple target users.

8. A diabetic complication meditation behavior management system applied to the meditation behavior management method for diabetic complications according to any one of claims 1-7, characterized in that, It includes a real-time meditation posture monitoring module and a real-time feedback correction module, wherein the real-time meditation posture monitoring module is connected to the real-time feedback correction module. The meditation posture real-time monitoring module is used to monitor the meditation behavior of the target user in real time, and obtain the analysis results and adjustment plan of the meditation behavior of the target user. The real-time feedback correction module is used to provide the analysis results and adjustment plan of the meditation behavior to the target user in real time.

9. An electronic device, characterized in that, include: Processor, memory, and bus, among which, The processor and the memory communicate with each other via the bus; The memory stores program instructions that can be executed by the processor, which can invoke the program instructions to perform the steps of the meditation behavior management method for diabetic complications as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions that cause the computer to perform the steps of the method for managing meditation behavior in cases of any one of claims 1 to 7.

Images