Individual transverse exercise amount feedback system and method in regionalized scene

Abstract

The invention provides an individual transverse motion amount feedback system and method in a regionalized scene, and the system comprises an individual motion tracking subsystem which is used for carrying out the motion capture and recognition of an individual in the regionalized scene through the combination of a target detection and time sequence analysis technology, and obtaining the motion tracking data of the individual; and the visual motion feedback subsystem is used for carrying out individual transverse motion amount evaluation based on the individual motion tracking data and carrying out visual feedback on an evaluation result. According to the technical scheme, the user can accurately obtain the exercise feedback in the regional scene, and optimization and health management of the individual exercise amount are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence (AI) health monitoring, and particularly relates to an individual horizontal exercise amount feedback system and method in a regionalized scenario. Background Art

[0002] With the rapid development of artificial intelligence (AI) technology, computer vision technology has been increasingly widely applied in various fields, especially in individual behavior recognition and motion analysis, and remarkable achievements have been made. Existing technologies are mainly applied in the fields of intelligent health monitoring, motion analysis, intelligent fitness, etc. However, the existing exercise amount feedback systems still have the following problems and deficiencies:

[0003] 1. The accuracy problem of individual action tracking

[0004] The mainstream image processing method of the current technology is to dynamically adjust the brightness and contrast of video frames by analyzing the scene light distribution in real time, combining histogram equalization and Retinex theory, so as to reduce the interference of sudden light changes on target detection. By combining RGB images with depth information (obtained through a stereo vision camera), a three-dimensional space feature map is constructed to enhance the target separation ability in complex backgrounds. Mask R-CNN is used for instance segmentation, and the optical flow method is combined to predict the motion trajectory of occluded targets, and the LSTM network (Long Short-Term Memory) is used to interpolate and complete the time-series occlusion segments to reduce missed detections and false detections. Based on the Gaussian mixture model (GMM), the background pixel distribution is updated in real time, and combined with motion area detection (such as the ViBe algorithm), dynamic backgrounds and individual actions are distinguished. For the exercise amount feedback system based on wearable devices, it is easy to interfere with the performance of athletes and cannot effectively feedback the exercise amount in real time.

[0005] 2. The limitations of data feedback

[0006] Most of the existing motion feedback platforms focus on the statistical analysis of users' personal exercise data, such as the number of steps, calorie consumption, exercise duration, etc., but lack multi-dimensional display of exercise amount and real-time interaction. This makes it difficult for users to obtain comprehensive feedback during exercise, unable to effectively perceive their own exercise status, and thus affects individual motion control and training effects.

[0007] 3. The application limitations of deep learning and image processing technologies

[0008] Although deep learning and image processing technologies have great potential in individual action recognition and analysis, traditional image recognition-based exercise volume feedback systems still face problems such as high system computing resource requirements and insufficient real-time feedback capabilities. Existing systems cannot provide efficient and accurate individual action recognition and real-time feedback in regional complex scenarios, and lack a personalized exercise volume display and visual feedback platform for different individuals.

[0009] 4. Insufficiency in individual exercise volume perception

[0010] Most traditional exercise volume monitoring systems use a unified standard to evaluate individual exercise volume, failing to consider individual physical conditions, exercise goals, and exercise habits. This results in low personalization and accuracy of exercise feedback, making it difficult to stimulate individuals' true perception and control of exercise volume. Summary of the Invention

[0011] To solve the problems existing in the prior art, the present invention provides an individual lateral exercise volume feedback system and system in a regional scenario, realizing precise tracking and real-time feedback of individual movements.

[0012] An individual lateral exercise volume feedback system in a regional scenario, the system includes:

[0013] An individual action tracking subsystem, configured to combine object detection and time series analysis technologies to capture and recognize the actions of an individual in a regional scenario, and obtain individual action tracking data;

[0014] A visual exercise feedback subsystem, configured to evaluate the individual lateral exercise volume based on the individual action tracking data and provide visual feedback on the evaluation results.

[0015] Preferably, the individual action tracking subsystem includes:

[0016] A multi-camera stereo vision module, configured to collect video information from all angles of a regional scenario in real time;

[0017] An action and event recognition module, configured to perform object detection on the video information, recognize individual actions, and perform time series processing on the individual actions using an LSTM network to obtain individual action tracking data.

[0018] A real-time feedback and decision-making assistance module, configured to feedback the individual action tracking data based on the inference ability of a deep learning model and provide a decision-making determination for the individual action tracking data;

[0019] A data processing and storage module, configured to encrypt and store the video information and the decision-making determination.

[0020] Preferably, the visual exercise feedback subsystem includes:

[0021] An individual motion data acquisition module for collecting and receiving the individual motion tracking data;

[0022] A motion quantity calculation and analysis module for calculating and analyzing the individual motion tracking data based on a motion quantity evaluation model to obtain a motion quantity evaluation result;

[0023] A personalized feedback display module for visualizing the motion quantity evaluation result;

[0024] A target setting and progress tracking module for a user to set a motion target and real-time track the progress of target completion;

[0025] A social interaction and community feedback module for social interaction and sharing of motion achievements among users;

[0026] A data report and health advice module for synthesizing the motion quantity evaluation results within a preset time period of a user, generating a motion report, and providing personalized health advice;

[0027] A user-defined setting module for a user to adjust system functions and display modes.

[0028] Preferably, in the motion quantity calculation and analysis module, the construction process of the motion quantity evaluation model includes:

[0029] Standardizing the number of steps, average acceleration, and heart rate in the individual motion tracking data to obtain standardized motion data;

[0030] Based on the standardized motion data, obtaining a basic motion quantity score by using a weighted fusion formula;

[0031] Introducing correction factors based on the individual physiological characteristics of the user, where the correction factors include a body mass index correction factor and a physical fitness level correction factor;

[0032] Processing the standardized motion data by using a pre-trained LSTM-attention neural network model to obtain an intensity classification label;

[0033] Based on the basic motion quantity score, the correction factors, and the intensity classification label, obtaining the motion quantity evaluation result and completing the construction of the motion quantity evaluation model.

[0034] Preferably, the target setting and progress tracking module includes:

[0035] A target setting unit for a user to set a motion target;

[0036] A progress display and prediction unit for displaying the completion degree of the moving target and predicting and reminding the target progress based on the user's exercise habits and performance; wherein, the completion degree of the moving target is calculated based on the exercise amount evaluation result and the moving target.

[0037] Preferably, the exercise report includes exercise amount, exercise intensity, calorie consumption, and exercise frequency.

[0038] Preferably, the personalized health advice is generated using a rule engine.

[0039] The present invention also provides an individual lateral exercise amount feedback method in a regionalized scenario, applying the system, including:

[0040] Combining object detection and time series analysis techniques to capture and identify the actions of an individual in a regionalized scenario to obtain individual action tracking data;

[0041] Evaluating the individual's lateral exercise amount based on the individual action tracking data and visually feedbacking the evaluation result.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows: By combining deep learning, computer vision, and multi-camera stereo vision technologies, the present invention realizes precise tracking and real-time feedback of individual movements. Through the visualization platform, users can obtain real-time personalized exercise amount feedback, improving the efficiency of exercise control and health management. The present invention has broad application prospects, especially in the fields of intelligent fitness, sports training, and health monitoring. Brief Description of the Drawings

[0043] In order to more clearly illustrate the technical solutions of the present invention, the accompanying drawings required for use in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other accompanying drawings without creative efforts.

[0044] Figure 1 It is a schematic structural diagram of the individual lateral exercise amount feedback system in the regionalized scenario of the embodiment of the present invention;

[0045] Figure 2 It is a flowchart of the image processing technology in the embodiment of the present invention;

[0046] Figure 3 It is a flowchart of the low-latency data transmission technology in the embodiment of the present invention. Detailed Embodiments

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Embodiment 1

[0050] As Figure 1 shown, a system for feedback of individual horizontal movement amount in a regionalized scenario, the system includes: an individual action tracking subsystem and a visual motion feedback subsystem.

[0051] The individual action tracking subsystem is used to capture and identify the actions of individuals in a regionalized scenario by combining object detection and time series analysis technologies, and obtain individual action tracking data:

[0052] (1) The individual action tracking data specifically includes:

[0053] Spatial data: individual position coordinates (X, Y, Z), movement speed, average acceleration (calculated from the displacement between consecutive frames).

[0054] Posture data: joint angles (such as the coordinates of 17 key points output by OpenPose), action type labels (such as running, jumping).

[0055] Time series features: action duration, frequency, and periodicity (hidden state vectors extracted by the LSTM network).

[0056] (2) Data source of the visualization subsystem:

[0057] The main data received by the visual motion feedback subsystem is the tracking data output by the individual action tracking subsystem, but the heart rate data needs to be supplemented in the following ways:

[0058] Remote Photoplethysmography (rPPG): Extract weak skin color change signals from facial area videos, and invert the heart rate value through Fourier transform and filtering technologies to achieve non-contact heart rate monitoring.

[0059] Optional wearable device integration: Support Bluetooth connection to third-party heart rate belts or smart bracelets, and synchronize heart rate data to the system through the API (the data source needs to be clearly stated in the user agreement).

[0060] A further implementation is that the individual motion tracking subsystem includes:

[0061] A multi-camera stereo vision module for real-time acquisition of video information from various angles of a regionalized scene; in this embodiment, a basketball court is used as the regionalized scene, and multiple cameras are arranged in the basketball court to achieve all-round motion capture of athletes and referees. The video of the athlete's face area is captured by multiple cameras, and the rPPG algorithm is used to extract the heart rate; at the same time, it supports synchronizing data with a smart bracelet (such as Polar H10) through the Bluetooth BLE protocol, and the data format is JSON, including timestamps and heart rates. Through reasonable camera angle arrangements, ensure sufficient field of view coverage and be able to clearly record every detail during the game. This system can obtain video information from all angles of the court in real time, providing comprehensive game scene data support. Specifically, the present invention adopts a data-driven method, combines deep learning and computer vision technologies, and performs data acquisition and processing in the following ways:

[0062] Data acquisition: High-resolution cameras are used to acquire multi-angle video data to ensure complete coverage of the motion scene.

[0063] Data cleaning and filtering: The collected raw data is denoised and preprocessed to improve data quality.

[0064] Data augmentation: Data augmentation techniques are used to expand the training set and enhance the generalization ability of the system.

[0065] An action and event recognition module (action recognition module) for performing object detection (such as YOLO, SSD) on the video information, identifying individual actions, and using an LSTM network to perform temporal processing on the individual actions to obtain individual motion tracking data. Specifically, combining object detection techniques (such as YOLO, SSD) and temporal analysis techniques (such as LSTM networks), the actions of athletes in a basketball game are analyzed and judged in real time. Through a deep learning model, various violations in the game can be automatically recognized, such as offensive timeouts, backcourt violations, etc., and accurate behavior recognition and event determination are provided. This technology monitors the game process in real time by analyzing video frames.

[0066] In this embodiment, the present invention combines deep learning and image processing technologies to achieve high-precision and dead-angle-free individual motion capture through a multi-camera stereo vision system. As Figure 2 shown, the main features in this aspect of the present invention are: The scene and the athlete's motion state are modeled in real time through a real-time motion acquisition system, key parameters of the motion state are extracted, and the extracted motion key parameters are analyzed in real time using relevant algorithms. Compared with traditional technologies, the system performs more stably in complex environments and can accurately identify individual motions in real time. The technical implementation of the real-time motion acquisition system includes:

[0067] (I) Real-time object detection: The YOLOv5 model is adopted, and a detection speed of more than 30 FPS is achieved through a lightweight network structure (such as the Focus layer and CSP module), and individual bounding boxes and class labels are output.

[0068] (II) Pose estimation and motion modeling: Human key points are extracted based on HRNet (High-Resolution Network), and the Kalman filter is combined to smooth the motion trajectory, and an individual motion state vector (including speed, acceleration, joint angular velocity) is constructed.

[0069] (III) Key parameter extraction algorithm:

[0070] 1. Calculation of exercise intensity: According to the average acceleration and joint movement amplitude, through the formula (θ max is the maximum change in joint angle).

[0071] 2. Event recognition: A two-stream network (spatial stream + temporal stream) is adopted to fuse RGB frames and optical flow maps to identify specific action events (such as shooting, violation).

[0072] (IV) Low-latency data transmission: Real-time transmission of video streams is realized based on the WebRTC protocol to ensure that the end-to-end latency is less than 200 ms, meeting the real-time requirements of the referee's penalty. The real-time feedback and decision-making assistance module is used to feedback individual action tracking data and provide decision-making judgments on individual action tracking data based on the inference ability of the deep learning model; the specific technologies for the specific deep learning model to assist the referee's penalty include: 1. Key event detection model: Train a temporal action classification model based on 3D CNN, with the input being 10 consecutive frames of RGB + optical flow maps, and output the probability of violation actions (such as traveling, double dribbling). 2. Real-time inference framework: TensorRT is used to quantize and accelerate the model, and it is deployed on edge computing devices (such as NVIDIA Jetson AGX) to achieve an inference latency of less than 50 ms. 3. Rule engine and decision tree: If the model output violation probability > 90%, trigger a voice prompt "Suggested penalty: traveling violation"; combined with historical penalty data (stored in the database), mine penalty association rules through the Apriori algorithm, and dynamically adjust the threshold to reduce false alarms. This system can support the referee to make more scientific and reasonable penalty decisions during the game and automatically record relevant penalty data and abnormal events. This not only improves the accuracy of the referee's penalty but also ensures the fairness and transparency of the game.

[0073] The data processing and storage module is used to encrypt and store video information and decision-making. Specifically, this module is responsible for the storage, management and post-analysis of video data and referee penalty information. All collected data is encrypted and stored, providing good privacy protection and ensuring that the data is long-term valid and traceable. This module supports the mining and analysis of historical data, and can provide data support for game optimization and referee training. Low-latency data transmission technology is used to ensure timely feedback of sports data. Figure 3 shown.

[0074] The present invention compresses the collected original video stream through Raspberry Pi, and then uses the RTSP protocol to upload the video stream to the sideline server through 10Gbps Ethernet in real time after adding a timestamp, thereby ensuring the high efficiency and real-time performance of the transmission. Even if there is a transmission delay caused by network fluctuations or workload, the timestamp can ensure the synchronization of the video stream and the accuracy of the analysis.

[0075] The visual motion feedback subsystem is used to evaluate the individual lateral motion based on the individual motion tracking data and to provide visual feedback of the evaluation results.

[0076] In a further embodiment, the visual motion feedback subsystem comprises:

[0077] The individual motion data acquisition module is used to collect and receive individual motion tracking data; through the combination of computer vision technology and deep learning models, the collected data will be accurately transmitted to the motion analysis module for further processing.

[0078] The exercise volume calculation and analysis module is used to calculate and analyze individual motion tracking data based on the exercise volume assessment model to obtain exercise volume assessment results; specifically, this module is responsible for calculating and analyzing the collected exercise data, and the main output includes but is not limited to: exercise volume, exercise intensity, calorie consumption, exercise type, etc. Through adaptive algorithms and machine learning models, this module can analyze the exercise of different individuals and provide personalized feedback.

[0079] A further implementation method is that, in the exercise volume calculation and analysis module, the process of constructing the exercise volume evaluation model includes:

[0080] The number of steps (steps), average acceleration (acc) and heart rate (hr) in the individual motion tracking data are standardized to obtain standardized motion data; specifically, motion data from different sources are processed uniformly through a standardized data format interface to ensure compatibility and scalability.

[0081] Step number normalization:

[0082] (Based on every thousand steps as the benchmark unit)

[0083] Acceleration normalization:

[0084]

[0085] Heart rate normalization:

[0086] (RHR = 60 is the resting heart rate, MHR = 220 - age is the maximum heart rate)

[0087] Based on the normalized exercise data, a basic exercise volume score is obtained using a weighted fusion formula; specifically, a basic exercise volume score is generated using the weighted fusion formula: (Weight coefficients α = 0.4; β = 0.3; γ = 0.3, which can be dynamically adjusted through the user's historical data).

[0088] A correction factor is introduced based on the user's individual physiological characteristics, where the correction factor includes a body mass index correction factor and a physical fitness level correction factor; noise filtering is performed on the collected raw data to ensure the accuracy of the analysis results. Specifically, body mass index correction:

[0089] (Enhance the sensitivity of the exercise volume assessment for overweight users)

[0090] Physical fitness level correction:

[0091] C fitness = 1 - 0.1×(L fitness - 1)L fitness ∈[1,5] (Physical fitness level calculated based on historical exercise data).

[0092] The pre-trained LSTM-attention neural network structure is as follows:

[0093] (1) Input layer: The time series of normalized exercise data (steps, acceleration, heart rate), with a time window of 60 seconds and a sampling rate of 10Hz (input dimension: 600×3).

[0094] (2) LSTM layer: A bidirectional LSTM structure, with 128 units in each layer, outputting the hidden state

[0095] (3) Attention mechanism: 1. Calculate the attention weights: a t = Softmax(W a ·h t + b a ). 2. Context vector:

[0096] (4) Classification layer: Fully connected layer (64 units) + Softmax to output intensity classification labels (low / medium / high intensity).

[0097] (5) Training strategy: Use Focal Loss to solve the class imbalance problem, and the pre-training dataset is COCO-Activity (100,000 labeled motion sequences).

[0098] Use a pre-trained LSTM-attention neural network model to process the standardized motion data to obtain intensity classification labels; specifically, use a pre-trained LSTM-attention neural network model to input the time series data Output intensity classification label: I label = Classifier(sep t-n:t )(The classification results include: 0 - low intensity, 1 - moderate, 2 - high intensity).

[0099] Based on the basic exercise score, correction factor, and intensity classification label, obtain the exercise amount evaluation result and complete the construction of the exercise amount evaluation model. Generate the final score by combining the correction factor and intensity label:

[0100] E final = E base × C BMI × C fitness × (1 + 0.2 × I label )

[0101] Personalized feedback display module, used to visualize the exercise amount evaluation result; this module is specifically responsible for displaying the user's motion data feedback in a graphical and intuitive way. Through various visualization methods (such as bar charts, pie charts, line charts, etc.), enable users to intuitively understand and master the exercise status. Specifically, use a variety of chart forms to display data in various dimensions such as exercise amount, intensity, and duration. For example, use a line chart to display the fluctuation of exercise intensity and a bar chart to display the change in daily steps.

[0102] Dynamic feedback: Dynamically update the exercise feedback information according to the user's real-time exercise performance to ensure its synchronization with the actual exercise status.

[0103] Goal setting and progress tracking module, used for users to set exercise goals and real-time track the progress of goal completion; this module allows users to set their own exercise goals and real-time track the progress of goal completion. Whether it is the number of steps, calories burned, or exercise duration, etc., can be personalized settings to help users conduct quantitative management. Customize the feedback display method according to the needs of different users (such as exercise intensity, calorie consumption, exercise goal completion degree, etc.).

[0104] A further implementation manner lies in that the goal setting and progress tracking module includes:

[0105] A target setting unit for a user to set a motion target;

[0106] A progress display and prediction unit for displaying the completion degree of a motion target and predicting and reminding the target progress based on the user's motion habits and performance; wherein, the completion degree of the motion target is calculated based on the motion amount evaluation result and the motion target. The progress of target completion is displayed in real time by means of a progress bar, a circular graph, etc., and the user is reminded whether the expected target is achieved during the motion process, helping the user to optimize the training plan. The process of calculating the target completion degree:

[0107] (E target (Automatically converted according to the motion target set by the user)

[0108] A social interaction and community feedback module for social interaction between users and sharing of motion achievements; this module provides a social interaction platform for users. Users can share their motion achievements (such as steps, motion duration, etc.) with other exercisers through this platform, participate in motion challenges or activities, enhancing the fun and sociality of the motion. Users can post motion problems or suggestions to other members and conduct interaction and experience exchange.

[0109] A data report and health advice module for generating a motion report by integrating the motion amount evaluation results within a preset time period of the user and providing personalized health advice; this module integrates the motion data of the user for a period of time to generate a motion report and provides personalized health advice according to the report, helping the user to achieve health goals while optimizing their motion plan. A further implementation manner is that the motion report includes motion amount, motion intensity, calorie consumption, and motion frequency.

[0110] A further implementation manner is that the personalized health advice is generated by a rule engine. According to the user's health condition, motion data, and target, personalized health advice is pushed, including adjustment of motion intensity, diet advice, recovery time, etc. In this embodiment, a rule engine is used to generate natural language advice:

[0111]

[0112] A user-defined setting module for a user to adjust the system functions and display modes. This module allows the user to adjust the functions and display modes of the platform according to their own needs and preferences, thereby providing a personalized user experience.

[0113] In practical applications, through multiple simulation tests and evaluation of actual motion data, the present invention mainly considers the following indicators:

[0114] Recognition accuracy: By comparing with manually calibrated data, evaluate the accuracy of action recognition and determination.

[0115] Real-time performance: Evaluate the response speed of the system to ensure real-time data transmission and feedback.

[0116] Stability and reliability: Test the performance of the system in different scenarios and different motion environments to ensure the stable operation of the system.

[0117] Embodiment 2

[0118] The present invention also provides a method for feedback of individual lateral movement amount in a regionalized scenario, which is applied to a system, including:

[0119] Combining object detection and time series analysis technologies, capture and recognize the actions of individuals in a regionalized scenario to obtain individual action tracking data;

[0120] Based on the individual action tracking data, evaluate the individual lateral movement amount and visually feedback the evaluation result.

[0121] The above-described embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A system for feedback of individual lateral motion in a regionalized scenario, characterized in that: The system comprises: The individual motion tracking subsystem is used to combine target detection and time series analysis technology to capture and identify the motion of individuals in regionalized scenes and obtain individual motion tracking data; The visual motion feedback subsystem is used to evaluate the individual lateral motion amount based on the individual motion tracking data, and to provide visual feedback of the evaluation results.

2. The system according to claim 1, characterized in that The individual motion tracking subsystem comprises: Multi-camera stereo vision module, used to collect video information from all angles of regional scenes in real time; The action and event recognition module is used to perform target detection on the video information, identify individual actions, and use an LSTM network to perform time series processing on the individual actions to obtain individual action tracking data; A real-time feedback and decision-making assistance module, which is used to feed back the individual motion tracking data based on the reasoning ability of the deep learning model and provide a determination decision of the individual motion tracking data; The data processing and storage module is used to encrypt and store the video information and the determination decision.

3. The system according to claim 1, characterized in that The visual motion feedback subsystem comprises: An individual motion data acquisition module, used to acquire and receive the individual motion tracking data; An exercise volume calculation and analysis module, used to calculate and analyze the individual motion tracking data based on an exercise volume evaluation model to obtain an exercise volume evaluation result; A personalized feedback display module, used to visualize the exercise volume assessment results; The goal setting and progress tracking module is used for users to set sports goals and track the progress of goal completion in real time; Social interaction and community feedback module, used for social interaction and sharing of sports results between users; The data report and health advice module is used to integrate the user's exercise assessment results within a preset time period, generate exercise reports, and provide personalized health advice; User-defined settings module is used for users to adjust system functions and display methods.

4. The system according to claim 3, characterized in that In the exercise volume calculation and analysis module, the construction process of the exercise volume evaluation model includes: Standardizing the number of steps, acceleration mean, and heart rate in the individual motion tracking data to obtain standardized motion data; Based on the standardized motion data, a weighted fusion formula is used to obtain a basic motion score; Introducing correction factors based on individual physiological characteristics of users, wherein the correction factors include body mass index correction factors and physical fitness level correction factors; Processing the standardized motion data using a pre-trained LSTM-attention neural network model to obtain intensity graded labels; Based on the basic exercise volume score, the correction factor and the intensity grading label, the exercise volume evaluation result is obtained to complete the construction of the exercise volume evaluation model.

5. The system according to claim 3, characterized in that The goal setting and progress tracking module includes: A target setting unit, used for users to set sports targets; The progress display and prediction unit is used to display the completion of the exercise goal and to predict and remind the user of the goal progress based on the user's exercise habits and performance; wherein the completion of the exercise goal is calculated based on the exercise volume evaluation result and the exercise goal.

6. The system according to claim 3, characterized in that The exercise report includes exercise volume, exercise intensity, calorie consumption and exercise frequency.

7. The system according to claim 3, characterized in that The personalized health advice is generated using a rule engine.

8. A method for feedback of individual lateral motion in a regionalized scene, using the system described in any one of claims 1 to 7, characterized in that: include: Combining target detection and time series analysis technology, the motion of individuals in regionalized scenes is captured and recognized to obtain individual motion tracking data; An individual lateral movement amount assessment is performed based on the individual motion tracking data, and the assessment result is visually fed back.

Images