User action execution standard degree scoring system of intelligent fitness equipment

Abstract

The invention relates to the technical field of image recognition, in particular to a user action execution standard degree scoring system for intelligent fitness equipment, and the system comprises the corresponding steps: determining a feature vector of a user action image based on a joint point pixel coordinate of the user action image, determining a feature vector difference degree between the user action image and the standard action image; utilizing the feature vector difference degree and the adjacent feature vector included angle to determine the joint motion difference degree of the user motion image in the target feature vector compared with the standard motion image; determining a deviation fluctuation coefficient and a motion deviation consistency degree of the user motion by using the joint motion difference degree, and obtaining a motion compensation index by using the deviation fluctuation coefficient and the motion deviation consistency degree; and obtaining an action execution standard degree score of the user by using the motion compensation index, and generating a targeted correction suggestion. By means of the technical scheme, the accuracy and reliability of action execution standard degree scoring are improved, and intelligent and personalized fitness guidance is achieved.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and more specifically to a user action execution standard scoring system for intelligent fitness equipment. Background Technology

[0002] With the increasing awareness of fitness among the general public, smart fitness equipment is becoming increasingly popular in homes and gyms. However, the average person has limited exercise knowledge, making it difficult to plan their training movements professionally and rationally, and to evaluate the quality of their execution. This makes it difficult to achieve good exercise results, and may even lead to physical injury, resulting in counterproductive outcomes. Therefore, how to realize the intelligent scoring function of fitness equipment and build a scientific and accurate movement evaluation system has become a key link in improving the safety and effectiveness of exercise for the general public.

[0003] In the process of scoring the standard of motion execution using smart fitness equipment, the score depends on the degree to which the user's actual movement matches the standard movement. When training with fitness equipment, users may experience transient errors due to momentary fluctuations in body posture, or systematic and persistent errors due to long-term deviations in their movement patterns. Furthermore, due to the body's compensatory mechanisms, temporary errors can be quickly self-regulated and confined to a single joint or localized area; however, for systematic errors, these compensatory mechanisms cannot effectively correct them, potentially exacerbating the transmission and accumulation of errors in multi-joint movements. This makes it difficult for existing scoring systems to accurately distinguish the nature and degree of error, thus reducing the accuracy and reliability of the standard of motion execution scoring and affecting users' effective evaluation and optimization of their training quality. Summary of the Invention

[0004] To address the technical problem that existing scoring systems struggle to accurately distinguish the nature and degree of error, resulting in low accuracy and reliability in scoring users' performance standards, this invention aims to provide a user performance standard scoring system for intelligent fitness equipment. The specific technical solution adopted is as follows: This invention provides a user movement execution standard scoring system for intelligent fitness equipment, the system comprising: The action recognition module is used to determine the feature vector of the user action image based on the key pixel coordinates of the user action image, and to determine the feature vector difference between the user action image and the standard action image of the corresponding frame. The deviation analysis module is used to determine the difference in joint motion between the user's motion image and the standard motion image in the target feature vector by using the difference in feature vectors and the angle between adjacent feature vectors; it uses the difference in joint motion to determine the deviation fluctuation coefficient and the degree of consistency of the user's motion deviation; and it uses the deviation fluctuation coefficient and the degree of consistency of the motion deviation to obtain the motion compensation index. The exercise evaluation module is used to obtain a score of the user's performance standard using the exercise compensation index and generate targeted correction suggestions.

[0005] Further, determining the feature vector of the user action image based on the keypoint pixel coordinates of the user action image includes: The feature vector sequence of the user action image is constructed by using the pixel coordinates of adjacent joints in the user action image according to the preset direction.

[0006] Further, determining the feature vector difference between the user's motion image and the standard motion image of the corresponding frame includes: Align the multi-frame user motion image sequence with the standard motion image sequence in time series to determine the standard motion image corresponding to the user motion image frame; Determine the cosine value of the target feature vector between the user action image and the standard action image of the corresponding frame, and use the cosine value to obtain the feature vector difference.

[0007] Furthermore, determining the joint motion difference between the user's motion image and the standard motion image in the target feature vector using the feature vector difference degree and the angle between adjacent feature vectors includes: The angle between adjacent feature vectors of a user is determined by using adjacent feature vectors in the user's action image, and the angle between adjacent feature vectors of a standard action image is determined by using adjacent feature vectors of a standard action image. By utilizing the feature vector difference, the angle between adjacent feature vectors of the user, and the angle between adjacent feature vectors of the standard, the joint motion difference of the user action image compared with the standard action image in the target feature vector is determined.

[0008] Further, determining the joint motion difference between the user's motion image and the standard motion image in the target feature vector using the feature vector difference degree, the angle between the user's adjacent feature vectors, and the angle between the standard adjacent feature vectors includes: By using the angle between adjacent feature vectors of the user and the angle between adjacent feature vectors of the standard, the difference in joint angles of the user action image and the standard action image in the target feature vector is determined. By utilizing feature vector difference and joint angle difference, the joint motion difference between the user motion image and the standard motion image in the target feature vector is determined.

[0009] Furthermore, the deviation fluctuation coefficient of the user's movement is determined using the joint motion variability, including: Determine the number of image frames in a multi-frame user action image sequence where the joint motion difference of the target feature vector is greater than zero; The error persistence of the target feature vector is determined by using the number of image frames and the total number of frames in the user action image sequence; By utilizing the degree of error persistence and the degree of difference in joint movement, the deviation fluctuation coefficient of the user's movement is determined.

[0010] Furthermore, determining the deviation fluctuation coefficient of the user's action using the degree of error persistence and the degree of joint movement difference includes: Determine the joint motion difference of each target feature vector and the minimum joint motion difference among them; The cumulative deviation during the user's movement is obtained by using the difference in motion of each joint and the minimum value of the difference in motion of the joints. The deviation fluctuation coefficient of the user's action is determined by using the degree of error persistence and the cumulative deviation.

[0011] Furthermore, the degree of consistency in user movement deviations is determined using joint motion variability, including: The deviation fluctuation coefficient of user action is determined by using the joint motion difference, and the maximum value of the deviation fluctuation coefficient of all feature vectors in the user action image is determined. By using the various deviation fluctuation coefficients and the maximum value of the deviation fluctuation coefficient, the degree of consistency of the user's action deviation is determined.

[0012] Furthermore, the method of obtaining a user's action execution standard score using the motion compensation index includes: The motion compensation index is quantified and converted into the degree of motion hazard. The degree of motion hazard and the preset full score standard are used to obtain the user's motion execution standard score.

[0013] Furthermore, the generation of targeted corrective suggestions includes: By utilizing the joint motion differences in user motion images, the locations of abnormal joints in user motion images can be identified and labeled. By utilizing the abnormal joint locations and user action types, corrective suggestions are generated for each user action type.

[0014] The present invention has the following beneficial effects: This invention acquires multiple frames of images during user movement to obtain feature vectors of multiple joints in the human body. These feature vectors are then matched with feature vectors from a standard movement template. The error characteristics of the multi-frame images are analyzed to quantify the degree of potential movement hazards and calculate the user's movement execution standard score. This improves the accuracy and reliability of the movement execution standard score, realizing a user movement execution standard score system for intelligent fitness equipment. Furthermore, by analyzing multiple frames of images during user movement, this invention quantifies the differences between the movement and the standard template, distinguishes between temporary and systematic errors, and calculates the degree of potential movement hazards. This generates scientific scoring results and targeted correction prompts, helping users to identify and correct incorrect movements in a timely manner, improve exercise effectiveness, prevent sports injuries, and achieve intelligent and personalized fitness guidance. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the steps of a user movement execution standard scoring system for an intelligent fitness device provided in one embodiment of the present invention. Figure 2 A detailed flowchart of step S1 in a user action execution standard scoring system for an intelligent fitness device provided in an embodiment of the present invention; Figure 3 A detailed flowchart of step S2 in a user action execution standard scoring system for an intelligent fitness device provided in an embodiment of the present invention; Figure 4 A detailed flowchart of step S22 in a user action execution standard scoring system for an intelligent fitness device provided in an embodiment of the present invention; Figure 5 A detailed flowchart of step S3 in a user action execution standard scoring system for an intelligent fitness device provided in an embodiment of the present invention; Figure 6 A detailed flowchart of step S3 in a user action execution standard scoring system for an intelligent fitness device provided in another embodiment of the present invention; Figure 7 This is a schematic diagram of the hardware operating environment of the user action execution standard scoring device for the intelligent fitness equipment involved in the embodiments of the present invention. Figure 8This is a schematic diagram of the framework structure of the user action execution standard scoring system for intelligent fitness equipment involved in the embodiments of the present invention; Figure 9 This is a schematic diagram of the main joints of the human body in the user movement execution standard scoring system of the intelligent fitness equipment involved in the embodiments of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a user movement performance standard scoring system for intelligent fitness equipment proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] Here is a brief explanation of the specific scenario in which this invention is addressed: When users train with fitness equipment, transient movement errors may occur due to momentary fluctuations in body posture, or systemic and persistent errors may result from long-standing habitual movement pattern deviations. Furthermore, due to the body's compensatory mechanisms, temporary errors can usually be limited to a single joint or localized area through rapid self-adjustment; however, systemic errors are not only difficult to correct effectively through compensatory mechanisms, but may also further lead to the transmission and accumulation of errors in multi-joint movements. This results in insufficient accuracy in the performance evaluation of movements, affecting users' effective assessment and optimization of their training quality.

[0020] Therefore, this invention analyzes multiple frames of images during a user's exercise process, quantifies the differences between the action and the standard template, distinguishes between temporary and systematic errors, and calculates the degree of potential movement hazards accordingly, thereby improving the accuracy of the scoring. It also marks locations with large deviations to help users identify and correct incorrect movements in a timely manner, improve exercise results, and achieve intelligent fitness guidance.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of a user action execution standard scoring system for intelligent fitness equipment provided by the present invention.

[0022] Example 1: For a user movement execution standard scoring system for an intelligent fitness equipment provided by this invention, please refer to [link / reference]. Figure 8 , Figure 8This is a schematic diagram of the framework structure of the user action execution standard scoring system for intelligent fitness equipment involved in the embodiments of the present invention.

[0023] The user movement execution standard scoring system of the intelligent fitness equipment (hereinafter referred to as the "user movement execution standard scoring system" or the "system") includes: The action recognition module A10 is used to determine the feature vector of the user action image based on the key pixel coordinates of the user action image, and to determine the feature vector difference between the user action image and the standard action image of the corresponding frame. The deviation analysis module A20 is used to determine the difference in joint motion between the user's motion image and the standard motion image in the target feature vector by using the difference in feature vectors and the angle between adjacent feature vectors; it uses the difference in joint motion to determine the deviation fluctuation coefficient and the degree of consistency of the user's motion deviation; and it uses the deviation fluctuation coefficient and the degree of consistency of the motion deviation to obtain the motion compensation index. The A30 motion evaluation module is used to obtain a score of the user's motion performance standard using the motion compensation index and generate targeted correction suggestions.

[0024] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a user movement execution standard scoring system for intelligent fitness equipment provided in an embodiment of the present invention.

[0025] The methods and steps corresponding to the user movement execution standard scoring system of the intelligent fitness equipment include: Step S1: Determine the feature vector of the user action image based on the key pixel coordinates of the user action image, and determine the feature vector difference between the user action image and the standard action image of the corresponding frame. In this embodiment, to implement a user movement execution standard scoring system for intelligent fitness equipment, it is necessary to add corresponding equipment. A high-resolution camera is used for recording to enhance the accuracy of posture estimation. The specific process is as follows: Arrange the equipment around the fitness equipment to capture images from both the front and side views, ensuring that the user's key joint movement trajectories are captured from all angles and avoiding visual obstruction.

[0026] During the user's action, multiple frames of user action image sequence are continuously acquired. Let's assume K frames are captured, denoted as K frames. .

[0027] Perform existing preprocessing operations such as denoising, distortion correction, and background segmentation on each frame (user action image).

[0028] Please refer to Figure 9 , Figure 9This is a schematic diagram of the main joints of the human body in the user movement execution standard scoring system of the intelligent fitness equipment involved in the embodiments of the present invention.

[0029] Integrated human pose estimation models (such as Open Pose and Media Pipe Pose) are used to analyze each frame of the preprocessed image to identify and locate key human joints (such as...). Figure 9 The pixel coordinates are shown in the figure.

[0030] Specifically, step S1, which determines the feature vector of the user action image based on the keypoint pixel coordinates of the user action image, includes: The feature vector sequence of the user action image is constructed by using the pixel coordinates of adjacent joints in the user action image according to the preset direction.

[0031] In this embodiment, based on the identified key points, a feature vector is constructed for the k-th frame of the user action image using the coordinates of adjacent key point pixels in a preset direction, such as from top to bottom. Assuming each frame has N feature vectors, the sequence (set) of feature vectors is denoted as follows: .

[0032] Specifically, please refer to Figure 2 Step S1, determining the feature vector difference between the user's motion image and the standard motion image of the corresponding frame, includes: Step S11: Align the multi-frame user motion image sequence with the standard motion image sequence in time series to determine the standard motion image of the frame corresponding to the user motion image. Step S12: Determine the cosine value of the target feature vector between the user action image and the standard action image of the corresponding frame, and use the cosine value to obtain the feature vector difference degree.

[0033] In this embodiment, due to differences in training experience, individual strength, and physical condition, the degree of standardization of users' movements varies during training with fitness equipment. This is manifested in the differences between multiple sets of joint feature vectors and the standard. The greater the difference, the less standard the movement is. Therefore, the degree of non-standard movement can be quantified by calculating the feature vector difference degree and then calculating the joint movement difference degree. The specific process is as follows: For the standard movements corresponding to the user's exercise type, collect high-frequency standard movement image sequences of multiple professional fitness personnel performing the standard movements (the sampling frequency should be higher than the user's sampling frequency to ensure the accuracy and smoothness of the template). Assuming a total of I frames are collected, and N joint feature vectors are extracted from each frame, then the feature vector sequence (set) of the i-th frame of the standard template is denoted as... .

[0034] Because the speed at which users perform actions may differ from the standard template, direct frame-to-frame comparison can amplify errors. Therefore, a Dynamic Time Warping (DTW) algorithm is needed to non-linearly align the user action image sequence with the standard action image sequence in the time dimension, finding the optimal matching path. This ensures that each frame of user data is compared with the most similar standard frame, thereby reducing errors caused by the potential inconsistency between the user's action speed and the standard template. Assume that the k-th frame of the user action image sequence corresponds to the i-th frame of the standard action image sequence.

[0035] Due to differences in training experience, individual strength, and physical condition, users' execution of movements varies during fitness equipment training. This difference is reflected in the discrepancy between the user's joint feature vector and the standard feature vector. The greater the discrepancy, the less standard the user's movements are. The feature vector difference is denoted as... The calculation formula is as follows: In the formula, This represents the difference in the nth feature vector of the user action image in the kth frame. This represents the nth feature vector of the user action image in the kth frame (as a target feature vector, it refers to any feature vector). This represents the nth feature vector of the i-th frame of the standard motion image. express The cosine of the angle between two eigenvectors. It should be noted that in this embodiment... The angle between two eigenvectors is in the range [0, 180°), excluding the extreme case of 180°. The 1 in the denominator is a reasonable constant set based on the range of any cosine value of the angle in the range [-1, 1], used to reflect that the larger the angle between the two eigenvectors, the greater the difference between them, and the greater the degree of difference between the eigenvectors.

[0036] The consistency of the user's direction is measured by calculating the cosine of the aligned user joint feature vector and the standard feature vector. The larger the value, the closer the two directions are, meaning the smaller the angle between them and the smaller the difference from the standard movement. Conversely, a smaller value indicates that the user's movement is less standard.

[0037] Step S2: Using the feature vector difference degree and the angle between adjacent feature vectors, determine the joint motion difference degree of the user action image compared with the standard action image in the target feature vector; When users train with fitness equipment, transient movement errors may occur due to momentary fluctuations in body posture (such as a brief sway followed by a rapid return to the correct movement trajectory), characterized by a short duration of deviation. Systematic and persistent errors may also result from non-standard movements, characterized by a long duration of deviation and significant fluctuations. Furthermore, due to the body's compensatory mechanisms, these two types of errors manifest differently during exercise. For transient errors, the body can quickly self-regulate to confine the error to a single joint or localized area; however, for systematic errors, the compensatory mechanism persists in the erroneous movement, failing to effectively correct it. This leads to further exacerbation of error transmission and accumulation in multi-joint linkages, manifesting as a persistent global deviation across multiple joints and a high degree of consistency in multi-joint movement deviations. By analyzing the error characteristics of multiple frames of images, the degree of potential movement hazards can be quantified.

[0038] Specifically, please refer to Figure 3 Step S2 includes: Step S21: Determine the angle between adjacent feature vectors of the user using adjacent feature vectors in the user action image, and determine the angle between adjacent feature vectors of the standard action image using adjacent feature vectors. Step S22: Using the feature vector difference degree, the angle between adjacent feature vectors of the user and the angle between adjacent feature vectors of the standard, determine the joint motion difference degree of the user action image compared with the standard action image in the target feature vector.

[0039] More specifically, please refer to Figure 4 Step S22 includes: Step S221: Using the angle between adjacent feature vectors of the user and the angle between adjacent feature vectors of the standard, determine the difference in joint angles of the user action image and the standard action image in the target feature vector. Step S222: Using feature vector difference degree and joint angle difference degree, determine the joint motion difference degree of the user action image compared with the standard action image in the target feature vector.

[0040] In this embodiment, regardless of a person's height, weight, or distance, as long as they assume the same posture, the calculated joint angles will be consistent. Therefore, the angle between the target feature vector and its neighboring feature vectors can be calculated, and the difference from the standard angle can be calculated to reflect the degree of joint movement variation, denoted as . The calculation formula is as follows: In the formula, It represents the angle between the nth feature vector of the user action image in the kth frame and the (n+1)th adjacent feature vector (the angle between adjacent user feature vectors). This represents the nth feature vector of the user action image in the kth frame. This represents the (n+1)th feature vector of the user action image in the k-th frame. Represents the magnitude of a vector. It represents the inverse function of the cosine of the angle between two vectors.

[0041] In the formula, This represents the joint angle difference of the nth feature vector in the k-th frame of the user motion image (compared to the i-th frame of the standard motion image). The angle between the a-th feature vector and the (a+1)-th feature vector represents the difference between the user motion image in frame k and the standard motion image in frame i. The angle between the two is denoted as the angle between adjacent user feature vectors and the angle between adjacent standard feature vectors, respectively.

[0042] Because the joint angles differ little under the same posture, the joint angles are calculated using adjacent feature vectors and compared with the joint angles under standard motion. The smaller the value, the more consistent the posture and the more standardized the movement. Further, the degree of difference in joint movement is obtained: In the formula, This represents the joint motion difference of the nth feature vector in the k-th frame of the user motion image (compared to the i-th frame of the standard motion image). This represents the joint angle difference of the nth feature vector in the kth frame of the user motion image. The eigenvector difference represents the eigenvector difference of the nth eigenvector of the user action image in the kth frame.

[0043] The degree of joint motion difference is obtained by combining the consistency of vector direction and angular deviation. The larger the value, the more serious the deviation of the joint movement from the standard movement, and the worse the user's performance of the movement.

[0044] Step S3: Determine the deviation fluctuation coefficient and the degree of consistency of the user's movement using the joint movement difference, and obtain the movement compensation index using the deviation fluctuation coefficient and the degree of consistency of the movement deviation. In the evaluation of the standard of movement execution, not all "errors" have the same nature and impact. During exercise using fitness equipment, users may experience momentary fluctuations in body posture (such as a brief sway followed by a rapid return to the correct movement trajectory), resulting in temporary errors in the duration of the movement that are much shorter than a complete movement cycle. Furthermore, due to the user's own non-standard movements, such as knee valgus, the deviation will be continuous and repetitive during the exercise.

[0045] Due to the compensatory mechanisms of physical exercise, temporary errors are quickly adjusted to limit the error to a single aspect, manifesting as localized deviations in a single joint or limb, without affecting the overall movement pattern. However, for systematic errors, the compensatory mechanism cannot effectively correct them and may even exacerbate the error, manifesting as persistent deviations across multiple joints. Therefore, by quantifying the compensatory mechanisms for different types of errors and calculating a compensation index, we can distinguish between temporary and systematic errors.

[0046] Specifically, in one embodiment, please refer to Figure 5 Step S3, which uses joint motion variability to determine the deviation fluctuation coefficient of the user's action, includes: Step S31: Determine the number of image frames in a multi-frame user action image sequence where the joint motion difference of the target feature vector is greater than zero; Step S32: Using the number of image frames and the total number of frames in the user action image sequence, determine the degree of error persistence of the target feature vector; In this embodiment, during exercise using fitness equipment, users may experience temporary movement errors due to momentary fluctuations in body posture. These errors are short-lived, but due to the user's non-standard movements, they may manifest as continuous and repetitive deviations during exercise.

[0047] Therefore, the persistence of the error can be reflected by counting the number of image frames where the joint motion difference is greater than zero during the motion process. Let's assume the number of image frames where the joint motion difference of the nth feature vector is greater than zero is denoted as... The degree of error persistence is denoted as The calculation formula is as follows: In the formula, This indicates the persistence of error in the nth eigenvector. This represents the joint motion difference of the nth feature vector in the user action image sequence. The number of image frames when the value is greater than zero. This indicates that a total of K frames (total number of frames) were obtained for the corresponding user action image sequence during the motion process.

[0048] The formula above reflects the proportion of time during which the nth feature vector deviates during the motion. The larger the value, the more persistent the deviation of the feature vector is throughout the entire motion cycle, and the more likely it is to be a systematic error.

[0049] Step S33: Determine the deviation fluctuation coefficient of the user's action by using the degree of error persistence and the degree of difference in joint movement.

[0050] More specifically, step S33 includes: Determine the joint motion difference of each target feature vector and the minimum joint motion difference among them; The cumulative deviation during the user's movement is obtained by using the difference in motion of each joint and the minimum value of the difference in motion of the joints. The deviation fluctuation coefficient of the user's action is determined by using the degree of error persistence and the cumulative deviation.

[0051] In this embodiment, during exercise using fitness equipment, users may experience momentary fluctuations in body posture (such as a brief sway followed by a rapid return to the correct movement trajectory), resulting in temporary errors with a duration much shorter than a complete movement cycle. These errors manifest as brief spikes in the data. Conversely, due to non-standard user movements, such as knee valgus, the errors appear as persistent and repetitive high deviations in the data. Therefore, a deviation fluctuation coefficient is calculated based on the error fluctuation of a specific joint throughout the entire movement sequence, denoted as [missing information]. This reflects the degree of deviation from the standard, and the specific calculation formula is as follows: In the formula, This represents the deviation fluctuation coefficient of the nth eigenvector. This indicates the persistence of error in the nth eigenvector. This represents the joint motion difference of the nth feature vector in the kth frame of the user motion image. The minimum value of the joint motion difference of the nth eigenvector.

[0052] This represents the cumulative deviation during the movement; the larger the value, the greater the degree to which the user's movements deviate from the standard. The larger the value, the greater the persistence of the error and the greater the degree of error.

[0053] Specifically, in another embodiment, please refer to Figure 6 Step S3, which uses joint motion difference to determine the consistency of user movement deviation, includes: Step S301: Determine the deviation fluctuation coefficient of the user's action using the joint motion difference, and determine the maximum value of the deviation fluctuation coefficient of all feature vectors in the user's action image. Step S302: Using each deviation fluctuation coefficient and the maximum value of the deviation fluctuation coefficient, determine the degree of consistency of the user's action deviation.

[0054] In this embodiment, during movement, since systematic errors typically manifest as coordinated deviations of multiple joints, while temporary errors are mostly local and isolated phenomena, the degree of consistency in movement deviation can be used to assess the coordination of multiple joints in error performance, reflecting whether the error has global characteristics. That is, the degree of consistency in movement deviation is used to determine whether it is a systematic error; the better the consistency, the more likely it is to be a systematic error. The degree of consistency in movement deviation is denoted as... The calculation formula is as follows: In the formula, Indicates the degree of consistency in action deviation. This represents the deviation fluctuation coefficient of the nth eigenvector. This represents the maximum deviation fluctuation coefficient of N eigenvectors. The 1 in the denominator is used to limit... The maximum value does not exceed 1.

[0055] This represents the difference in the deviation fluctuation coefficient of N feature vectors. The smaller the value, the better the consistency of the error of the feature vectors, that is, the better the coordination of multiple joints in error performance, indicating a better degree of consistency in motion deviation. In this case, the greater the possibility of systematic error.

[0056] Further, step S3, which uses the deviation fluctuation coefficient and the degree of consistency of movement deviation to obtain the motion compensation index, includes: Compensation mechanisms are adaptive responses of the body during exercise or daily activities. When the function of a certain part is limited (such as muscle weakness, insufficient joint mobility, or injury), the body compensates for the functional loss by having other parts (muscles, joints, or posture) work extra to make up for the functional deficiency. For temporary errors, this is a localized deviation of a single joint or limb; for systematic errors, it is a global deviation of multiple joints or the entire limb. Therefore, the exercise compensation index can be calculated by the degree of consistency and the overall deviation during exercise, denoted as [insert index here]. The calculation formula is as follows: In the formula, This indicates the compensatory index during exercise. Indicates the degree of consistency in action deviation. This represents the deviation fluctuation coefficient of the nth eigenvector.

[0057] The motion compensation index represents the motion compensation process and reflects the persistence of error in time, the coordination in space, and the degree of accumulation of deviation. The larger the value, the more likely the error is to be a systematic error, the failure of the compensation mechanism, and the existence of global and continuous problems in the execution of the action.

[0058] Step S4: Use the motion compensation index to obtain the user's action execution standard score and generate targeted correction suggestions.

[0059] Specifically, step S4, which uses the motion compensation index to obtain a score on the user's action execution standard, includes: The motion compensation index is quantified and converted into the degree of motion hazard. The degree of motion hazard and the preset full score standard are used to obtain the user's motion execution standard score.

[0060] In this embodiment, people may find it difficult to achieve good exercise results during exercise, and may even suffer physical injury due to improper movements, leading to counterproductive results. This is addressed through the exercise compensation index. This index reflects the degree of potential movement hazards during exercise. A higher value indicates that the error is more likely to be a systematic error, meaning it's caused by the user's incorrect form. Prolonged incorrect exercise can lead to injury, requiring correction. Therefore, the degree of movement hazard is calculated using a quantified movement compensation index, denoted as [insert index here]. The calculation formula is as follows: In the formula, Indicates the degree of potential hazard in the action. This indicates the compensatory index during exercise. is a natural constant. The 1 in the denominator is used to limit... The maximum value does not exceed 1.

[0061] The degree of potential motion hazards is calculated by quantifying the motion compensation index. The larger the value, the more likely the error is to be a systematic error. In this case, there is a global and continuous problem in the user's action execution, which should be prompted for correction first.

[0062] Furthermore, based on the calculated degree of hazard... The data is converted into a score of the user's action execution standard, and images and regions with large action errors are located, marked, and corrective prompts are provided. The specific process is as follows: Calculation-based risk level of action Calculate the performance standard score of the action, denoted as S, using the following formula: In the formula, Score the standard of user action execution. Indicates the degree of potential hazard in the action. The standard for a perfect score is preset and can be adjusted.

[0063] It can be seen that the higher the risk level of the action, the lower the score.

[0064] Specifically, step S4, generating targeted corrective suggestions, includes: By utilizing the joint motion differences in user motion images, the locations of abnormal joints in user motion images can be identified and labeled. By utilizing the abnormal joint locations and user action types, corrective suggestions are generated for each user action type.

[0065] In this embodiment, based on the joint motion difference calculated in the aforementioned embodiments, joints with significant errors are marked with red highlighted circles on the user's motion image to indicate the locations of the abnormal joints, and the specific difference values ​​are displayed. It should be noted that "significant error" here can be defined as comparing the joint motion difference with a threshold corresponding to the current user's motion type; an error greater than this threshold is considered significant.

[0066] Based on the error localization results (abnormal joint positions), targeted correction suggestions are generated in conjunction with the user's movement type (such as squats, bench presses, etc.). For example, if the user's knee flexion angle during a squat is significantly larger than the standard movement, specifically more than 10° greater than the standard angle, then the user should be advised to continue squatting to reach the standard flexion angle.

[0067] By integrating the parameter calculation and scoring generation in the above embodiments into the user action execution standard scoring system of fitness equipment, a user action execution standard scoring system for intelligent fitness equipment is realized.

[0068] This invention acquires multiple frames of images during user movement to obtain feature vectors of multiple joints in the human body. These feature vectors are then matched with feature vectors from a standard movement template. The error characteristics of the multi-frame images are analyzed to quantify the degree of potential movement hazards and calculate the user's movement execution standard score. This improves the accuracy and reliability of the movement execution standard score, realizing a user movement execution standard score system for intelligent fitness equipment. Furthermore, by analyzing multiple frames of images during user movement, this invention quantifies the differences between the movement and the standard template, distinguishes between temporary and systematic errors, and calculates the degree of potential movement hazards. This generates scientific scoring results and targeted correction prompts, helping users to identify and correct incorrect movements in a timely manner, improve exercise effectiveness, prevent sports injuries, and achieve intelligent and personalized fitness guidance.

[0069] Example 2: This invention also proposes a user movement execution standard scoring device for intelligent fitness equipment. The device can be a data processing device such as a computer or server, or a combination of multiple devices.

[0070] like Figure 7 As shown, Figure 7This is a schematic diagram of the hardware operating environment of the user action execution standard scoring device for the intelligent fitness equipment involved in the embodiments of the present invention.

[0071] like Figure 7 As shown, the user movement execution standard scoring device of this intelligent fitness equipment may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display or an input unit such as a control panel; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed RAM or a stable, non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001. The memory 1005, as a computer storage medium, may include a user movement execution standard scoring program.

[0072] Those skilled in the art will understand that Figure 7 The hardware structure shown does not constitute a limitation on the device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0073] Continue to refer to Figure 7 , Figure 7 The memory 1005, which is a computer-readable storage medium, may include an operating device, a user interface module, a network communication module, and a user action execution standard scoring program.

[0074] exist Figure 7 In this context, the network communication module is mainly used to connect to the server and can communicate with the server for data; while the processor 1001 can call the user actions stored in the memory 1005 to execute the standard score program and execute the steps in the above embodiments.

[0075] Based on the hardware structure of the user movement execution standard scoring device for the above-mentioned intelligent fitness equipment, various embodiments of the user movement execution standard scoring system for the intelligent fitness equipment of the present invention are implemented.

[0076] Furthermore, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a user action execution standard scoring program, wherein when the user action execution standard scoring program is executed by a processor, it implements the steps of the method corresponding to the user action execution standard scoring system for the intelligent fitness equipment described above.

[0077] The method implemented when the user action execution standard score program is executed can be referred to in various embodiments of the user action execution standard score system of the intelligent fitness equipment of the present invention, and will not be repeated here.

[0078] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0080] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] The above description is only a preferred embodiment of the present invention and does not limit the scope of protection of the present invention. All equivalent structural / method transformations made under the inventive concept of the present invention using the contents of the present invention specification and drawings, or direct / indirect applications in other related technical fields, are included within the scope of protection of the present invention.

Claims

1. A user movement execution standard scoring system for intelligent fitness equipment, characterized in that, The system includes: The action recognition module is used to determine the feature vector of the user action image based on the key pixel coordinates of the user action image, and to determine the feature vector difference between the user action image and the standard action image of the corresponding frame. The deviation analysis module is used to determine the difference in joint motion between the user's motion image and the standard motion image in the target feature vector by using the difference in feature vectors and the angle between adjacent feature vectors; it uses the difference in joint motion to determine the deviation fluctuation coefficient and the degree of consistency of the user's motion deviation; and it uses the deviation fluctuation coefficient and the degree of consistency of the motion deviation to obtain the motion compensation index. The exercise evaluation module is used to obtain a score of the user's performance standard using the exercise compensation index and generate targeted correction suggestions.

2. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, The process of determining the feature vector of the user action image based on the keypoint pixel coordinates of the user action image includes: The feature vector sequence of the user action image is constructed by using the pixel coordinates of adjacent joints in the user action image according to the preset direction.

3. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, The determination of the feature vector difference between the user's motion image and the standard motion image of the corresponding frame includes: Align the multi-frame user motion image sequence with the standard motion image sequence in time series to determine the standard motion image corresponding to the user motion image frame; Determine the cosine value of the target feature vector between the user action image and the standard action image of the corresponding frame, and use the cosine value to obtain the feature vector difference.

4. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, The method of determining the joint motion difference between a user's motion image and a standard motion image in the target feature vector, using the feature vector difference degree and the angle between adjacent feature vectors, includes: The angle between adjacent feature vectors of a user is determined by using adjacent feature vectors in the user's action image, and the angle between adjacent feature vectors of a standard action image is determined by using adjacent feature vectors of a standard action image. By using the feature vector difference degree, the angle between adjacent feature vectors of the user and the angle between adjacent feature vectors of the standard, the joint motion difference degree of the user action image compared with the standard action image in the target feature vector is determined.

5. The user movement execution standard scoring system for intelligent fitness equipment according to claim 4, characterized in that, The method of determining the joint motion difference between the user's motion image and the standard motion image in the target feature vector by utilizing the feature vector difference degree, the angle between adjacent user feature vectors, and the angle between adjacent standard feature vectors includes: By using the angle between adjacent feature vectors of the user and the angle between adjacent feature vectors of the standard, the difference in joint angles of the user action image and the standard action image in the target feature vector is determined. By utilizing feature vector difference and joint angle difference, the joint motion difference between the user motion image and the standard motion image in the target feature vector is determined.

6. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, Determining the deviation fluctuation coefficient of user movements using joint motion variability includes: Determine the number of image frames in a multi-frame user action image sequence where the joint motion difference of the target feature vector is greater than zero; The error persistence of the target feature vector is determined by using the number of image frames and the total number of frames in the user action image sequence; By utilizing the degree of error persistence and the degree of difference in joint movement, the deviation fluctuation coefficient of the user's movement is determined.

7. The user movement execution standard scoring system for intelligent fitness equipment according to claim 6, characterized in that, The method of determining the deviation fluctuation coefficient of user movements by utilizing the degree of error persistence and the degree of joint movement difference includes: Determine the joint motion difference of each target feature vector and the minimum joint motion difference among them; The cumulative deviation during the user's movement is obtained by using the difference in motion of each joint and the minimum value of the difference in motion of the joints. The deviation fluctuation coefficient of the user's action is determined by using the degree of error persistence and the cumulative deviation.

8. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, Determining the consistency of user movement deviations using joint motion variability includes: The deviation fluctuation coefficient of user action is determined by using the joint motion difference, and the maximum value of the deviation fluctuation coefficient of all feature vectors in the user action image is determined. By using the various deviation fluctuation coefficients and the maximum value of the deviation fluctuation coefficient, the degree of consistency of the user's action deviation is determined.

9. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, The method of obtaining a user's action execution standard score using the motion compensation index includes: The motion compensation index is quantified and converted into the degree of motion hazard. The degree of motion hazard and the preset full score standard are used to obtain the user's motion execution standard score.

10. The user movement execution standard scoring system for intelligent fitness equipment according to claim 1, characterized in that, The generation of targeted corrective suggestions includes: By utilizing the joint motion differences in user motion images, the locations of abnormal joints in user motion images can be identified and labeled. By utilizing the abnormal joint locations and user action types, corrective suggestions are generated for each user action type.

Images