Methods, devices, equipment, and media for computer vision-based muscle fitness testing

Abstract

This application discloses a method, apparatus, device, and medium for muscle fitness testing based on computer vision, relating to the field of computer vision technology. The method includes: determining the user's posture parameters during several identical movements in a muscle fitness test, and determining the total output power of each movement based on the posture parameters; determining the user's initial power and power decay rate based on the total output power of each movement, and determining the user's 1RM based on the subjective score and total output power of each movement, as well as the initial power; determining the user's muscle strength score based on the 1RM, and determining the user's muscle endurance score based on the total output power and power decay rate of each movement; determining the user's muscle coordination score and quality score for completing the muscle fitness test, and determining and outputting the user's muscle fitness test results based on the muscle strength score, muscle endurance score, muscle coordination score, and quality score. This application can improve the accuracy of muscle fitness testing.

Description

Technical Field

[0001] This invention relates to the field of computer vision technology, and in particular to a method, apparatus, device, and medium for testing muscle fitness based on computer vision. Background Technology

[0002] Muscle fitness is a core physiological indicator for measuring human muscle strength, endurance, and functional ability, and is crucial for maintaining bone health, metabolic levels, daily activity levels, and overall bodily integrity. Traditional muscle fitness testing relies on cumbersome specialized muscle function tests, requiring strict adherence to multiple procedures including aerobic warm-up, equipment familiarization, and standardized movements. In muscle strength testing, 1RM is a core indicator, but traditional testing procedures are complex, requiring multiple increases in resistance with rest intervals. Furthermore, test results are highly dependent on equipment type, with data from different manufacturers lacking interoperability, and some testing equipment no longer available, leading to poor data reliability. Muscle endurance testing also has scientific shortcomings; commonly used tests such as push-ups and crunches have weak correlations with the endurance of the target muscle groups and cannot accurately reflect actual endurance levels.

[0003] Furthermore, existing testing methods have significant limitations: firstly, a single test can only target local muscle groups and cannot comprehensively assess whole-body muscle function; secondly, standardization is insufficient, with a lack of unified testing protocols for specific subjects, and equipment differences further affect the accuracy of results; thirdly, strong reliance on specialized equipment limits its application scenarios and makes data acquisition difficult; and fourthly, it cannot scientifically demonstrate the correlation between muscle fitness testing and results. These problems result in significant deficiencies in the scientific validity, consistency, safety, and practical application value of traditional testing methods, making it difficult to meet the testing needs of users of different ages and health statuses.

[0004] In conclusion, improving the safety, universality, and reliability of muscle fitness testing is a pressing technical issue that needs to be addressed. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for muscle fitness testing based on computer vision, which can improve the safety, universality, and reliability of muscle fitness testing. The specific solution is as follows:

[0006] In a first aspect, this application provides a computer vision-based muscle fitness testing method applied to a computer device, comprising:

[0007] The system uses computer vision technology to collect motion videos of users performing muscle fitness tests, processes the motion videos to determine the posture parameters of the user when performing the same action several times in the muscle fitness test, and uses the SMPL model corresponding to the user to generate target motion trajectories based on the posture parameters corresponding to each action, so as to determine the total output power of the user when performing each action based on the target motion trajectories.

[0008] The user's initial power and power decay rate are determined based on the total output power corresponding to each action, and the user's 1RM is determined based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays when performing the muscle fitness test.

[0009] The user's muscle strength score is determined based on the 1RM, and the user's muscle endurance score is determined based on the total output power and the power decay rate corresponding to each movement.

[0010] The system determines the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and determines and outputs the corresponding muscle fitness test result for the user based on the muscle strength score, muscle endurance score, muscle coordination score, and quality score.

[0011] Optionally, the step of acquiring motion videos of the user performing a muscle fitness test based on computer vision technology and processing the motion videos to determine the user's posture parameters during several identical movements in the muscle fitness test includes:

[0012] The SMPL model corresponding to the user is pre-constructed, and the motion video of the user performing the same action several times in the muscle fitness test is collected based on computer vision technology.

[0013] The motion sequence of joints corresponding to each action is extracted from the motion video by a preset human posture estimation algorithm, and the posture parameters corresponding to the user when performing the same action in several times in the muscle fitness test are determined based on the joint motion sequence of each action.

[0014] Optionally, determining the total output power of the user when performing each action based on each of the target motion trajectories includes:

[0015] Based on the target motion trajectory, the angular velocity and angular acceleration of each joint of the user when performing each action are determined, and based on the transformation matrix corresponding to the posture parameters of each action, the angular velocity of each joint of each action is converted into the angular velocity of each limb.

[0016] Using the Zatsiorsky model, the mass of each limb is determined based on the user's height, weight, and the proportion of each limb, and the moment of inertia of each limb is determined based on the mass of each limb.

[0017] The net joint torque of each joint is determined based on the rotational inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint when the user performs each action.

[0018] Based on the net joint torque and angular velocity of each joint when the user performs each action, determine the instantaneous output power of each joint when the user performs each action.

[0019] The total output power of the user during each action is determined based on the instantaneous output power of each joint when the user performs each action.

[0020] Optionally, after determining the net joint torque of each joint when the user performs each action based on the rotational inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint, the method further includes:

[0021] Based on the target motion trajectory corresponding to each action performed by the user, the muscle activation state of the user during each action is determined;

[0022] Based on the muscle activation state and net joint torque of each joint when the user performs each action, the muscle group strength of each muscle when the user performs each action is determined, and the muscle activation degree of the user when performing each action is determined based on the muscle group strength of each muscle.

[0023] The user's muscle activation level during each action is visualized using a pre-set visualization platform.

[0024] Optionally, determining the 1RM corresponding to the user based on the subjective score corresponding to each action, the total output power, and the initial power includes:

[0025] If the ratio between the total output power of the user when performing the current action and the initial power is less than or equal to a preset power decrease ratio threshold, and the subjective score corresponding to the current action is greater than or equal to a preset subjective score threshold, then the target number of movements corresponding to the current action in the muscle fitness test is determined.

[0026] The target load corresponding to the current action is determined based on the net joint torque of each joint when the user performs the current action, and the 1RM corresponding to the user is determined based on the target load and the target number of movements.

[0027] Optionally, determining the user's muscle strength score based on the 1RM and determining the user's muscle endurance score based on the total output power and the power decay rate corresponding to each movement includes:

[0028] The user's muscle strength score is determined based on the ratio between the target load and the 1RM.

[0029] Based on the total output power corresponding to each movement and the number of movements corresponding to the muscle fitness test, the average output power corresponding to the user completing the muscle fitness test is determined, and the standard deviation of the output power is determined based on the total output power corresponding to each movement and the average output power.

[0030] A first score is determined based on the ratio between the average output power and the standard deviation, and a second score is determined based on the ratio between the power attenuation rate and a preset power attenuation rate threshold.

[0031] The first score and the second score are weighted and calculated to obtain the user's muscle endurance score.

[0032] Optionally, determining the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test includes:

[0033] Based on the muscle group strength of each muscle when the user performs each action, the muscle coordination score corresponding to the user's completion of the muscle fitness test is determined.

[0034] Determine the preset standard movement trajectory template and preset trajectory deviation threshold corresponding to the muscle fitness test, and determine the trajectory deviation between each target movement trajectory and the preset standard movement trajectory template;

[0035] The third score is determined by using a dynamic time warping method based on the ratio between each trajectory deviation and the preset trajectory deviation threshold.

[0036] Determine several poses to be verified when the user performs each action, and determine the cosine similarity between each pose to be verified and a preset standard pose template;

[0037] The fourth score is determined based on the cosine similarity, and the third and fourth scores are weighted to obtain the quality score corresponding to the user's completion of the muscle fitness test.

[0038] Secondly, this application provides a computer vision-based muscle fitness testing device, applied to a computer device, comprising:

[0039] The total output power determination module is used to acquire motion videos of a user performing a muscle fitness test based on computer vision technology, process the motion videos to determine the posture parameters of the user when performing several identical movements in the muscle fitness test, and use the SMPL model corresponding to the user to generate target motion trajectories based on the posture parameters corresponding to each movement, so as to determine the total output power of the user when performing each movement based on each target motion trajectory.

[0040] The power decay rate determination module is used to determine the user's initial power and power decay rate based on the total output power corresponding to each action, and to determine the user's 1RM based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays when performing the muscle fitness test.

[0041] A muscle endurance rating determination module is used to determine the user's muscle strength rating based on the 1RM, and to determine the user's muscle endurance rating based on the total output power and the power decay rate corresponding to each movement;

[0042] The test result determination and output module is used to determine the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and to determine and output the corresponding muscle fitness test result of the user based on the muscle strength score, the muscle endurance score, the muscle coordination score and the quality score.

[0043] Thirdly, this application provides an electronic device, comprising:

[0044] Memory, used to store computer programs;

[0045] A processor is used to execute the computer program to implement the aforementioned computer vision-based muscle fitness testing method.

[0046] Fourthly, this application provides a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned computer vision-based muscle fitness testing method.

[0047] In this application, firstly, motion videos of a user performing a muscle fitness test are acquired using computer vision technology. These videos are then processed to determine the user's posture parameters during several identical movements in the muscle fitness test. Using the user's corresponding SMPL model, target motion trajectories are generated based on the posture parameters for each movement. The total output power of the user during each movement is then determined based on these target motion trajectories. Subsequently, the user's initial power and power decay rate are determined based on the total output power for each movement. Finally, the user's subjective score for each movement, the total output power, and the initial power are used to determine the... The corresponding 1RM; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays during the muscle fitness test; then, the user's muscle strength score is determined based on the 1RM, and the user's muscle endurance score is determined based on the total output power and the power decay rate for each movement; finally, the user's muscle coordination score and quality score for completing the muscle fitness test are determined, and the corresponding muscle fitness test result for the user is determined and output based on the muscle strength score, the muscle endurance score, the muscle coordination score, and the quality score. As can be seen from the above, this application first uses computer vision technology to collect and process motion videos of users performing muscle fitness tests, obtaining posture parameters of the user performing several identical movements in the muscle fitness test. Combined with the user's SMPL model, a target motion trajectory is generated, thereby determining the total output power for each movement. Then, based on the total output power, an initial power and power decay rate are fitted, and the 1RM value is determined by combining the subjective score of each movement with the total output power. Subsequently, muscle strength scores are calculated based on the 1RM, muscle endurance scores are calculated based on the total output power and power decay rate, and muscle coordination and movement quality scores are determined simultaneously, ultimately obtaining and outputting the muscle fitness test results. In this way, firstly, this application does not rely on expensive professional equipment, achieving non-contact testing using only a regular camera; secondly, this application achieves high-precision, dynamic, and personalized muscle output simulation through the SMPL model, and its simulation results show a significant correlation with real physiological data; and thirdly, this application can comprehensively cover core dimensions such as muscle strength, endurance, coordination, and movement quality. In this way, by leveraging standardized data processing procedures and personalized model fitting, this application improves the scientific rigor and accuracy of test results, making it suitable for users of different ages and health conditions, and significantly expanding the application scenarios and universality of muscle fitness testing. Attached Figure Description

[0048] To more clearly illustrate the technical solutions 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0049] Figure 1 A flowchart of a computer vision-based muscle fitness testing method provided for this application;

[0050] Figure 2 A schematic diagram of a specific SMPL model and musculoskeletal model provided for this application;

[0051] Figure 3 A specific visualization result diagram provided for this application;

[0052] Figure 4 A schematic diagram of a specific exponential decay model provided in this application;

[0053] Figure 5 A schematic diagram of a muscle fitness testing device based on computer vision provided in this application;

[0054] Figure 6 This application provides a structural diagram of an electronic device. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Muscle fitness is a core physiological indicator for measuring human muscle strength, endurance, and functional ability, and is crucial for maintaining bone health, metabolic levels, daily activity levels, and the integrity of bodily functions. Traditional muscle fitness testing relies on cumbersome specialized muscle function tests, requiring strict adherence to multiple procedures such as aerobic warm-up, equipment familiarization, and standardized movements. Furthermore, existing testing methods have significant limitations: firstly, a single test can only target a local muscle group, failing to comprehensively assess overall muscle function; secondly, standardization is insufficient, lacking a unified testing protocol for specific subjects, and equipment differences further affect the accuracy of results; thirdly, the strong reliance on specialized equipment limits its application scenarios and makes data acquisition difficult; and fourthly, it cannot scientifically demonstrate the correlation between muscle fitness testing and results. These problems result in significant deficiencies in the scientific validity, consistency, safety, and practical application value of traditional testing methods, making it difficult to meet the testing needs of users of different ages and health statuses. Therefore, this application provides a computer vision-based muscle fitness testing scheme that can improve the safety, universality, and reliability of muscle fitness testing.

[0057] See Figure 1 As shown, this embodiment of the invention discloses a computer vision-based muscle fitness testing method, applied to a computer device, which may include:

[0058] Step S11: Collect motion videos of the user performing a muscle fitness test based on computer vision technology, and process the motion videos to determine the posture parameters of the user when performing several identical movements in the muscle fitness test. Using the SMPL model corresponding to the user, generate target motion trajectories based on the posture parameters corresponding to each movement, and determine the total output power of the user when performing each movement based on each target motion trajectory.

[0059] In this embodiment, a user-specific SMPL (Skinned Multi-Person Linear) model is pre-constructed. The core advantage of the SMPL model lies in its ability to generate a high-precision 3D human body mesh based on the input image and a limited set of inputs, such as height and weight. Then, motion videos of the user performing several identical movements during a muscle fitness test are collected using computer vision technology. Subsequently, a pre-defined human pose estimation algorithm is used to extract the joint motion sequence corresponding to each movement from the motion videos. Based on this joint motion sequence, the pose parameters corresponding to the user performing several identical movements during the muscle fitness test are determined.

[0060] Specifically, a camera can capture motion videos of a user performing the same movements several times during a muscle fitness test. Then, 2D / 3D human pose estimation algorithms, such as OpenPose and ViTPose, can be used to extract the joint motion sequence corresponding to each movement from the motion video to obtain the corresponding pose parameters. As can be seen, this embodiment only requires a regular camera, or even a mobile phone, for motion capture, and uses computer vision algorithms to extract human motion information from the video, eliminating the dependence on specific physical devices. All calculations and data analysis are completed in the cloud or within the local algorithm.

[0061] It should be noted that this embodiment can utilize the SMPL model to reconstruct a high-precision 3D human body model and motion trajectory based on the posture parameters corresponding to each action. Specifically, the SMPL model satisfies the following formula:

[0062] ;

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] in, The shape parameter is typically a 10-dimensional linear shape coefficient vector used to control individual differences in the human body, such as height, weight, and build. As a posture parameter, the relative joint rotation of the human skeleton is defined, and the entire mesh is deformed by linear blending (LBS). To be based on shape with posture The generated body shape and skeleton; This is the adjusted joint center position; It is the skinning weight of each vertex with respect to each joint. . These are template vertices, typically consisting of 6890 fixed points, representing a general human body shape template. The direction of the shape can be understood as the direction of "getting fatter", "getting taller", or "getting wider shoulders"; Using shape as the basis, the 10 PCA principal component directions are linearly combined. . It is a pose-dependent correction term used to compensate for non-rigid deformations caused by joint rotation, such as bending the elbow or knee, such as muscle bulging or skin stretching. `k` represents the global transformation of the k-th joint, which is the combined rotation and translation of the parent and child skeletons. Its function is to transform the k-th joint from the local skeleton coordinate system to the global coordinate system. Typically, SMPL models define 24 joints. `t` represents the global translation, usually the position of the root node or in the camera coordinate system, indicating the position of the entire body model in the world coordinate system or camera coordinate system.

[0068] In this embodiment, the total output power of the user when performing each action is determined based on the target motion trajectories. The specific process may include: first, determining the angular velocity and angular acceleration of each joint of the user when performing each action based on the target motion trajectories, and converting the angular velocity of each joint of each action into the angular velocity of each limb based on the transformation matrix corresponding to the posture parameters of each action; then, using the Zatsiorsky model, determining the mass of each limb based on the user's height, weight, and the proportion of each limb, and determining the moment of inertia of each limb based on the mass of each limb; subsequently, determining the net joint torque of each joint of the user when performing each action based on the moment of inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint; then, determining the instantaneous output power of each joint of the user when performing each action based on the net joint torque and the angular velocity of each joint; finally, determining the total output power of the user when performing each action based on the instantaneous output power of each joint of the user.

[0069] Specifically, the joint position information of the time series is first obtained from the SMPL model, as shown below:

[0070] ;

[0071] in, This represents the spatial position vector of joint i at time t, i.e., the position of the i-th joint.

[0072] Next, in order to calculate the angular velocity and angular acceleration of each joint, the local joint axis is first defined using adjacent joints, and the joint angles are calculated as follows:

[0073] ;

[0074] in, It is the bone vector between adjacent joints, representing the vector from the i-th joint to the (i+1)-th joint, describing the direction of the bone.

[0075] Subsequently, the joint angles were time-differentially analyzed to obtain the angular velocity and angular acceleration of each joint, as shown below:

[0076] ;

[0077] in, Let be the angular velocity of joint i at time t. Let be the angular acceleration of joint i at time t.

[0078] Then, the transformation matrix corresponding to the pose parameters of each action can be used. Convert the angular velocities of each joint into the angular velocities of each limb. Subsequently, the Zatsiorsky model can be used to estimate the mass distribution and inertial parameters of each limb based on the user's height, weight, limb proportions, etc., as shown below:

[0079] ;

[0080] in, Let M be the mass of the j-th limb, and M be the total mass of the user. Let j represent the percentage of the j-th limb, such as the thigh, which accounts for approximately 10.5% of the body mass.

[0081] It should be noted that, in order to determine the moment of inertia of each limb based on its mass, for a simplified rigid body model, the following formula can be used to approximate the moment of inertia of each human limb, such as the upper arm, thigh, and lower leg. The moment of inertia can be understood as the "anti-rotational ability" of an object relative to a certain axis of rotation, as shown below:

[0082] ;

[0083] in, Let be the moment of inertia of the j-th limb; Let be the segment length of the j-th limb, representing the geometric length of that limb, such as from hip to knee, knee to ankle, shoulder to elbow, etc. It is an empirical coefficient used to take into account the shape of the mass distribution of human limbs, not all of which are at the center.

[0084] Subsequently, inverse dynamics can be used to determine the net joint torque of each joint during each user action based on the rotational inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint. The basic form is rigid body dynamics, as shown below:

[0085] ;

[0086] in, Let be the net joint torque of the i-th joint; Let be the rotational inertia of the i-th limb; Let be the angular velocity of the i-th limb.

[0087] Then, based on the net joint torque and angular velocity of each joint, the instantaneous output power of each joint can be determined, as shown below:

[0088] ;

[0089] in, The instantaneous output power of the i-th joint is used to evaluate explosive power, dynamic capabilities, etc. Let be the net joint torque of the i-th joint; Let be the angular velocity of the i-th joint.

[0090] Finally, the total output power of the user during each action can be determined based on the instantaneous output power of each joint during each action, as shown below:

[0091] ;

[0092] in, This represents the total output power at time t when the user performs the i-th action. This represents the instantaneous output power of each joint at time t when the user performs the i-th action. That is, the system simultaneously measures the power of multiple joints, such as the hip, knee, and ankle, at time t. Then, add them together to get the total output power of the whole body.

[0093] It should be noted that, after determining the net joint torque of each joint when the user performs each action based on the rotational inertia, angular velocity, and angular acceleration of each limb, the process may further include: first, determining the muscle activation state of the user when performing each action based on the target motion trajectory corresponding to each action; then, determining the muscle group strength of each muscle when the user performs each action based on the muscle activation state and the net joint torque of each joint, and determining the muscle activation degree of the user when performing each action based on the muscle group strength of each muscle; finally, visually displaying the muscle activation degree of the user when performing each action through a preset visualization platform.

[0094] Specifically, this embodiment can calculate the muscle activation state based on the target motion trajectory corresponding to each action performed by the user, including position, velocity, acceleration, etc., and the cost function is as follows:

[0095] ;

[0096] in, This represents the amount of muscle activation. Used to punish the degree of muscle activation, preventing overactivation to conserve energy / reduce fatigue; ETM is an ideal torque motor. The control input for punishing external auxiliary devices, such as exoskeletons and ideal torque motors, is used to minimize the auxiliary torque. Indicates the desired joint position, joint velocity, and joint acceleration; This represents the actual joint position, joint velocity, and joint acceleration. The goal is to determine the optimal muscle activation state and the ideal torque motor control signal within the musculoskeletal system.

[0097] Since the muscular system is redundant—meaning that a joint torque can be generated by multiple muscles in different combinations—this embodiment can utilize the solved muscle activation levels. The system's dynamic model will determine the net joint torques required to drive the body to complete the desired movement at each joint. This is broken down into specific muscles. Therefore, in the muscle redundancy problem (i.e., one torque is achieved by multiple muscles), an optimization method is needed to inversely calculate the output force of each muscle group. The muscle strength optimization problem is a static optimization, and the optimization objective is to minimize the sum of squares of the activation levels of all muscles, as shown below:

[0098] make Given the force vector of the muscle group, solve the following formula:

[0099] ;

[0100] condition: ;

[0101] Where R is the muscle lever arm matrix, This represents the net joint torque for each joint. The constraint here is that the sum of the torques generated by the muscles equals the known net joint torque, and the force of each muscle cannot exceed its physiological power limit. In this way, the muscle group strength of each muscle group can be calculated when the user performs each movement.

[0102] Then, the muscle activation level of the user during each movement can be determined using the following formula and the muscle group strength of each muscle, as shown in the formula below:

[0103] ;

[0104] in, , representing the muscle activation level of the i-th muscle group. It should be noted that, see [reference needed]. Figure 2 As shown, for a user's SMPL model, musculoskeletal model parameters can be calculated, and then the musculoskeletal model can be used for visualization. See also... Figure 3As shown, in this embodiment, muscle activation can be visualized on a preset visualization platform using a musculoskeletal model. Specifically, it can be used to construct heatmaps, muscle group participation spectra, etc., to meet the needs of exercisers for obtaining information about the muscle groups they are exercising. In this way, this embodiment can use the SMPL model to obtain accurate geometric information of the user's entire body, drive the musculoskeletal model to perform dynamic inversion, and the musculoskeletal model can be personalized and scaled based on the SMPL model.

[0105] Step S12: Determine the user's initial power and power decay rate based on the total output power corresponding to each action, and determine the user's 1RM based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays when performing the muscle fitness test.

[0106] In this embodiment, the total output power corresponding to each action can be collected. The user's fatigue attenuation rate, i.e., power attenuation rate k, and the initial power are obtained through fitting. Among these, the power decay rate quantifies the rate of individual fatigue, while the initial power reflects the individual's true potential. Muscle fatigue can manifest as a decrease in output power; see [link to relevant documentation]. Figure 4 As shown, an exponential decay model is used for modeling, and the formula is as follows:

[0107] ;

[0108] in, The total output power when the user performs the i-th repetitive action; Initial power can be considered an indirect indicator of an individual’s maximum muscle strength in the power dimension, i.e., the user’s maximum instantaneous work capacity before performing a muscle fitness test; k is the power decay rate, which is individual-related. This is to account for measurement error. Therefore, the fatigue attenuation rate k and the initial power can be obtained through modeling. .

[0109] It should be noted that the 1RM corresponding to the user is determined based on the subjective score and total output power corresponding to each action, as well as the initial power. The specific process may include: if the ratio between the total output power of the user when performing the current action and the initial power is less than or equal to a preset power decrease ratio threshold, and the subjective score corresponding to the current action is greater than or equal to a preset subjective score threshold, then the target number of movements corresponding to the current action in the muscle fitness test is determined; then the target load corresponding to the current action is determined based on the net joint torque of each joint when the user performs the current action, and the 1RM corresponding to the user is determined based on the target load and the target number of movements.

[0110] Specifically, in this embodiment, a threshold for exhaustion is defined, which is a comprehensive judgment condition for "approaching exhaustion," as shown below:

[0111] ;

[0112] in, A subjective rating for the user's i-th action; Preset a subjective rating threshold, such as 9 or 10; A preset power reduction rate threshold, such as 40%, is set. If the current action meets the judgment condition, the system records the current action. In muscle fitness tests, the target number of repetitions is taken as the point near the limit, that is, the point near exhaustion, also known as the "near-exhaustion point".

[0113] Next, the target load corresponding to the current action is determined by the net joint torque of each joint when the user performs the current action, as shown below:

[0114] ;

[0115] Where L represents the target load corresponding to the current action, that is, the load the user is performing when executing the first action. The load during the movement; BW is the body weight above the target joint center, combined with an empirical coefficient for the mass distribution shape of human limb segments. Calculated; It is the equivalent force arm, which is the horizontal distance from the center of the target joint to the line of force of the load.

[0116] Finally, the theoretical 1RM (one maximum repetition) can be estimated from the target number of repetitions corresponding to the "near-exhaustion point" using the Epley formula or other empirical formulas. 1RM represents the maximum resistance encountered within the full range of motion of the joint under correct posture and certain rules, and has become a standard evaluation indicator for dynamic strength. The 1RM estimation formula is shown below:

[0117] ;

[0118] in, The target number of repetitions is the number of times the user completes the action when they are close to exhaustion.

[0119] Step S13: Determine the user's muscle strength score based on the 1RM, and determine the user's muscle endurance score based on the total output power and the power decay rate corresponding to each movement.

[0120] In this embodiment, the user's muscle strength score is determined based on 1RM, and the user's muscle endurance score is determined based on the total output power and power decay rate corresponding to each movement. The specific process may include: first, determining the user's muscle strength score based on the ratio between the target load and the 1RM; then, determining the average output power corresponding to the user completing the muscle fitness test based on the total output power corresponding to each movement and the number of movements corresponding to the muscle fitness test, and determining the standard deviation of the output power based on the total output power and the average output power corresponding to each movement; subsequently, determining a first score based on the ratio between the average output power and the standard deviation, and determining a second score based on the ratio between the power decay rate and a preset power decay rate threshold; finally, performing a weighted calculation on the first score and the second score to obtain the user's muscle endurance score.

[0121] Specifically, in this embodiment, the ratio of the user's actual force level in a total of N repetitions of the muscle fitness test to their maximum capacity (1RM) is measured. This ratio reflects the user's short-term explosive power or maximum muscle output capacity, resulting in the user's muscle strength score. As shown below:

[0122] ;

[0123] It should be noted that if multiple joints are involved simultaneously, such as in a squat, composite indicators of multiple muscle groups or joints can be collected to obtain the user's muscle strength score.

[0124] Subsequently, to assess whether the user's output remained stable, whether the posture was out of control, and whether there was a significant fatigue decline curve during continuous repetitive movements, a muscle endurance score was obtained. First, calculate the coefficient of variation of the output power change to obtain the first score. As shown below:

[0125] ;

[0126] in: The coefficient of variation represents the change in output power; This represents the total output power for each action; This represents the average level of total output power over n consecutive actions by the user, i.e., the average output power, as shown below:

[0127] ;

[0128] This represents the fluctuation in the power output by the user during n consecutive actions, i.e., the power value relative to the average value. The degree of dispersion is used to obtain the standard deviation of the output power, as shown below:

[0129] .

[0130] Understandably, a high CV value means that a user's power fluctuates greatly relative to its average level, indicating poor endurance. A low CV value means that even if a user's average power is high or low, its output is very stable, indicating good endurance.

[0131] Then, the second score can be determined based on the ratio between the power attenuation rate and the preset power attenuation rate threshold. The formula is shown below:

[0132] ;

[0133] Where k represents the power attenuation rate, This represents the maximum reference descent slope set by the system, i.e., the preset power attenuation rate threshold, which is obtained through individual experience.

[0134] Next, based on the goodness of fit To adjust the first score Second score Regarding the weight allocation, if the fatigue curve fits poorly, it indicates that k is unreliable, and the weight of the second score should be reduced. It should be noted that... The calculation process is as follows:

[0135] 1. Calculate the total sum of squares (SST), which measures the total variability of the power data itself:

[0136] ;

[0137] in, This is the average of the power observations, as shown below:

[0138] ;

[0139] 2. Calculate the sum of squared residuals (SSR). SSR measures the degree of variability that the model failed to explain.

[0140] ;

[0141] in, The power values ​​predicted by the model are shown below:

[0142] ;

[0143] 3. Calculation :

[0144] ;

[0145] Finally, the first and second scores are weighted according to their assigned weights to obtain the user's muscle endurance score, as shown in the formula below:

[0146] .

[0147] Step S14: Determine the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and determine and output the corresponding muscle fitness test result of the user based on the muscle strength score, the muscle endurance score, the muscle coordination score and the quality score.

[0148] In this embodiment, the process of determining the muscle coordination score and quality score corresponding to a user's completion of a muscle fitness test can include: first, determining the muscle coordination score based on the muscle group strength of each muscle group when the user performs each movement; then, determining a preset standard movement trajectory template and a preset trajectory deviation threshold corresponding to the muscle fitness test, and determining the trajectory deviation between each target movement trajectory and the preset standard movement trajectory template; subsequently, using a dynamic time warping method, determining a third score based on the ratio between each trajectory deviation and the preset trajectory deviation threshold; then, determining several postures to be verified when the user performs each movement, and determining the cosine similarity between each posture to be verified and the preset standard posture template; finally, determining a fourth score based on the cosine similarity, and performing a weighted calculation on the third score and the fourth score to obtain the quality score corresponding to the user's completion of the muscle fitness test.

[0149] Specifically, this embodiment assesses the coordinated movement ability between multiple joints / muscle groups, such as whether the thigh / hip / knee exert force synchronously and whether the posture remains stable, to obtain a muscle coordination score. The force output curves of the two main muscle groups during the movement are shown below. The formula for calculating the muscle coordination score is as follows:

[0150] ;

[0151] Here, corr is the Pearson correlation coefficient. It should be noted that if there are two or more major muscle groups involved in the movement, a correlation matrix can be constructed, and the average correlation value or the mean of the main diagonal can be extracted to calculate the muscle coordination score.

[0152] Next, a quality score was obtained to measure whether the form, rhythm, and stability of the user's completed actions closely resembled the standard template. The third score can be calculated first using the DTW (Dynamic Time Warping) method. The DTW method measures the time alignment error between the user's action sequence and a standard sample, i.e., a preset standard action trajectory template, and determines the third score based on a preset trajectory deviation threshold. The formula is shown below:

[0153] ;

[0154] in, The DTW distance between the user's target motion trajectory and the preset standard motion trajectory template, i.e., the trajectory deviation; This is a preset trajectory deviation threshold.

[0155] It's important to note that the DTW (Dynamic Time Management) method measures the smoothness and temporal consistency of a user's overall movement process, assessing coordination, fluidity, and overall pattern, emphasizing the process itself. Examples include "walking gait," "a set of Tai Chi movements," and "the continuity of a fitness exercise routine." The quality of these movements is reflected in the smoothness of the entire process, not just a single momentary posture. Furthermore, since older adults or rehabilitation patients often complete movements more slowly, DTW can effectively align their slow movements with standard fast movements for comparison, without penalizing them for slow speed.

[0156] Subsequently, several poses to be verified by the user during each action can be determined, and the cosine similarity between each pose and a preset standard pose template can be determined. The cosine similarity is then used as the action similarity index to obtain the fourth score. The formula is shown below:

[0157] ;

[0158] in, The posture is yet to be verified; Preset standard posture template; .

[0159] It's important to note that motion similarity metrics can measure the static standardization of a user's posture at a specific moment, such as a keyframe, assessing the degree of execution of the movement and the accuracy of posture at extreme positions, focusing on the result. For example, it can assess whether the knees buckle inward at the lowest point of a squat, whether the torso leans excessively forward, whether the barbell touches the chest at the lowest point of a bench press, and whether the elbow angle is appropriate. In rehabilitation training, when doctors ask patients to "raise their arms to shoulder height," it's only necessary to compare whether the posture at the highest point of the arm is up to standard.

[0160] Finally, the third and fourth scores can be weighted to obtain the quality score corresponding to the user's completion of the muscle fitness test. The quality score can be used for rehabilitation / health assessments, etc., and the calculation formula is shown below:

[0161] ;

[0162] It should be noted that the weighting of the third and fourth scores depends on the assessment focus: if the emphasis is on the smoothness and rhythm of the movement process, such as in rehabilitation and dance, then [the following criteria are applied]. Higher weighting, such as 0.7; if the standardization of key postures is emphasized, such as powerlifting or achieving the required posture, then assign... A higher weighting, such as 0.7; if equal attention is paid to the process and the result, such as fitness, it is recommended to allocate them evenly, such as 0.5 each.

[0163] In this embodiment, the user's corresponding muscle fitness test results are determined and output based on muscle strength score, muscle endurance score, muscle coordination score, and quality score, thus obtaining the user's overall fitness score. . It is a single, comprehensive score calculated after a user completes n repetitions of a standard movement in a muscle fitness test, used to assess the user's overall muscle fitness level during that training session. Each indicator can be standardized to... The scores are divided into points, weighted, and combined to form the overall fitness score:

[0164] ;

[0165] in, It can be 30%. It can be 25%. It can be 20%. It can be 25%.

[0166] As can be seen from the above, this embodiment first acquires motion videos of the user performing a muscle fitness test based on computer vision technology, and processes the motion videos to determine the user's posture parameters when performing several identical movements in the muscle fitness test. Then, using the user's corresponding SMPL model, target motion trajectories are generated based on the posture parameters corresponding to each movement, and the total output power of the user during each movement is determined based on the target motion trajectories. Next, the user's initial power and power decay rate are determined based on the total output power corresponding to each movement, and the subjective score, total output power, and initial power are used to determine the... The 1RM corresponding to the user is described; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays during the muscle fitness test; then, the user's muscle strength score is determined based on the 1RM, and the user's muscle endurance score is determined based on the total output power and the power decay rate for each movement; finally, the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test are determined, and the muscle fitness test result corresponding to the user is determined and output based on the muscle strength score, the muscle endurance score, the muscle coordination score, and the quality score. As can be seen from the above, this embodiment first uses computer vision technology to collect and process motion videos of the user during muscle fitness testing, obtaining posture parameters of the user performing several identical movements in the muscle fitness test. Combined with the user's SMPL model, a target motion trajectory is generated, and the total output power of each movement is determined. Then, based on the total output power, an initial power and power decay rate are fitted, and the 1RM value is determined by combining the subjective score of each movement with the total output power. Subsequently, muscle strength scores are calculated based on the 1RM, muscle endurance scores are calculated based on the total output power and power decay rate, and muscle coordination and movement quality scores are determined simultaneously. Finally, the muscle fitness test results are obtained and output. In this way, firstly, this embodiment does not rely on expensive professional equipment, achieving non-contact testing using only a regular camera; secondly, this embodiment uses the SMPL model to achieve high-precision, dynamic, and personalized muscle output simulation, and the simulation results show a significant correlation with real physiological data; and thirdly, this embodiment comprehensively covers core dimensions such as muscle strength, endurance, coordination, and movement quality. In this way, this embodiment improves the scientific nature and accuracy of the test results by using standardized data processing procedures and personalized model fitting, making it suitable for users of different ages and health conditions, and significantly expanding the application scenarios and universality of muscle fitness testing.

[0167] Accordingly, see Figure 5 As shown in the illustration, this application also provides a computer vision-based muscle fitness testing device, applied to a computer device, which may include:

[0168] The total output power determination module 11 is used to acquire motion videos of a user performing a muscle fitness test based on computer vision technology, process the motion videos to determine the posture parameters of the user when performing several identical movements in the muscle fitness test, and use the SMPL model corresponding to the user to generate target motion trajectories based on the posture parameters corresponding to each movement, so as to determine the total output power of the user when performing each movement based on each target motion trajectory.

[0169] The power decay rate determination module 12 is used to determine the user's initial power and power decay rate based on the total output power corresponding to each action, and to determine the user's 1RM based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate of power decay when the user performs the muscle fitness test.

[0170] The muscle endurance rating determination module 13 is used to determine the user's muscle strength rating based on the 1RM, and to determine the user's muscle endurance rating based on the total output power and the power decay rate corresponding to each movement;

[0171] The test result determination and output module 14 is used to determine the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and to determine and output the corresponding muscle fitness test result of the user based on the muscle strength score, the muscle endurance score, the muscle coordination score and the quality score.

[0172] In some specific embodiments, the total output power determination module 11 may include:

[0173] The motion video acquisition unit is used to pre-build the SMPL model corresponding to the user, and acquire motion videos of the user performing the same movements in several times during the muscle fitness test based on computer vision technology.

[0174] The posture parameter determination unit is used to extract the joint motion sequence corresponding to each action from the motion video through a preset human posture estimation algorithm, and to determine the posture parameters corresponding to the user when performing the same action in several times in the muscle fitness test based on the joint motion sequence corresponding to each action.

[0175] In some specific embodiments, the total output power determination module 11 may include:

[0176] An angular velocity conversion unit is used to determine the angular velocity and angular acceleration of each joint of the user when performing each action based on the target motion trajectory, and to convert the angular velocity of each joint of each action into the angular velocity of each limb based on the transformation matrix corresponding to the posture parameters of each action.

[0177] The moment of inertia determination unit is used to determine the mass of each limb based on the user's height, weight and the proportion of each limb using the Zatsiorsky model, and to determine the moment of inertia of each limb based on the mass of each limb.

[0178] The joint net torque determination unit is used to determine the joint net torque of each joint when the user performs each action based on the rotational inertia of each limb, the angular velocity of each limb and the angular acceleration of each joint.

[0179] An instantaneous output power determination unit is used to determine the instantaneous output power of each joint when the user performs each action, based on the net joint torque and angular velocity of each joint when the user performs each action.

[0180] The total output power determination unit is used to determine the total output power of the user when performing each action based on the instantaneous output power of each joint when the user performs each action.

[0181] In some specific embodiments, the computer vision-based muscle fitness testing device may further include:

[0182] A muscle activation state determination unit is used to determine the muscle activation state of the user when performing each action based on the target motion trajectory corresponding to the user when performing each action.

[0183] The muscle activation determination unit is used to determine the muscle group strength of each muscle when the user performs each action based on the muscle activation state and the net joint torque of each joint, and to determine the muscle activation degree of the user when performing each action based on the muscle group strength of each muscle.

[0184] The visualization unit is used to visualize the muscle activation level of the user when performing each action through a preset visualization platform.

[0185] In some specific embodiments, the power attenuation rate determination module 12 may include:

[0186] The target number of movement determination unit is used to determine the target number of movement corresponding to the current action in the muscle fitness test if the ratio between the total output power of the user when performing the current action and the initial power is less than or equal to a preset power decrease ratio threshold, and the subjective score corresponding to the current action is greater than or equal to a preset subjective score threshold.

[0187] The target load determination unit is used to determine the target load corresponding to the current action based on the net joint torque of each joint when the user performs the current action, and to determine the 1RM corresponding to the user based on the target load and the target number of movements.

[0188] In some specific embodiments, the muscle endurance scoring module 13 may include:

[0189] A muscle strength rating determination unit is used to determine the user's muscle strength rating based on the ratio between the target load and the 1RM;

[0190] The standard deviation determination unit is used to determine the average output power of the user completing the muscle fitness test based on the total output power corresponding to each action and the number of actions corresponding to the muscle fitness test, and to determine the standard deviation of the output power based on the total output power and the average output power corresponding to each action.

[0191] The second score determination unit is used to determine a first score based on the ratio between the average output power and the standard deviation, and to determine a second score based on the ratio between the power attenuation rate and a preset power attenuation rate threshold.

[0192] A muscle endurance scoring unit is used to perform a weighted calculation on the first score and the second score to obtain the user's muscle endurance score.

[0193] In some specific embodiments, the test result determination and output module 14 may include:

[0194] A muscle coordination score determination unit is used to determine the muscle coordination score corresponding to the user's completion of the muscle fitness test based on the muscle group strength of each muscle when the user performs each action.

[0195] The trajectory deviation determination unit is used to determine the preset standard movement trajectory template and preset trajectory deviation threshold corresponding to the muscle fitness test, and to determine the trajectory deviation between each target movement trajectory and the preset standard movement trajectory template.

[0196] The third score determination unit is used to determine the third score based on the ratio between each trajectory deviation and the preset trajectory deviation threshold using a dynamic time warping method.

[0197] The cosine similarity determination unit is used to determine several poses to be verified when the user performs each action, and to determine the cosine similarity between each pose to be verified and a preset standard pose template.

[0198] The quality score determination unit is used to determine a fourth score based on the cosine similarity, and to perform a weighted calculation on the third score and the fourth score to obtain the quality score corresponding to the user's completion of the muscle fitness test.

[0199] Furthermore, embodiments of this application also disclose an electronic device, Figure 6 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the computer vision-based muscle fitness testing method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0200] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0201] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0202] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the computer vision-based muscle fitness testing method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0203] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned computer vision-based muscle fitness testing method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0204] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0205] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0206] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0207] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0208] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A computer vision-based muscle fitness testing method, characterized in that, Applied to computer devices, including: The system uses computer vision technology to collect motion videos of users performing muscle fitness tests, processes the motion videos to determine the posture parameters of the user when performing the same action several times in the muscle fitness test, and uses the SMPL model corresponding to the user to generate target motion trajectories based on the posture parameters corresponding to each action, so as to determine the total output power of the user when performing each action based on the target motion trajectories. The user's initial power and power decay rate are determined based on the total output power corresponding to each action, and the user's 1RM is determined based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays when performing the muscle fitness test. The user's muscle strength score is determined based on the 1RM, and the user's muscle endurance score is determined based on the total output power and the power decay rate corresponding to each movement. Determine the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and determine and output the corresponding muscle fitness test result of the user based on the muscle strength score, the muscle endurance score, the muscle coordination score and the quality score; The formula for calculating the quality score is as follows: ； in, To rate the quality, It is the third score. , The DTW distance between the user's target motion trajectory and the preset standard motion trajectory template represents the trajectory deviation. The preset trajectory deviation threshold; It is the fourth score. , The posture to be verified. To preset standard posture templates, , The weight of the third score, The weight for the fourth score.

2. The computer vision-based muscle fitness testing method according to claim 1, characterized in that, The process of acquiring motion videos of users performing muscle fitness tests using computer vision technology and processing these videos to determine the user's posture parameters during several identical movements in the muscle fitness test includes: The SMPL model corresponding to the user is pre-constructed, and the motion video of the user performing the same action several times in the muscle fitness test is collected based on computer vision technology. The motion sequence of joints corresponding to each action is extracted from the motion video by a preset human posture estimation algorithm, and the posture parameters corresponding to the user when performing the same action in several times in the muscle fitness test are determined based on the joint motion sequence of each action.

3. The computer vision-based muscle fitness testing method according to claim 1, characterized in that, Determining the total output power of the user when performing each action based on the target motion trajectories includes: Based on the target motion trajectory, the angular velocity and angular acceleration of each joint of the user when performing each action are determined, and based on the transformation matrix corresponding to the posture parameters of each action, the angular velocity of each joint of each action is converted into the angular velocity of each limb. Using the Zatsiorsky model, the mass of each limb is determined based on the user's height, weight, and the proportion of each limb, and the moment of inertia of each limb is determined based on the mass of each limb. The net joint torque of each joint is determined based on the rotational inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint when the user performs each action. Based on the net joint torque and angular velocity of each joint when the user performs each action, determine the instantaneous output power of each joint when the user performs each action. The total output power of the user during each action is determined based on the instantaneous output power of each joint when the user performs each action.

4. The computer vision-based muscle fitness testing method according to claim 3, characterized in that, After determining the net joint torque of each joint when the user performs each action based on the rotational inertia of each limb, the angular velocity of each limb, and the angular acceleration of each joint, the method further includes: Based on the target motion trajectory corresponding to each action performed by the user, the muscle activation state of the user during each action is determined; Based on the muscle activation state and net joint torque of each joint when the user performs each action, the muscle group strength of each muscle when the user performs each action is determined, and the muscle activation degree of the user when performing each action is determined based on the muscle group strength of each muscle. The user's muscle activation level during each action is visualized using a pre-set visualization platform.

5. The computer vision-based muscle fitness testing method according to claim 4, characterized in that, The determination of the user's 1RM based on the subjective score corresponding to each action, the total output power, and the initial power includes: If the ratio between the total output power of the user when performing the current action and the initial power is less than or equal to a preset power decrease ratio threshold, and the subjective score corresponding to the current action is greater than or equal to a preset subjective score threshold, then the target number of movements corresponding to the current action in the muscle fitness test is determined. The target load corresponding to the current action is determined based on the net joint torque of each joint when the user performs the current action, and the 1RM corresponding to the user is determined based on the target load and the target number of movements.

6. The computer vision-based muscle fitness testing method according to claim 5, characterized in that, The process of determining the user's muscle strength score based on the 1RM and determining the user's muscle endurance score based on the total output power and the power decay rate for each movement includes: The user's muscle strength score is determined based on the ratio between the target load and the 1RM. Based on the total output power corresponding to each movement and the number of movements corresponding to the muscle fitness test, the average output power corresponding to the user completing the muscle fitness test is determined, and the standard deviation of the output power is determined based on the total output power corresponding to each movement and the average output power. A first score is determined based on the ratio between the average output power and the standard deviation, and a second score is determined based on the ratio between the power attenuation rate and a preset power attenuation rate threshold. The first score and the second score are weighted and calculated to obtain the user's muscle endurance score.

7. The computer vision-based muscle fitness testing method according to any one of claims 4 to 6, characterized in that, Determining the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test includes: Based on the muscle group strength of each muscle when the user performs each action, the muscle coordination score corresponding to the user's completion of the muscle fitness test is determined. Determine the preset standard movement trajectory template and preset trajectory deviation threshold corresponding to the muscle fitness test, and determine the trajectory deviation between each target movement trajectory and the preset standard movement trajectory template; The third score is determined by using a dynamic time warping method based on the ratio between each trajectory deviation and the preset trajectory deviation threshold. Determine several poses to be verified when the user performs each action, and determine the cosine similarity between each pose to be verified and a preset standard pose template; The fourth score is determined based on the cosine similarity, and the third and fourth scores are weighted to obtain the quality score corresponding to the user's completion of the muscle fitness test.

8. A computer vision-based muscle fitness testing device, characterized in that, Applied to computer devices, including: The total output power determination module is used to acquire motion videos of a user performing a muscle fitness test based on computer vision technology, process the motion videos to determine the posture parameters of the user when performing several identical movements in the muscle fitness test, and use the SMPL model corresponding to the user to generate target motion trajectories based on the posture parameters corresponding to each movement, so as to determine the total output power of the user when performing each movement based on each target motion trajectory. The power decay rate determination module is used to determine the user's initial power and power decay rate based on the total output power corresponding to each action, and to determine the user's 1RM based on the subjective score corresponding to each action, the total output power, and the initial power; the initial power represents the user's maximum instantaneous work capacity before performing the muscle fitness test, and the power decay rate represents the rate at which the user's power decays when performing the muscle fitness test. A muscle endurance rating determination module is used to determine the user's muscle strength rating based on the 1RM, and to determine the user's muscle endurance rating based on the total output power and the power decay rate corresponding to each movement; The test result determination and output module is used to determine the muscle coordination score and quality score corresponding to the user's completion of the muscle fitness test, and to determine and output the corresponding muscle fitness test result of the user based on the muscle strength score, the muscle endurance score, the muscle coordination score and the quality score; The formula for calculating the quality score is as follows: ； in, To rate the quality, It is the third score. , The DTW distance between the user's target motion trajectory and the preset standard motion trajectory template represents the trajectory deviation. The preset trajectory deviation threshold; It is the fourth score. , The posture to be verified. To preset standard posture templates, , The weight of the third score, The weight for the fourth score.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory; wherein the memory is used to store a computer program, which is loaded and executed by the processor to implement the computer vision-based muscle fitness testing method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the computer vision-based muscle fitness testing method as described in any one of claims 1 to 7.

Images