An intelligent comparison and analysis method for prompting improvement of fitness actions and a terminal

Abstract

This invention discloses an intelligent comparison and analysis method and terminal for prompting improvement of fitness movements, including the following steps: S1, obtaining the three-dimensional coordinates of human joints in the coach's movement video and the user's movement video using a three-dimensional pose estimation method; S2, determining the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent joints in the joint coordinates; S3, aligning the three-dimensional angle differences of the limb polar coordinates in both using a dynamic time warping method, summing the absolute values ​​of the polar coordinate angle differences corresponding to each limb in all frames to obtain the cumulative sum of angle differences for each limb, and identifying the limb with the largest cumulative sum of angle differences as the limb requiring movement improvement; S4, providing intelligent prompts for the limb requiring movement improvement. This overcomes problems such as different shooting angles from 2D cameras, asynchronous user movements, and abnormal prompts concentrated in the forearm.

Description

Technical Field

[0001] This invention relates to the field of pose estimation technology in computer vision, and in particular to an intelligent comparison and analysis method and terminal for prompting improvement of fitness movements. Background Technology

[0002] In recent years, pose estimation technology has developed rapidly, making significant progress from single-person pose recognition to multi-person pose recognition, and from 2D pose recognition to 3D pose recognition. However, some problems still exist in its application. Most existing devices are based on 3D binocular cameras, but in applications using conventional 2D cameras, including mobile phone cameras, tablet cameras, and laptop cameras, there is one less dimension of information compared to the 3D binocular cameras in motion-sensing game devices, making processing more difficult. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide an intelligent comparison and analysis method for prompting improvement of fitness movements, thereby overcoming or solving a series of problems existing in movement comparison, including different perspectives of 2D camera shooting, asynchronous movements, and abnormal prompts being concentrated in the forearm with a high probability.

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0005] A smart comparative analysis method for suggesting improvements in fitness movements includes the following steps:

[0006] S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video.

[0007] S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates.

[0008] S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved.

[0009] S4. Provide intelligent prompts for the limbs whose movements need improvement.

[0010] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is as follows:

[0011] A smart comparison and analysis terminal that provides suggestions for improving fitness movements includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs the following steps:

[0012] S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video.

[0013] S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates.

[0014] S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved.

[0015] S4. Provide intelligent prompts for the limbs whose movements need improvement.

[0016] The beneficial effects of this invention are as follows: It provides an intelligent comparison and analysis method and terminal for prompting improvement of fitness movements. It uses a dynamic time warping algorithm to compare the differences in the three-dimensional angles of the limbs in the coach's video and the user's video, optimizes the three-dimensional angle difference data of the forearm, and overcomes problems such as different shooting angles of 2D cameras, asynchronous user movements, and abnormal prompts concentrated in the forearm. Attached Figure Description

[0017] Figure 1 This is a flowchart of an intelligent comparison and analysis method for prompting improvement of fitness movements according to an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of an intelligent comparison and analysis terminal that provides suggestions for improving fitness movements, according to an embodiment of the present invention. Detailed Implementation

[0019] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0020] Definitions:

[0021] Dynamic Time Warping (RTW) is a non-linear warping technique that measures the similarity between two time series of different lengths based on the nearest neighbor principle.

[0022] MediaPipe is a framework for building machine learning pipelines to process time-series data such as video and audio.

[0023] A smart comparative analysis method for suggesting improvements in fitness movements includes the following steps:

[0024] S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video.

[0025] S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates.

[0026] S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved.

[0027] S4. Provide intelligent prompts for the limbs whose movements need improvement.

[0028] As can be seen from the above description, the beneficial effects of the present invention are as follows: It provides an intelligent comparison and analysis method for prompting improvement of fitness movements, which uses a dynamic time warping algorithm to compare the differences in the three-dimensional angles of the limbs in the coach's movement video and the user's movement video, optimizes the three-dimensional angle difference data of the forearm, and overcomes problems such as different shooting angles of 2D cameras, asynchronous user movements, and abnormal prompts concentrated in the forearm.

[0029] Furthermore, after step S4, the method further includes:

[0030] S5. Construct a scoring function using activation functions to score the accuracy of the user's overall actions relative to the coach's actions;

[0031] The scoring function is: Where tanh is the activation function, Δθ i Due to differences in orientation angles, For distance difference, Time is the video duration, and k is a constant value that needs to be optimized.

[0032] As described above, by comprehensively analyzing a set of intelligent comparison data of fitness movements, users are provided with a more intuitive understanding of the accuracy of their fitness movements, making it easier for them to self-test their exercise results and adjust their fitness plans.

[0033] Further, step S1 specifically includes:

[0034] The first joint coordinates are obtained through a pre-recorded video of the coach's movements;

[0035] The second joint coordinates are obtained from a video recording of the user's actions on-site.

[0036] As described above, joint coordinates are pre-calculated using pre-recorded coaching video, and data such as limb coordinates and three-dimensional angle differences are obtained and saved, reducing redundant calculations.

[0037] Furthermore, the three-dimensional angle difference in step S3 specifically refers to:

[0038] The calculation of the angle difference for each limb used for dynamic time warping alignment includes subtracting the difference between the average angles of the first limb polar coordinates and the second limb polar coordinates corresponding to that limb:

[0039] Δθ i =θ i,coach -θ i,user -(θ average,coach -θ average,user ),

[0040]

[0041] Where, θ i,coach -θ i,user , For the angular differences corresponding to each limb, θ average,coach , θ is the average angle of the first limb in polar coordinates. average,user , The average angle of the second limb in polar coordinates.

[0042] As described above, when calculating the angular differences corresponding to each limb, the difference in the average angle between the coach's and user's movements is subtracted, eliminating the influence of the coach and user's different body orientations in the video and the influence of the shooting angle. This overcomes the difficulties in comparing movements and the asynchrony between user and coach movements caused by the different shooting angles of 2D cameras.

[0043] Further, step S3 specifically includes:

[0044] When calculating the cumulative sum of angle differences in the forearm, subtract the corresponding three-dimensional angle difference data of the arm.

[0045] As described above, this reduces the difference in the basic angle of the forearm, which has a large range of motion, and solves the problem that the difference in motion comparison is highly concentrated in the forearm.

[0046] A smart comparison and analysis terminal that provides suggestions for improving fitness movements includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs the following steps:

[0047] S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video.

[0048] S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates.

[0049] S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved.

[0050] S4. Provide intelligent prompts for the limbs whose movements need improvement.

[0051] As can be seen from the above description, the beneficial effects of the present invention are as follows: It provides an intelligent comparison and analysis terminal that prompts improvement of fitness movements. It uses a dynamic time warping algorithm to compare the differences in the three-dimensional angles of the limbs in the coach's movement video and the user's movement video, optimizes the three-dimensional angle difference data of the forearm, and overcomes problems such as different shooting angles of 2D cameras, asynchronous user movements, and abnormal prompts concentrated in the forearm.

[0052] Furthermore, after step S4, the method further includes:

[0053] S5. Construct a scoring function using activation functions to score the accuracy of the user's overall actions relative to the coach's actions;

[0054] The scoring function is: Where tanh is the activation function, Δθ i Due to differences in orientation angles, For distance difference, Time is the video duration, and k is a constant value that needs to be optimized.

[0055] As described above, by comprehensively analyzing a set of intelligent comparison data of fitness movements, users are provided with a more intuitive understanding of the accuracy of their fitness movements, making it easier for them to self-test their exercise results and adjust their fitness plans.

[0056] Further, step S1 specifically includes:

[0057] The first joint coordinates are obtained through a pre-recorded video of the coach's movements;

[0058] The second joint coordinates are obtained from a video recording of the user's actions on-site.

[0059] As described above, joint coordinates are pre-calculated using pre-recorded coaching video, and data such as limb coordinates and three-dimensional angle differences are obtained and saved, reducing redundant calculations.

[0060] Furthermore, the three-dimensional angle difference in step S3 specifically refers to:

[0061] The calculation of the angle difference for each limb used for dynamic time warping alignment includes subtracting the difference between the average angles of the first limb polar coordinates and the second limb polar coordinates corresponding to that limb:

[0062] Δθ i =θ i,coach -θ i,user -(θ average,coach -θ average,user ),

[0063]

[0064] Where, θ i,coach-θ i,user , For the angular differences corresponding to each limb, θ average,coach , θ is the average angle of the first limb in polar coordinates. average,user , The average angle of the second limb in polar coordinates.

[0065] As described above, when calculating the angular differences corresponding to each limb, the difference in the average angle between the coach's and user's movements is subtracted, eliminating the influence of the coach and user's different body orientations in the video and the influence of the shooting angle. This overcomes the difficulties in comparing movements and the asynchrony between user and coach movements caused by the different shooting angles of 2D cameras.

[0066] Further, step S3 specifically includes:

[0067] When calculating the cumulative sum of angle differences in the forearm, subtract the corresponding three-dimensional angle difference data of the arm.

[0068] As described above, this reduces the difference in the basic angle of the forearm, which has a large range of motion, and solves the problem that the difference in motion comparison is highly concentrated in the forearm.

[0069] The intelligent comparison and analysis method and terminal for prompting improvement of fitness movements described above in this invention can overcome problems such as different shooting angles of 2D cameras, asynchronous user movements, and abnormal prompts concentrated on the forearm. The following is a detailed description of the specific implementation:

[0070] Example 1

[0071] A smart comparative analysis method for suggesting improvements in fitness movements includes the following steps:

[0072] S1. Obtain the first joint coordinates and the second joint coordinates using three-dimensional pose estimation technology. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video.

[0073] The first joint coordinates are obtained from a pre-recorded video of the coach's movements; the second joint coordinates are obtained from a video of the user's movements recorded on-site.

[0074] Specifically, the comparative analysis and prompts for fitness movements involve comparing pre-recorded instructor movements with live user movements. Both instructor and user video images utilize 3D pose estimation technology, such as MediaPipe's pose estimation function, to obtain the 3D coordinates of human joints. The 3D polar coordinate angle of the limb corresponding to the line connecting every two joints is calculated, thus obtaining the 3D polar coordinate angle of each limb. This is then categorized into instructor's and user's movements. and users The above-mentioned method of estimating and calculating the three-dimensional polar coordinate angle of each limb from the coach's video images can be pre-calculated and saved, so that it does not need to be calculated repeatedly each time.

[0075] S2. Obtain the polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinates are: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two joint points in the first joint coordinates; the second limb polar coordinates are: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two joint points in the second joint coordinates.

[0076] S3. The dynamic time warping method is used to align the three-dimensional angle differences between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute values ​​of the angle differences corresponding to each limb in all frames are summed to obtain the cumulative sum of the angle differences of that limb.

[0077] Specifically, to handle comparisons when the coach and user's movements are out of sync, a dynamic time wrapping comparison method is used. The three-dimensional polar coordinate angles of each limb of the coach and user are aligned before calculating the angle differences. Furthermore, when calculating the difference in the three-dimensional polar coordinate angles of the coach and user, the difference in the average angles of all limbs of the two individuals is subtracted to eliminate the influence of different orientations of the two individuals and the influence of the shooting angle. The average angles of all limbs of the coach and user are calculated as follows:

[0078] θ average,coach =Σ(θ) 1,coach +θ 2,coach +θ 3,coach +...) / n,

[0079]

[0080] θ average,user =Σ(θ) 1,user +θ 2,user +θ 3,user +...) / n,

[0081]

[0082] To calculate the difference in the three-dimensional polar coordinate angles of each limb between the coach and the user, the difference in the average angles of all limbs of both individuals must be subtracted, as follows:

[0083] Δθ1=θ 1,coach -θ 1,user -(θ average,coach -θ average,user )

[0084] Δθ2=θ 2,coach -θ 2,user -(θ average,coach -θ average,user )

[0085] Δθ3=θ 3,coach -θ 3,user -(θ average,coach -θ average,user ) ...

[0087]

[0088]

[0089]

[0090]

[0091] Based on the above calculations, for each limb, the absolute values ​​of the angle differences across all frames are summed to obtain the cumulative sum of the angle differences for that limb. For example, for limb 1, the cumulative sum across all frames is: Calculate the cumulative sum of angle differences for limb 1; for limb 2, sum the cumulative differences across all frames: The cumulative sum of the angle differences of limb 2 is obtained;

[0092] S4. Provide intelligent prompts for limbs whose movements need improvement, wherein the limbs whose movements need improvement are the limbs with the largest cumulative sum of angle differences.

[0093] The limb with the largest cumulative angle difference is selected as the limb whose movement needs improvement and given intelligent prompts. For example, in the calculation results above, assuming that limb 5 has the largest cumulative angle difference, the user is prompted to improve the movement of limb 5.

[0094] S5. Use activation functions to score the accuracy of the user's overall actions relative to the coach's actions.

[0095] An activation function is used to score the accuracy of the user's overall movements relative to the coach's movements. The scoring function is designed as follows: The design utilizes the principle that the maximum value of the activation function tanh can only infinitely approach 1, where Time is the video duration and k is a constant value that needs to be optimized.

[0096] Example 2

[0097] The difference between this embodiment and Embodiment 1 is that the calculation range of the three-dimensional angle difference data in step S3 is further defined:

[0098] The three-dimensional angle difference is calculated as follows: the angle difference for each limb used for dynamic time warping alignment includes subtracting the difference between the average angles of the first limb polar coordinates and the second limb polar coordinates corresponding to that limb.

[0099] Regarding the three-dimensional angle differences of the forearm, the preferred method is to subtract the three-dimensional angle differences of the arm, which is equivalent to reducing the basic angle differences of the forearm. This overcomes the problem that the large range of motion of the forearm leads to a high probability of the action comparison difference being concentrated in the forearm.

[0100] Example 3

[0101] A smart comparison and analysis terminal that provides suggestions for improving fitness movements includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in Embodiments 1 and 2.

[0102] In summary, the present invention provides an intelligent comparison and analysis method and terminal for improving fitness movements. By using a dynamic time warping algorithm to compare the three-dimensional angle differences of limbs in coach videos and user videos, it overcomes problems such as different shooting angles of 2D cameras, asynchronous user movements, and abnormal prompts concentrated on the forearm.

[0103] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0104] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0105] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A smart comparative analysis method for suggesting improvements in fitness movements, characterized in that, Includes the following steps: S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video. S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates. S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved. The specific differences in three-dimensional angles are as follows: The calculation of the angle difference for each limb used for dynamic time warping alignment includes subtracting the difference between the average angles of the first limb polar coordinates and the second limb polar coordinates corresponding to that limb: ， ， Where, θ i,coach -θ i,user , For the angular differences corresponding to each limb, θ average,coach , θ is the average angle of the first limb in polar coordinates. average,user , The average angle of the second limb in polar coordinates; When calculating the cumulative sum of angle differences in the forearm, subtract the corresponding three-dimensional angle difference data of the arm. S4. Provide intelligent prompts for the limbs whose movements need improvement.

2. The intelligent comparative analysis method for prompting improvement of fitness movements according to claim 1, characterized in that, The process following step S4 also includes: S5. Construct a scoring function using activation functions to score the accuracy of the user's overall actions relative to the coach's actions; The scoring function is: In the formula, tanh is the activation function, and Δθ i Due to differences in orientation angles, For azimuth angle differences, Time is the video duration, and k is a constant value that needs to be optimized.

3. The intelligent comparative analysis method for prompting improvement of fitness movements according to claim 1, characterized in that, Step S1 specifically involves: The first joint coordinates are obtained through a pre-recorded video of the coach's movements; The second joint coordinates are obtained from a video recording of the user's actions on-site.

4. An intelligent comparison and analysis terminal for suggesting improvements in fitness movements, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it performs the following steps: S1. Obtain the first joint coordinates and the second joint coordinates using the three-dimensional pose estimation method. The first joint coordinates are the three-dimensional coordinates of the human joints in the coach's action video, and the second joint coordinates are the three-dimensional coordinates of the human joints in the user's action video. S2. Calculate the corresponding polar coordinates of the first limb and the second limb based on the first joint coordinates and the second joint coordinates; the first limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the first joint coordinates; the second limb polar coordinate is: the three-dimensional polar coordinate angle of a limb corresponding to the line connecting every two adjacent three-dimensional coordinates of the second joint coordinates. S3. The dynamic time warping method is used to align the three-dimensional angle difference between the polar coordinates of the first limb and the polar coordinates of the second limb in time, and the absolute value of the polar coordinate angle difference corresponding to each limb in all frames is summed to obtain the cumulative sum of angle differences for each limb, and the limb with the largest cumulative sum of angle differences is taken as the limb whose action needs to be improved. The specific differences in three-dimensional angles are as follows: The calculation of the angle difference for each limb used for dynamic time warping alignment includes subtracting the difference between the average angles of the first limb polar coordinates and the second limb polar coordinates corresponding to that limb: Dth i =θ i,coach -θ i,user -(θ average,coach -θ average,user ), ， Where, θ i,coach -θ i,user , For the angular differences corresponding to each limb, θ average,coach , θ is the average angle of the first limb in polar coordinates. average,user , The average angle of the second limb in polar coordinates; When calculating the cumulative sum of angle differences in the forearm, subtract the corresponding three-dimensional angle difference data of the arm. S4. Provide intelligent prompts for the limbs whose movements need improvement.

5. The intelligent comparison and analysis terminal for prompting improvement of fitness movements according to claim 4, characterized in that, The process following step S4 also includes: S5. Construct a scoring function using activation functions to score the accuracy of the user's overall actions relative to the coach's actions; The scoring function is: Where tanh is the activation function, and Δθ i Due to differences in orientation angles, For azimuth angle differences, Time is the video duration, and k is a constant value that needs to be optimized.

6. The intelligent comparison and analysis terminal for prompting improvement of fitness movements according to claim 4, characterized in that, Step S1 specifically involves: The first joint coordinates are obtained through a pre-recorded video of the coach's movements; The second joint coordinates are obtained from a video recording of the user's actions on-site.

Images