A method, system, and medium for guiding motion analysis

Abstract

The application provides a guide technique action analysis method, system and medium, the method comprises the following steps: detecting hand key points and limb key points corresponding to each frame of image, calculating and generating a human comprehensive point for each frame of image based on a comprehensive point analysis model; using a category feature analysis model, and combining the bone vector included angle of the student, calculating the corresponding category feature calculation result to determine the technique matching result between the student's practiced technique and the standard technique; analyzing the hand key points and limb key points corresponding to each frame of image to obtain key point analysis results; comparing the bone vector included angle of the student to obtain the bone vector included angle difference; analyzing the comprehensive points in the human comprehensive point frequency heat map and the standard comprehensive point frequency heat map to obtain comprehensive point analysis results; obtaining the guide technique action analysis result through the technique matching result, the key point analysis result, the bone vector included angle difference and the comprehensive point analysis result.

Description

Technical Field

[0001] This invention relates to the field of human motion analysis technology, and more specifically, to a method, system, and medium for analyzing guided movement techniques. Background Technology

[0002] Traditional Chinese medicine (TCM) guiding exercises, based on the theories of meridians and qi and blood, combine breathing techniques, physical movements, and mental regulation to unblock meridians and promote the smooth flow of qi and blood. This has significant implications for disease prevention and the management of chronic diseases. Promoting TCM guiding exercises requires multifaceted efforts and cooperation. This can be achieved through various means, including strengthening publicity and education, establishing professional institutions and teams, enhancing scientific research and academic exchange, and cultivating professional talent. Furthermore, education and training can help build a professional team with specialized knowledge and skills in TCM guiding exercises, providing a talent pool for the promotion of these exercises.

[0003] It should be noted that learning and mastering Traditional Chinese Medicine (TCM) guiding exercises requires certain professional knowledge and skills, which can only be acquired through systematic learning and practice. However, educators often emphasize theory over practice and form over implementation, which negatively impacts the learning and promotion of TCM guiding exercises. To address this issue, educators should focus on combining the theoretical knowledge of TCM guiding exercises with practical application, enabling students to grasp the basic theory while also understanding and practicing the specific methods and techniques of TCM guiding exercises. Furthermore, educators should promptly understand students' learning progress and needs, addressing any problems and shortcomings to improve and refine the teaching, thereby ensuring teaching quality and effectiveness.

[0004] However, due to a lack of professional training methods and equipment, most students learn by following videos or in person with a teacher when practicing guiding exercises. This often leads to strains due to improper practice and makes it difficult for them to judge whether their movements are correct. This situation causes students to lose motivation to practice consistently, resulting in an inability to effectively master and apply traditional Chinese medicine guiding exercises. Consequently, the adoption rate of traditional Chinese medicine guiding exercises at the grassroots level is low.

[0005] In order to solve the above problems, people have been seeking an ideal technological solution. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by providing a method, system, and medium for analyzing the movements of guided exercises. This invention explores a novel learning method based on the holistic view of Traditional Chinese Medicine, integrating motion capture and tracking technology, artificial intelligence technology, and human-computer interaction technology. It incorporates modern science to develop quantitative evaluation standards for guided exercises and a comprehensive human body point analysis method, providing learners with quantitative and visual analysis of the learning process and evaluation results. This enhances the learning experience and effectively reduces the difficulty of promoting and practicing guided exercises, thus having significant implications for guiding the practice of guided exercises.

[0007] To achieve the above objectives, the first aspect of the present invention provides a method for analyzing guided exercises movements, comprising: reading each frame of an image of a student practicing guided exercises movements; analyzing each frame of an image using a pre-configured convolutional neural network model to detect key hand points and limb key points corresponding to each frame of an image; acquiring a pre-configured comprehensive point analysis model; calculating and generating a human comprehensive point for each frame of an image based on the comprehensive point analysis model and the key hand and limb key points corresponding to each frame of an image; calculating the corresponding category feature calculation result using a pre-configured category feature analysis model, combined with the student's corresponding bone vector angle, human comprehensive point feature value, and human key point feature value; determining the movement matching result between the student's practiced movements and standard movements based on the comparison result between the category feature calculation result and the pre-configured standard category feature result; analyzing the key hand and limb key points corresponding to each frame of an image, and the standard key points associated with the movement matching result, to obtain key point analysis results; wherein, the key point analysis results include human body key points. The differences in spatial vector magnitude, velocity, and acceleration corresponding to key points are analyzed. The standard key points refer to pre-configured human body key points of the inheritor. The bone vector angles corresponding to the trainee are compared with the standard bone vector angles associated with the move matching results to obtain bone vector angle differences. The standard bone vector angles refer to the pre-configured bone vector angles of the inheritor. After the trainee completes the guidance practice for each move, all human body integration points are displayed on a pre-selected trainee image to obtain a frequency heatmap of human body integration points corresponding to each move. The frequency heatmap of human body integration points is analyzed in conjunction with the integration points in the pre-configured standard integration point frequency heatmap to obtain integration point analysis results. The standard integration point frequency heatmap represents the distribution location of integration points corresponding to the inheritor. The integration point analysis results include differences in spatial vector magnitude, velocity, and acceleration corresponding to human body integration points. Through the move matching results, the key point analysis results, the bone vector angle differences, and the integration point analysis results, the guidance movement analysis results are obtained.

[0008] To achieve the above objectives, a second aspect of the present invention provides a guided exercise movement analysis system, comprising an omnidirectional motion capture and tracking module, a human key point analysis module, a guided exercise technique classification module, and a comprehensive analysis module. The omnidirectional motion capture and tracking module includes a markerless point motion capture device, a first support mechanism disposed below the markerless point motion capture device, an omnidirectional moving chassis disposed below the first support mechanism, and a motion control unit for controlling the omnidirectional moving chassis. The markerless point motion capture device is used to capture each frame of the student's practice movements, and the motion control unit is used to adjust the orientation and angle of the omnidirectional moving chassis. The markerless motion capture device is always positioned directly in front of the trainee; the human key point analysis module is used to analyze each frame of image using a pre-configured convolutional neural network model to detect the corresponding hand and limb key points; it is also used to acquire a pre-configured comprehensive point analysis model, and based on the comprehensive point analysis model and the corresponding hand and limb key points of each frame of image, calculate and generate a human comprehensive point for each frame of image; the guiding technique classification module is used to use a pre-configured category feature analysis model, combined with the trainee's corresponding bone vector angle, human comprehensive point feature value, and human key point feature value, to calculate the corresponding category feature calculation result; it also uses... Based on the comparison between the calculated results of the category features and the pre-configured standard category feature results, the matching result between the moves practiced by the trainee and the standard moves is determined. The comprehensive analysis module is used to analyze the hand and limb key points corresponding to each frame of the image, and the standard key points associated with the move matching result, to obtain key point analysis results. The key point analysis results include the spatial vector magnitude difference, velocity difference, and acceleration difference corresponding to the human body key points. The standard key points refer to the pre-configured human body key points of the inheritor. The comprehensive analysis module is also used to analyze the angle of the bone vector corresponding to the trainee and the angle of the standard bone vector associated with the move matching result. The difference in bone vector angles is obtained through comparison; the standard bone vector angle refers to the pre-configured bone vector angle of the inheritor; the comprehensive analysis module is also used to display all human body comprehensive points on a pre-selected student image after the student completes the guidance practice of each move, and obtain a frequency heatmap of human body comprehensive points corresponding to each move; the comprehensive points in the human body comprehensive point frequency heatmap are analyzed with the pre-configured standard comprehensive point frequency heatmap to obtain comprehensive point analysis results; the standard comprehensive point frequency heatmap is used to represent the distribution position of the comprehensive points corresponding to the inheritor, and the comprehensive point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body comprehensive points;The comprehensive analysis module is also used to obtain the guiding technique movement analysis results through the move matching results, the key point analysis results, the bone vector angle differences, and the comprehensive point analysis results.

[0009] To achieve the above objectives, a third aspect of the present invention provides a readable storage medium having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the guidance action analysis method described above.

[0010] The beneficial effects of this invention are as follows:

[0011] This invention reads each frame of an image of a student practicing guided exercises, detects the key points of the hands and limbs corresponding to each frame using artificial intelligence technology, and pre-configures a comprehensive point analysis model and a category feature analysis model to obtain comprehensive point analysis results and move matching results. It also analyzes the key points of the student's body and the angle between bone vectors to obtain the key point analysis results and the differences between bone vector angles. Finally, it obtains the guided exercise movement analysis results, which include move matching results, key point analysis results, differences between bone vector angles, and comprehensive point analysis results.

[0012] Therefore, this invention takes the assessment of guided exercises as its core, and formulates professional and universal quantitative assessment standards for guided exercises and a comprehensive human body point analysis method. It supports online and offline analysis, and realizes the quantitative and visual analysis of the learning process and assessment results of guided exercises, helping students to effectively master and apply traditional Chinese medicine guided exercises. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating the guided movement analysis method of the present invention;

[0014] Figure 2 This is a schematic diagram of the guiding movement analysis method of the present invention;

[0015] Figure 3 This is a visual schematic diagram of the key points of the hand, the key points of the limbs, and the comprehensive points of the human body in this invention;

[0016] Figure 4 This is a schematic block diagram of the guiding movement analysis system of the present invention;

[0017] Figure 5 This is a schematic diagram of the guiding movement analysis system of the present invention;

[0018] In the diagram: 1. Display module; 2. Second support mechanism; 3. Motion capture device; 4. First support mechanism; 5. Array-type spectrum sensing radar; 6. Lifting platform shell; 7. Lifting platform inner shell; 8. Omnidirectional moving chassis; 9. Omnidirectional wheels. Detailed Implementation

[0019] The technical solution of the present invention will be further described in detail below through specific embodiments.

[0020] To facilitate understanding, the interactive parties and / or terms and / or custom terms involved in this invention will first be explained in conjunction with the technical solution of this invention:

[0021] Daoyin exercises refer to the various types of movements included in traditional Chinese medicine Daoyin exercises. A single movement may include a series of actions, including but not limited to Baduanjin, Tai Chi, and Wuqinxi. Among them, Wuqinxi includes the first movement, Tiger Play, the second movement, Deer Play, the third movement, Bear Play, the fourth movement, Monkey Play, and the fifth movement, Bird Play.

[0022] Pre-configured convolutional neural network models are used to detect hand and limb keypoints in each frame of the image. For example, MediaPipe Hands is used to detect 21 hand keypoints in each frame, totaling 42 keypoints for both hands; BlazePose is used to detect 33 limb keypoints in each frame. It should be noted that convolutional neural networks for detecting hand keypoints include, but are not limited to, MediaPipe Hands, and convolutional neural network models for detecting limb keypoints include, but are not limited to, BlazePose.

[0023] Each frame of the image corresponds to 21 hand key points: specifically including WRIST (wrist joint), THUMB_CMC (thumb carpometacarpophalangeal joint), THUMB_MCP (thumb metacarpophalangeal joint), THUMB_IP (thumb interphalangeal joint), THUMB_TIP (thumb fingertip), INDEX_FINGER_MCP (index finger carpometacarpophalangeal joint), INDEX_FINGER_PIP (index finger metacarpophalangeal joint), INDEX_FINGER_DIP (index finger interphalangeal joint), INDEX_FINGER_TIP (index fingertip), MIDDLE_FINGER_MCP (middle finger carpometacarpophalangeal joint), MIDDLE_FI NGER_PIP (middle finger metacarpophalangeal joint), MIDDLE_FINGER_DIP (middle finger interphalangeal joint), MIDDLE_FINGER_TIP (middle finger tip), RING_FINGER_MCP (ring finger carpometacarpophalangeal joint), RING_FINGER_PIP (ring finger metacarpophalangeal joint), RING_FINGER_DIP (ring finger interphalangeal joint), RING_FINGER_TIP (ring finger tip), PINKY_MCP (little finger carpometacarpophalangeal joint), PINKY_PIP (little finger metacarpophalangeal joint), PINKY_DIP (little finger interphalangeal joint), and PINKY_TIP (little finger tip).

[0024] Limb key points corresponding to each frame of image: For example, BlazePose detects 33 limb key points corresponding to each frame of image, specifically including 8 hand key points and 25 key points of other parts of the body. This embodiment mainly utilizes the 25 key points of other parts of the body, as shown in the attached figure. Figure 3 As shown.

[0025] Comprehensive Point Analysis Model: A pre-configured analysis model related to the holistic view of Traditional Chinese Medicine; the comprehensive points of the human body are the calculation results of the comprehensive point analysis model. It should be noted that different guiding exercises can target different body parts. This application discovered that the area with the highest density of comprehensive points corresponding to the inheritors of guiding exercises in the standard comprehensive point frequency heatmap corresponds to the body parts exercised by the corresponding movements. Therefore, this application uses the comprehensive point analysis model to analyze the key points of the trainees' hands and limbs to obtain comprehensive point information of the human body. Then, after a trainee completes the practice of a certain guiding exercise movement, the comprehensive point frequency heatmap of the human body can be used to reflect whether the trainee's movements are standard (or whether the practice of a certain guiding exercise movement meets the expected training effect of that movement).

[0026] Category Feature Analysis Model: Another pre-configured analysis model used to identify the type of moves practiced by trainees, such as the first move of the Five Animal Frolics, Tiger Frolics, or the second move of the Five Animal Frolics, Deer Frolics, etc.

[0027] Bone vector angles: This application mainly uses 12 bone vector angles, as shown in the attached figure. Figure 3 As shown; the angles of these 12 bone vectors are calculated from the key points of the hand and limbs corresponding to each frame of the image. The specific calculation formula is based on existing technology and will not be elaborated here.

[0028] Human body composite point feature values: mainly include the spatial vector magnitude, velocity, acceleration, etc. corresponding to the human body composite points. (X,Y,Z) represents the three-dimensional coordinates of the human body composite point corresponding to a certain frame of image; the velocity of the composite point = the difference in the spatial vector magnitude of the composite point between image frames / the time difference between frames, and the acceleration of the composite point = the velocity difference of the human body composite point between image frames / the time difference between frames.

[0029] Human body key point feature values: mainly include the spatial vector magnitude, velocity, acceleration, etc. corresponding to human body key points. (x,y,z) represents the three-dimensional coordinates of the human body key points corresponding to a certain frame of the image; key point velocity = difference in spatial vector magnitude of key points between image frames / time difference between frames, key point acceleration = difference in velocity of human body key points between image frames / time difference between frames.

[0030] Key point analysis results include differences in spatial vector magnitude, velocity, and acceleration corresponding to key points on the human body. The difference in spatial vector magnitude corresponding to key points on the human body is the difference in spatial vector magnitude at the same key point in each frame of the image when the student and the inheritor are practicing the same move. The difference in velocity corresponding to key points on the human body is the difference in velocity at the same key point in each frame of the image when the student and the inheritor are practicing the same move. The difference in acceleration corresponding to key points on the human body is the difference in acceleration at the same key point in each frame of the image when the student and the inheritor are practicing the same move.

[0031] Bone vector angle difference: refers to the difference in bone vector angle at the same bone vector angle in each frame of the image when the student and the inheritor are practicing the same move.

[0032] The results of the comprehensive point analysis include differences in the spatial vector magnitude, velocity, and acceleration corresponding to the comprehensive points on the human body. The difference in the spatial vector magnitude corresponding to the comprehensive points on the human body equals the difference in the spatial vector magnitude of the comprehensive points in each frame of the image when the student and the inheritor practice the same move. This can be represented by the magnitude of the difference in the three-dimensional coordinates between the comprehensive points, such as... (X, Y, Z) represents the three-dimensional coordinates of the trainee's body's integrated points, and (X0, Y0, Z0) represents the three-dimensional coordinates of the inheritor's body's integrated points. After classifying the techniques, if the magnitude of the difference between the three-dimensional coordinates of the integrated points is less than the preset value, it indicates that the trainee's practiced movements are relatively standard and not much different from the inheritor's movements, thus achieving the training objective. If the magnitude of the difference between the three-dimensional coordinates of the integrated points is greater than or equal to the preset value, it indicates that the trainee's practiced movements are not standard and differ significantly from the inheritor's movements, thus failing to achieve the training objective. The preset value can be set according to accuracy requirements and is not limited here.

[0033] The speed difference corresponding to the human body's comprehensive point = the difference in speed at the comprehensive point between the student and the inheritor when practicing the same move. The acceleration difference corresponding to the human body's comprehensive point = the difference in acceleration at the comprehensive point between the student and the inheritor when practicing the same move. The specific calculation formula will not be elaborated here.

[0034] Example 1

[0035] This embodiment provides a specific implementation method for analyzing guided movement techniques, as shown in the attached figure. Figure 1 As shown; the guided movement analysis method includes:

[0036] The system reads each frame of the student's practice of the guiding movements, analyzes each frame using a pre-configured convolutional neural network model, and detects the corresponding key points of the hands and limbs in each frame.

[0037] Obtain a pre-configured comprehensive point analysis model, and based on the comprehensive point analysis model and the hand key points and limb key points corresponding to each frame image, calculate and generate a human body comprehensive point for each frame image;

[0038] A pre-configured category feature analysis model is used, combined with the student's corresponding bone vector angle, human body comprehensive point feature value and human body key point feature value, to calculate the corresponding category feature calculation result; based on the comparison result between the category feature calculation result and the pre-configured standard category feature result, the move matching result between the move practiced by the student and the standard move is determined.

[0039] For each frame of image, the key points of the hands and limbs are analyzed, and the standard key points associated with the matching results of the moves are analyzed to obtain key point analysis results. The key point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body key points. The standard key points refer to the pre-configured human body key points of the inheritor.

[0040] The bone vector angle corresponding to the trainee is compared with the standard bone vector angle associated with the matching result of the move to obtain the difference in bone vector angle; wherein, the standard bone vector angle refers to the bone vector angle of the pre-configured successor.

[0041] After the trainees complete the guided practice of each move, all human body integration points are displayed on a pre-selected trainee image to obtain a frequency heatmap of human body integration points corresponding to each move. The frequency heatmap of human body integration points is then analyzed with the integration points in a pre-configured standard integration point frequency heatmap to obtain integration point analysis results. The standard integration point frequency heatmap is used to represent the distribution location of integration points corresponding to the inheritor. The integration point analysis results include differences in spatial vector magnitude, velocity, and acceleration corresponding to human body integration points.

[0042] The analysis results of the guiding technique are obtained by combining the move matching results, the key point analysis results, the bone vector angle difference, and the comprehensive point analysis results.

[0043] It should be noted that the key points of the inheritor's hands and limbs are pre-defined as standard key points, and the bone vector angles of the inheritor are defined as standard bone vector angles, and these are stored in association with the corresponding move categories. After determining the move category represented by the move matching result, the key points of the hands, limbs, and bone vector angles corresponding to each frame of the image when the inheritor performs the same move are read according to the move category, preparing for subsequent key point analysis and bone vector angle comparison.

[0044] It should also be noted that the results of the key point analysis, the difference in the bone vector angle, and the results of the comprehensive point analysis can be calculated separately based on the corresponding move practice data of the trainees and the inheritors. The specific calculation formulas will not be elaborated here.

[0045] It should also be noted that both the human body comprehensive point frequency heatmap and the standard comprehensive point frequency heatmap are used to display the human body comprehensive points of each frame of an image corresponding to a certain move; the difference is that the human body comprehensive point frequency heatmap corresponds to the human body comprehensive points corresponding to the movements performed by the student, while the standard comprehensive point frequency heatmap corresponds to the human body comprehensive points corresponding to the movements performed by the inheritor of the guiding technique.

[0046] It should also be noted that, based on the holistic view of traditional Chinese medicine, the comprehensive points of the human body in each frame of the image are displayed in the same image. This image can be a human body image of a pre-selected student / inheritor of the guiding technique in a resting state.

[0047] In some embodiments, the pre-configured comprehensive point analysis model uses the following formula:

[0048]

[0049] Where (X,Y,Z) represents the three-dimensional coordinates of the human body composite point, n p ω represents the total number of key points in the hand and limbs. i This represents the pre-set weights corresponding to hand or limb keypoints, (x i ,y i, z i () represents the three-dimensional coordinates corresponding to key points of the hand or limb.

[0050] It can be understood that the total number of hand and limb key points n used in the calculation process for the human body composite points corresponding to each frame image is... p It can be 67, or modified accordingly based on actual needs.

[0051] Specifically, this embodiment assigns different weights to each key point, with pre-set weights ω corresponding to hand or limb key points. i The value range is [0,1].

[0052] In some embodiments, the pre-configured category feature analysis model uses the following formula:

[0053]

[0054] Where P represents the calculated category feature of the corresponding move, ω angle This represents the pre-set weight of the bone vector angle detection, n. aθ represents the total number of pre-set bone vector angles. a This represents the angle between the bone vectors corresponding to the student. This represents the pre-set weights of the bone vector angles;

[0055] ω composite This represents the pre-set weight of the human body comprehensive point detection, n. c β represents the total number of pre-defined human body feature categories. c λ represents the comprehensive feature value of the human body corresponding to the student. c This represents the pre-set weights of the human body's comprehensive feature values;

[0056] ω single n represents the pre-set weight of human keypoint detection. s n represents the total number of pre-set human key point feature data. j This represents the total number of pre-set human body key points, γ s,j μ represents the key feature value of the human body corresponding to the trainee. s,j This represents the pre-set weights of key human body feature values.

[0057] It should be noted that this embodiment uses a pre-configured category feature analysis model to classify moves, and combines the student's corresponding bone vector angle, human body comprehensive point feature value and human body key point feature value to comprehensively analyze and obtain the category feature calculation result. The category feature calculation result is compared with the standard category feature result to determine what specific move the student is practicing, thereby quickly and accurately classifying moves, and then using move classification to assist in the analysis of the student's practiced moves.

[0058] It should also be noted that the pre-configured standard category feature results include different category feature results corresponding to different moves within the same guiding technique category; for example, Baduanjin includes eight moves, and the standard category feature results corresponding to Baduanjin include the first move category feature result P1, the second move category feature result P2, ..., the eighth move category feature result P8. The value ranges between P1 and P8 do not overlap. If the difference between the category feature calculation result and the second move category feature result P2 is within a preset range, then the move matching result is the second move of Baduanjin.

[0059] It should also be noted that for the acquired M-frame images, the angle value θ of each bone vector is... a This can be the variance or average of the angles between a series of bone vectors corresponding to M frames of images, or the human body composite point feature value β. c It can be the variance or average of a series of comprehensive point feature values ​​(spatial vector magnitude, velocity, acceleration) corresponding to M frames of images, and the human key point feature value γ. s,jIt can be represented as the variance or average of a series of key point feature values ​​(spatial vector magnitude, velocity, acceleration) corresponding to M frames of images.

[0060] It is understandable that the pre-set weight ω for bone vector angle detection... angle Bone vector angle weights The weight ω of human body comprehensive point detection composite Human body comprehensive point feature value weight λ c The weight ω of human key point detection single Human body key point feature value weight μ s,j The value range is [0,1], and it can be dynamically adjusted according to actual needs;

[0061] For example, if trainees are more concerned about the overall bone vector angle, the weight ω of the pre-set bone vector angle detection can be increased. angle Weight of the angle between the bone vector and the bone vector The selected guidance exercise category contains more lower limb movements, which can also increase the weight of the bone vector angle related to the lower limbs; the selected guidance exercise category contains more upper limb movements, which can also increase the weight of the bone vector angle related to the upper limbs.

[0062] For example, if trainees focus more on the overall human body contours, the weight given to human body contour detection can be increased. composite Weight λ of human body integrated point features c ;

[0063] For example, if trainees focus more on each fine movement, the weight given to human key point detection can be increased. single and human key point feature weights μ s,j .

[0064] Specifically, the pre-set total number of bone vector angles n a The maximum number of human body feature categories can be 12, where n is the pre-set total number of human body feature categories. c The maximum number of human body key points can be 3, and the preset total number of key points n is 3. j It can be 67 (including 42 hand key points and 25 limb key points); n a n c n s and n j The specific values ​​can also be dynamically adjusted according to actual needs.

[0065] In one specific implementation, the total number of bone vector angles n of the acquired 10 frames of images a Take 12, the total number of human body comprehensive feature categories n c The total number of human body key point feature data n s All are taken as 3, and the total number of key points on the human body is n.j Take 67;

[0066] Correspondingly, the bone vector angle value θ1 is the variance of the bone vector angle ∠1 at the same location corresponding to 10 frames of images, and so on, θ 12 The variance of the angle ∠12 between bone vectors at the same location corresponding to 10 frames of images;

[0067] Correspondingly, the human body composite point feature value β1 is the variance of the spatial vector magnitude of the composite point corresponding to 10 frames of images, the human body composite point feature value β2 is the variance of the velocity of the composite point corresponding to 10 frames of images, and the human body composite point feature value β3 is the variance of the acceleration of the composite point corresponding to 10 frames of images.

[0068] Correspondingly, the human body key point feature value γ 1,1 The variance of the spatial vector magnitudes of the first keypoint is given by γ, and so on, representing the eigenvalues ​​of human keypoints. 1,67 The variance of the magnitudes of the 10 spatial vectors of the 67th keypoint; human keypoint eigenvalue γ 2,1 The variance of the nine velocity values ​​at the first keypoint, and so on, represents the characteristic value γ of the human keypoint. 2,67 The variance of the nine velocity values ​​at the 67th keypoint; human keypoint eigenvalue γ 3,1 The variance of the eight acceleration values ​​at the first keypoint, and so on, represents the characteristic value γ of the human keypoint. 3,67 The variance of the eight acceleration values ​​at the 67th key point.

[0069] In some embodiments, the guided movement analysis method further includes:

[0070] In each frame of the image, key points of the hand, key points of the limbs, and comprehensive points of the human body are drawn to visualize human motion information.

[0071] And / or, based on the human body composite points corresponding to each frame of image and the acquisition time, draw the trajectory of the human body composite points;

[0072] And / or, display the split-screen comparison information of the actions of the trainee and the inheritor, wherein the split-screen comparison information of the actions includes the difference in the angle between bone vectors, the difference in the spatial vector magnitude of each key point of the human body and the comprehensive point of the human body, the difference in the velocity of each key point of the human body and the comprehensive point of the human body, and the difference in the acceleration of each key point of the human body and the comprehensive point of the human body.

[0073] Example 2

[0074] Based on Example 1, this example provides a specific implementation method for another guided movement analysis method.

[0075] The difference between this embodiment and Embodiment 1 is that the guided movement analysis method further includes updating the weights corresponding to the key hand points or key limb points in the comprehensive point analysis model based on the calculation results of the category features.

[0076] It should be noted that, in Example 1, the weights ω corresponding to the hand key points or limb key points used in the comprehensive point analysis model are... i The initial default value;

[0077] To improve the accuracy of the calculation results of the three-dimensional coordinates of the human body's composite points, different keypoint weights are pre-assigned for different moves, and the keypoint weights ω corresponding to different moves are assigned... i The data is stored in association with the move category; in this embodiment, after calculating the category features, the weights ω corresponding to the hand or limb key points used in the comprehensive point analysis model are assigned. i Update the weight ω i Updated to a set of weights associated with the move represented by the category feature calculation results.

[0078] In other embodiments, these weights can also be dynamically adjusted based on the move classification results. For example, when there are many hand movements in the moves corresponding to the detected move type, the weight corresponding to the hand key points can be increased; when there are many limb movements in the moves corresponding to the detected move type, the weight corresponding to the limb key points can also be increased.

[0079] In one specific implementation, when a student practices a certain move for the first time, the student selects a guiding technique. However, this selection does not involve choosing a specific move, but rather a general category of guiding techniques; for example, selecting the Eight Pieces of Brocade, but not specifying which particular movement within the Eight Pieces of Brocade. Therefore, in Example 1, the weight ω corresponding to the key hand points or limb key points used in the comprehensive point analysis model... i The initial default value is a set of weights corresponding to a general category, such as the initial default weight value for Baduanjin. After obtaining the category feature calculation results, the weights are updated according to the movement category detected by the category feature calculation results. For example, after detecting that the student is practicing the third movement of Baduanjin based on the previous few frames of images, the weights are updated to a set of weights associated with the third movement of Baduanjin, thereby improving the detection accuracy of subsequent comprehensive point analysis, etc.

[0080] Example 3

[0081] Based on the above embodiments, this embodiment provides a specific implementation method for another guided movement analysis method, as shown in the appendix. Figure 2 As shown.

[0082] In some embodiments, the method for analyzing guided exercises further includes: before a student practices a certain guided exercise move, selecting a breathing mode through a breathing mode selection interface, wherein the breathing mode includes natural breathing, abdominal breathing, pelvic floor muscle contraction breathing, and breath-holding.

[0083] It should be noted that natural breathing refers to breathing without changing the student's normal breathing pattern, but rather breathing naturally; abdominal breathing refers to breathing through the movement of the diaphragm, including normal abdominal breathing and reverse abdominal breathing; pelvic floor muscle contraction breathing refers to contracting the muscles of the anus and perineum during inhalation and relaxing the muscles of the anus and perineum during exhalation; and breath-holding refers to pausing briefly between or after inhalation and exhalation before resuming inhalation or exhalation.

[0084] In some embodiments, after selecting a breathing mode, the method further includes:

[0085] Using an array-type spectrum sensing radar and a radar information processing server, combined with signal processing and spectrum sensing technology, the micro-motion information of the trainee's body surface in a resting state is detected in a surround manner. Based on the micro-motion information of the body surface, it is detected whether the trainee's current breathing mode is a pre-selected breathing mode, so as to perform non-contact detection of breathing mode.

[0086] When the student's current breathing method matches the pre-selected breathing pattern, a guide technique selection interface pops up, allowing the student to select the specific technique.

[0087] When the student's current breathing method is inconsistent with the pre-selected breathing mode, the voice prompt will prompt the student to adjust their breathing method and perform a breathing method test again.

[0088] It should be noted that after the student selects a breathing mode, this embodiment uses an array-type spectrum sensing radar 5 to detect the micro-movement information of the student's body surface in a surrounding (up, down, left, right, front, and back) manner, and performs non-contact detection of breathing mode and breathing state.

[0089] It should also be noted that the radar information processing server receives the micro-motion information of the body surface acquired by the array-type spectrum sensing radar 5, calculates and feeds back indicators such as the smoothness of the student's breathing, the accuracy of the breathing method, and the breathing frequency in the resting state, and generates resting state detection results and a judgment on whether the student can start practicing the guiding technique.

[0090] It should be noted that by detecting the key hand points corresponding to each frame of the image, the system can also identify the student's gesture information based on the key hand points. By comparing this information with the pre-stored gesture command information, the system can perform non-contact operation of various functions, including pausing, starting, ending, and changing the type of guidance exercise in real time.

[0091] In some embodiments, when selecting a breathing mode and specific techniques of the guiding exercise, the following is performed:

[0092] The system collects the trainee's hand gestures and compares them with the hand gesture command information corresponding to the pre-configured breathing mode. If the hand gesture matches the hand gesture command information corresponding to a certain breathing mode, the corresponding breathing mode is selected as the target breathing mode.

[0093] The system collects the trainees' hand gestures and compares them with the pre-configured gesture command information corresponding to a certain type of guided exercise. If the hand gesture matches the gesture command information corresponding to a certain type of guided exercise, the corresponding guided exercise type is selected for practice.

[0094] Example 4

[0095] Based on the above embodiments, this embodiment provides a specific implementation of a guided movement analysis system, as shown in the appendix. Figure 4 and attached Figure 5 As shown;

[0096] The guided movement analysis system includes a comprehensive motion capture and tracking module, a human key point analysis module, a guided movement classification module, and a comprehensive analysis module.

[0097] The omnidirectional motion capture and tracking module includes a markerless point motion capture device 3, a first support mechanism 4 (for supporting the motion capture device) disposed below the markerless point motion capture device 3, an omnidirectional moving chassis 8 disposed below the first support mechanism 4, and a motion control unit for controlling the omnidirectional moving chassis 8; wherein, the markerless point motion capture device 3 is used to capture each frame of the student's practice movements, and the motion control unit is used to adjust the orientation and angle of the omnidirectional moving chassis 8 so that the markerless point motion capture device 3 is always directly in front of the student;

[0098] The human body key point analysis module is used to analyze each frame of image through a pre-configured convolutional neural network model to detect the hand key points and limb key points corresponding to each frame of image; it is also used to obtain a pre-configured comprehensive point analysis model, and calculate and generate a human body comprehensive point for each frame of image based on the comprehensive point analysis model and the hand key points and limb key points corresponding to each frame of image.

[0099] The guided movement classification module is used to calculate the corresponding category feature calculation results by using a pre-configured category feature analysis model and combining the student's corresponding bone vector angle, human body comprehensive point feature value and human body key point feature value; it is also used to determine the movement matching result between the student's practiced movement and the standard movement based on the comparison result between the category feature calculation result and the pre-configured standard category feature result.

[0100] The comprehensive analysis module is used to analyze the key points of the hands and limbs corresponding to each frame of the image, as well as the standard key points associated with the move matching results, to obtain key point analysis results. The key point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body key points. The standard key points refer to the pre-configured human body key points of the inheritor.

[0101] The comprehensive analysis module is also used to compare the bone vector angle corresponding to the trainee with the standard bone vector angle associated with the move matching result to obtain the bone vector angle difference; wherein, the standard bone vector angle refers to the pre-configured bone vector angle of the inheritor;

[0102] The comprehensive analysis module is also used to display all human body integration points on a pre-selected student image after the student completes the guided practice of each move, thereby obtaining a frequency heatmap of human body integration points corresponding to each move; and to analyze the integration points in the human body integration point frequency heatmap and the pre-configured standard integration point frequency heatmap to obtain integration point analysis results; wherein, the standard integration point frequency heatmap is used to represent the distribution position of the integration points corresponding to the inheritor, and the integration point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body integration points;

[0103] The comprehensive analysis module is also used to obtain the analysis results of the guiding technique movements through the matching results of the moves, the analysis results of the key points, the differences in the bone vector angles, and the comprehensive point analysis results.

[0104] It should be noted that the markerless motion capture device 3 can also acquire the orientation and spatial coordinates of the student's limbs and construct the spatial coordinate relationship between the student and the module; the motion control unit uses spatial positioning technology to drive the omnidirectional wheels 9 and automatically adjust the orientation and angle of the omnidirectional moving chassis 8 so that the markerless motion capture device 3 is always in front of the student, which facilitates more accurate capture of the student's limb movement information.

[0105] In some embodiments, the guided movement analysis system further includes a display module 1, which is mounted on the lifting platform housing 6 via a second support mechanism 2. The lifting platform housing 6 is mounted on an omnidirectional moving chassis 8 via an inner lifting platform housing 7. The display module 1 is used for:

[0106] In each frame of the image, key points of the hand, key points of the limbs, and comprehensive points of the human body are drawn to visualize human motion information.

[0107] And / or, based on the human body composite points corresponding to each frame of image and the acquisition time, draw the trajectory of the human body composite points;

[0108] And / or, display the split-screen comparison information of the actions of the trainee and the inheritor, wherein the split-screen comparison information of the actions includes the difference in the angle between bone vectors, the difference in the spatial vector magnitude of each key point of the human body and the comprehensive point of the human body, the difference in the velocity of each key point of the human body and the comprehensive point of the human body, and the difference in the acceleration of each key point of the human body and the comprehensive point of the human body.

[0109] It should be noted that the display module 1 is also used to display the frequency domain differences in sequence data information of key points and comprehensive points of the human body, as well as to display the evaluation report of guided exercises, etc.

[0110] The frequency domain information of the sequence data can be obtained from the time domain information through digital signal processing techniques, such as discrete Fourier transform, wavelet transform, and power spectral density estimation. This allows for multi-dimensional and accurate evaluation and intelligent error correction of the trainee's movements. Combined with correlation analysis, the degree of fit between the trainee's and the inheritor's movements is obtained, resulting in an evaluation result categorized into four levels: excellent, good, average, and poor. The guided exercise movement evaluation report includes the movement matching results, key point analysis results, bone vector angle differences, and comprehensive point analysis results, and is displayed separately after the trainee completes the movements and chooses to exit the practice session.

[0111] In some embodiments, the guided movement analysis system further includes a human-computer interaction module, which is used for:

[0112] The system collects the trainee's hand gestures and compares them with the hand gesture command information corresponding to the pre-configured breathing modes. If the hand gesture matches the hand gesture command information corresponding to a certain breathing mode, the corresponding breathing mode is selected as the target breathing mode. The breathing modes include natural breathing, abdominal breathing, pelvic floor breathing, and breath-holding.

[0113] The system collects the trainees' hand gestures and compares them with the pre-configured gesture command information corresponding to a certain type of guided exercise. If the hand gesture matches the gesture command information corresponding to a certain type of guided exercise, the corresponding guided exercise type is selected for practice.

[0114] In some embodiments, the guided movement analysis system further includes a voice prompt module, which is used to: prompt the student to adjust their breathing mode when the student's current breathing mode is inconsistent with the pre-selected breathing mode.

[0115] In one specific implementation, after all modules of the system are initialized and started, the system waits for the student to make a system start gesture to enter the breathing mode selection system. If the system receives the start command, a breathing mode selection interface pops up, allowing the student to select a mode using gestures, including four modes: natural breathing, abdominal breathing, pelvic floor muscle breathing, and breath-holding.

[0116] After the student has made their selection, the system will start to detect the student's resting state, determine whether the student's breathing method and breathing state are correct, generate the test results, and provide prompts in conjunction with voice broadcast. If it does not meet the requirements, the student needs to adjust or change to a smoother breathing method in time.

[0117] If the student's resting state meets the system requirements, a selection interface for the type of guided exercise will pop up for the student to choose from. After selection, the student must follow the system's voice prompts to begin practicing. At the same time, the omnidirectional motion capture and tracking module will automatically capture the student's limb movement information and continuously adjust the angle and orientation of the omnidirectional moving chassis 8 to ensure that the system is always directly in front of the student's limbs.

[0118] During the student's practice, the system displays the student's limb and hand movements and their simulated images (key points of the hands, key points of the limbs, and comprehensive points of the human body, etc.), and the guided movement analysis images (screen comparison information of the student's and the guided movement inheritor's corresponding movements). It can also control the system to pause, start, end, and change guided movements in real time based on the student's gestures.

[0119] Meanwhile, the guided exercises classification module will identify and classify the names of the exercises practiced by the student in real time, and can also adopt different analysis schemes according to different exercises. If the student makes a finishing gesture or has completed the guided exercises, the system will automatically display the human body comprehensive point trajectory and frequency heat map of each exercise practiced by the student, generate and display the exercise evaluation report, and then terminate the system operation.

[0120] Example 5

[0121] Based on the above embodiments, this embodiment provides a readable storage medium storing instructions that, when executed by one or more processors, cause the processors to perform the guided movement analysis method as described in Embodiment 1, 2, or 3.

[0122] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer non-transitory readable storage media (including, but not limited to, disk storage, v3D-ROM, optical storage, etc.) containing computer program code.

[0123] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0125] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. A method for analyzing the movements of guided exercises, characterized in that, include: The system reads each frame of the student's practice of the guiding movements, analyzes each frame using a pre-configured convolutional neural network model, and detects the corresponding key points of the hands and limbs in each frame. Obtain a pre-configured comprehensive point analysis model, and based on the comprehensive point analysis model and the hand key points and limb key points corresponding to each frame image, calculate and generate a human body comprehensive point for each frame image; The pre-configured comprehensive point analysis model uses the following formula: Where (X,Y,Z) represents the three-dimensional coordinates of the human body composite point, n p ω represents the total number of key points in the hand and limbs. i This represents the pre-set weights corresponding to hand or limb keypoints, (x i , y i, z i () represents the three-dimensional coordinates corresponding to key points of the hand or limb; A pre-configured category feature analysis model is used, combined with the student's corresponding bone vector angle, human body comprehensive point feature value and human body key point feature value, to calculate the corresponding category feature calculation result; based on the comparison result between the category feature calculation result and the pre-configured standard category feature result, the move matching result between the move practiced by the student and the standard move is determined. The pre-configured category feature analysis model uses the following formula: Where P represents the calculated category feature of the corresponding move, ω angle This represents the pre-set weight of the bone vector angle detection, n. a θ represents the total number of pre-set bone vector angles. a φ represents the angle between the bone vectors corresponding to the student. a This represents the pre-set weights of the bone vector angles; ω composite n represents the pre-set weight of the human body comprehensive point detection. c β represents the total number of pre-defined human body feature categories. c λ represents the human body composite feature value corresponding to the student. c This represents the pre-set weights of the human body's comprehensive feature values; ω single n represents the pre-set weight of human keypoint detection. s n represents the total number of pre-set human body key point feature data. j This represents the total number of pre-set human body key points, γ s,j μ represents the key feature value of the human body corresponding to the trainee. s,j This represents the pre-set weights of key human body feature values; For each frame of image, the key points of the hands and limbs are analyzed, and the standard key points associated with the matching results of the moves are analyzed to obtain key point analysis results. The key point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body key points. The standard key points refer to the pre-configured human body key points of the inheritor. The bone vector angle corresponding to the trainee is compared with the standard bone vector angle associated with the matching result of the move to obtain the difference in bone vector angle; wherein, the standard bone vector angle refers to the bone vector angle of the pre-configured successor. After the trainees complete the guided practice of each move, all human body integration points are displayed on a pre-selected trainee image to obtain a frequency heatmap of human body integration points corresponding to each move. The frequency heatmap of human body integration points is then analyzed with the integration points in a pre-configured standard integration point frequency heatmap to obtain integration point analysis results. The standard integration point frequency heatmap is used to represent the distribution location of integration points corresponding to the inheritor. The integration point analysis results include differences in spatial vector magnitude, velocity, and acceleration corresponding to human body integration points. The analysis results of the guiding technique are obtained by combining the move matching results, the key point analysis results, the bone vector angle difference, and the comprehensive point analysis results.

2. The method for analyzing guided movements according to claim 1, characterized in that, Also includes: In each frame of the image, key points of the hand, key points of the limbs, and comprehensive points of the human body are drawn to visualize human motion information. And / or, based on the human body composite points corresponding to each frame of image and the acquisition time, draw the trajectory of the human body composite points; And / or, display the split-screen comparison information of the actions of the trainee and the inheritor, wherein the split-screen comparison information of the actions includes the difference in the angle between bone vectors, the difference in the spatial vector magnitude of each key point of the human body and the comprehensive point of the human body, the difference in the velocity of each key point of the human body and the comprehensive point of the human body, and the difference in the acceleration of each key point of the human body and the comprehensive point of the human body.

3. The method for analyzing guided movements according to claim 1, characterized in that, Also includes: Based on the calculation results of the category features, the weights corresponding to the key hand points or key limb points in the comprehensive point analysis model are updated, and the updated weights are associated with the moves represented by the calculation results of the category features.

4. The method for analyzing guided movements according to claim 1, characterized in that, Also includes: Before students practice a certain guiding technique, they can select a breathing mode through the breathing mode selection interface. The breathing modes include natural breathing, abdominal breathing, Kegel breathing, and breath-holding.

5. The method for analyzing guided movements according to claim 4, characterized in that, After selecting a breathing mode, the following are also included: Using an array-type spectrum sensing radar and a radar information processing server, combined with signal processing and spectrum sensing technology, the micro-motion information of the trainee's body surface in a resting state is detected in a surround manner. Based on the micro-motion information of the body surface, it is detected whether the trainee's current breathing mode is a pre-selected breathing mode, so as to perform non-contact detection of breathing mode. When the student's current breathing method matches the pre-selected breathing pattern, a guide technique selection interface pops up, allowing the student to select the specific technique. When the student's current breathing method is inconsistent with the pre-selected breathing mode, the voice prompt will prompt the student to adjust their breathing method and perform a breathing method test again.

6. The method for analyzing guided movements according to claim 5, characterized in that, When selecting a breathing mode and specific techniques for the guiding exercise, execute: The system collects the trainee's hand gestures and compares them with the hand gesture command information corresponding to the pre-configured breathing mode. If the hand gesture matches the hand gesture command information corresponding to a certain breathing mode, the corresponding breathing mode is selected as the target breathing mode. The system collects the trainees' hand gestures and compares them with the pre-configured gesture command information corresponding to a certain type of guided exercise. If the hand gesture matches the gesture command information corresponding to a certain type of guided exercise, the corresponding guided exercise type is selected for practice.

7. A guided exercise movement analysis system, characterized in that: It includes a comprehensive motion capture and tracking module, a human key point analysis module, a guiding technique classification module, and a comprehensive analysis module. The omnidirectional motion capture and tracking module includes a markerless point motion capture device, a first support mechanism disposed below the markerless point motion capture device, an omnidirectional moving chassis disposed below the first support mechanism, and a motion control unit for controlling the omnidirectional moving chassis; wherein, the markerless point motion capture device is used to capture each frame of the student's practice movements, and the motion control unit is used to adjust the orientation and angle of the omnidirectional moving chassis so that the markerless point motion capture device is always directly in front of the student; The human body key point analysis module is used to analyze each frame of image through a pre-configured convolutional neural network model to detect the hand key points and limb key points corresponding to each frame of image; it is also used to obtain a pre-configured comprehensive point analysis model, and calculate and generate a human body comprehensive point for each frame of image based on the comprehensive point analysis model and the hand key points and limb key points corresponding to each frame of image. The pre-configured comprehensive point analysis model uses the following formula: Where (X,Y,Z) represents the three-dimensional coordinates of the human body composite point, n p ω represents the total number of key points in the hand and limbs. i This represents the pre-set weights corresponding to hand or limb keypoints, (x i , y i, z i () represents the three-dimensional coordinates corresponding to key points of the hand or limb; The guided movement classification module is used to calculate the corresponding category feature calculation results by using a pre-configured category feature analysis model and combining the student's corresponding bone vector angle, human body comprehensive point feature value and human body key point feature value; it is also used to determine the movement matching result between the student's practiced movement and the standard movement based on the comparison result between the category feature calculation result and the pre-configured standard category feature result. The pre-configured category feature analysis model uses the following formula: Where P represents the calculated category feature of the corresponding move, ω angle This represents the pre-set weight of the bone vector angle detection, n. a θ represents the total number of pre-set bone vector angles. a φ represents the angle between the bone vectors corresponding to the student. a This represents the pre-set weights of the bone vector angles; ω composite n represents the pre-set weight of the human body comprehensive point detection. c β represents the total number of pre-defined human body feature categories. c λ represents the human body composite feature value corresponding to the student. c This represents the pre-set weights of the human body's comprehensive feature values; ω single n represents the pre-set weight of human keypoint detection. s n represents the total number of pre-set human body key point feature data. j This represents the total number of pre-set human body key points, γ s,j μ represents the key feature value of the human body corresponding to the trainee. s,j This represents the pre-set weights of key human body feature values; The comprehensive analysis module is used to analyze the key points of the hands and limbs corresponding to each frame of the image, as well as the standard key points associated with the move matching results, to obtain key point analysis results. The key point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body key points. The standard key points refer to the pre-configured human body key points of the inheritor. The comprehensive analysis module is also used to compare the bone vector angle corresponding to the trainee with the standard bone vector angle associated with the move matching result to obtain the bone vector angle difference; wherein, the standard bone vector angle refers to the pre-configured bone vector angle of the inheritor; The comprehensive analysis module is also used to display all human body integration points on a pre-selected student image after the student completes the guided practice of each move, thereby obtaining a frequency heatmap of human body integration points corresponding to each move; and to analyze the integration points in the human body integration point frequency heatmap and the pre-configured standard integration point frequency heatmap to obtain integration point analysis results; wherein, the standard integration point frequency heatmap is used to represent the distribution position of the integration points corresponding to the inheritor, and the integration point analysis results include the differences in spatial vector magnitude, velocity, and acceleration corresponding to the human body integration points; The comprehensive analysis module is also used to obtain the analysis results of the guiding technique movements through the matching results of the moves, the analysis results of the key points, the differences in the bone vector angles, and the comprehensive point analysis results.

8. A readable storage medium, characterized in that: It stores instructions that, when executed by one or more processors, cause the processors to perform the guided movement analysis method as described in any one of claims 1 to 6.

Images