Motion assessment methods, devices, equipment and storage media

Abstract

This application discloses a motion assessment method, apparatus, device, and storage medium. The method includes: acquiring motion data of a target object; extracting actual key point data from the motion data; determining assessment index values ​​based on the actual key point data; and determining the motion assessment result of the target object based on the assessment index values. This method accurately extracts the actual key point data of the target object from the motion data and determines the assessment index values, thereby determining the motion assessment result based on the assessment index values. This reduces the subjective error of manual assessment, improves the accuracy and reliability of motion assessment, and is beneficial for providing effective personalized guidance for the motion movements of the target object.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision, and in particular to a motion evaluation method and device, equipment and a storage medium. BACKGROUND

[0002] With the development of the economy and society, people's living standards have improved, and the importance of health has also increased. More and more people participate in sports to achieve the purpose of exercising the body and enhancing physical fitness, which promotes the growing demand for sports training, such as ball games, and drives the development of related sports training technology.

[0003] Traditional motion evaluation methods rely on artificial subjective guidance and evaluation, which is easily affected by subjective factors and has low accuracy in evaluating the motion of athletes. SUMMARY

[0004] The present application provides a motion evaluation method, device, computer equipment and storage medium, which aims to improve the accuracy of motion evaluation.

[0005] In a first aspect, a motion evaluation method is provided, comprising:

[0006] obtaining motion data of a target object;

[0007] extracting actual key element data from the motion data;

[0008] determining an evaluation index value according to the actual key element data;

[0009] determining a motion evaluation result of the target object according to the evaluation index value.

[0010] In a second aspect, a motion evaluation device is provided, comprising:

[0011] an acquisition module configured to obtain motion data of a target object;

[0012] an extraction module configured to extract actual key element data from the motion data;

[0013] a first determination module configured to determine an evaluation index value according to the actual key element data;

[0014] a second determination module configured to determine a motion evaluation result of the target object according to the evaluation index value.

[0015] Optionally, in some embodiments of the present application, the evaluation index value includes one or a combination of the following: motion accuracy, motion coordination, and motion speed.

[0016] Optionally, in some embodiments of this application, the actual key element data includes one or more combinations of the following: actual key point data, actual key line data, and actual key region data.

[0017] Optionally, in some embodiments of this application, when the actual key element data includes the actual key point data, the first determining module is used to determine the evaluation index value based on the actual key point data; when the actual key element data includes the actual key line data and / or the actual key region data, the first determining module is used to determine at least one target key point data based on the actual key line data and / or the actual key region data; and to determine the evaluation index value based on the at least one target key point data.

[0018] Optionally, in some embodiments of this application, the first determining module includes:

[0019] The acquisition submodule is used to acquire standard key element data;

[0020] The first determining submodule is used to determine the accuracy of the action based on the actual key element data and the standard key element data.

[0021] Optionally, in some embodiments of this application, the first determining submodule includes:

[0022] The first calculation unit is used to calculate the first curvature data and / or the first angle data based on the actual key element data, and to calculate the second curvature data and / or the second angle data based on the preset standard key element data.

[0023] The second calculation unit is used to calculate a first similarity based on the first curvature data and the second curvature data and / or to calculate a second similarity based on the first angle data and the second angle data;

[0024] The first determining unit is configured to determine the accuracy of the action based on the first similarity and / or the second similarity.

[0025] Optionally, in some embodiments of this application, the apparatus further includes:

[0026] The first processing submodule is used to perform frame alignment processing on the actual key element data and the standard key element data.

[0027] Optionally, in some embodiments of this application, the first determining module includes:

[0028] The first calculation submodule is used to calculate multiple adjacent point velocity sequences based on each set of key element sequences in the actual key element data.

[0029] The second determining submodule is used to determine the acceleration peak time point of each set of key element sequences based on the velocity sequences of multiple adjacent points;

[0030] The third determining submodule is used to determine the degree of coordination of the action based on each of the acceleration peak time points.

[0031] Optionally, in some embodiments of this application, the third determining submodule includes:

[0032] A pairing unit is used to pair each of the peak time points based on a preset sequence of key element names;

[0033] The third calculation unit is used to calculate the time difference between each pair of peak time points;

[0034] The fourth calculation unit is used to calculate the average absolute deviation of each of the time differences;

[0035] The second determining unit is used to determine the degree of motion coordination based on the mean absolute deviation.

[0036] Optionally, in some embodiments of this application, the first determining module includes:

[0037] The second calculation submodule is used to calculate multiple adjacent point velocity sequences based on each set of key element sequences in the actual key element data.

[0038] The fourth determination submodule is used to determine the average velocity of each set of key element sequences based on the velocity sequences of multiple adjacent points;

[0039] The fifth determining submodule is used to determine the motion speed based on each of the average speeds.

[0040] Optionally, in some embodiments of this application, the apparatus further includes:

[0041] The processing module is used to normalize the actual key element data.

[0042] Optionally, in some embodiments of this application, the processing module includes:

[0043] The sixth determining submodule is used to determine the first positioning point and the second positioning point in the actual key element data;

[0044] The third calculation submodule is used to calculate the distance between the first positioning point and the corresponding second positioning point;

[0045] The second processing submodule is used to normalize the actual key element data according to the distance.

[0046] Optionally, in some embodiments of this application, the actual key element data includes human body key element data and / or hitting device key element data.

[0047] Optionally, in some embodiments of this application, the target object is a tennis player, and the sports data is data generated by the tennis player during tennis training.

[0048] Thirdly, a motion assessment system is provided, including data acquisition equipment, tennis service equipment, and display equipment;

[0049] The data acquisition device is used to acquire motion data of the target object and send it to the tennis service device; the tennis service device is used to extract actual key point data from the motion data; determine the evaluation index value based on the actual key point data; and determine the motion evaluation result of the target object based on the evaluation index value; the display device is used to provide real-time feedback of the motion evaluation result.

[0050] Fourthly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described action evaluation method.

[0051] Fifthly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described action evaluation method.

[0052] This application provides a motion assessment method, apparatus, computer device, and storage medium. The method involves acquiring motion data of a target object; extracting actual key point data from the motion data; determining assessment index values ​​based on the actual key point data; and determining the motion assessment result of the target object based on the assessment index values. In the motion assessment scheme provided by this application, by accurately extracting the actual key point data of the target object from the motion data and determining the assessment index values, the motion assessment result is determined based on the assessment index values. This reduces the subjective error of manual assessment and improves the accuracy and reliability of motion assessment. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1This is a diagram illustrating the application environment of the motion evaluation method provided in the embodiments of this application.

[0055] Figure 2 An application environment diagram for the motion evaluation method provided in another embodiment of this application;

[0056] Figure 3 A schematic diagram illustrating key points of the target object during golf training, as provided in an embodiment of this application;

[0057] Figure 4 A schematic diagram illustrating key points of the target object during table tennis training, as provided in the embodiments of this application;

[0058] Figure 5 A schematic diagram of key points of the target object during badminton training, provided in an embodiment of this application;

[0059] Figure 6 A flowchart of the motion evaluation method provided in the embodiments of this application;

[0060] Figure 7 A schematic diagram illustrating the evaluation index values ​​provided in the embodiments of this application;

[0061] Figure 8 This is a structural block diagram of the motion evaluation device provided in the embodiments of this application;

[0062] Figure 9 This is a structural block diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0065] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0066] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0067] The motion evaluation method provided in this invention can be applied to, for example... Figure 1 In this application environment, computer device 110 communicates with server 120 via network 130. Computer device 110 can acquire motion data of the target object; extract actual key point data from the motion data; determine evaluation index values ​​based on the actual key point data; determine the action evaluation result of the target object based on the evaluation index values, and display it through computer device 110. In this invention, by accurately extracting the actual key point data of the target object from the motion data and determining the evaluation index values, the action evaluation result is determined based on the evaluation index values, reducing the subjective error of manual evaluation and improving the accuracy and reliability of action evaluation. Computer device 110 may include, but is not limited to, various smartphones 110-1, tablet computers 110-2, and laptop computers 110-3. The invention will be described in detail below through specific embodiments.

[0068] The motion evaluation method provided in this invention can be applied to, for example... Figure 2 The motion evaluation system shown includes a data acquisition device 140, a tennis service device 160, and a display device 170. The data acquisition device 140 acquires motion data of a target object 150 and sends it to the tennis service device 160. The tennis service device 160 extracts actual key point data from the motion data, determines evaluation index values ​​based on the actual key point data, and determines the motion evaluation result of the target object 150 based on the evaluation index values. The display device 170 provides real-time feedback of the motion evaluation result.

[0069] The data acquisition device 140 acquires motion data of the target object 150 via network 130 and sends it to the tennis service device 160; the tennis service device 160 extracts actual key point data from the motion data via network 130, determines evaluation index values ​​based on the actual key point data, and determines the motion evaluation result of the target object 150 based on the evaluation index values; the display device 170 acquires the motion evaluation result from the tennis service device 160 in real time via network 130 and provides feedback.

[0070] The data acquisition device 140 can be a single camera, which can capture the movement of the target object from different angles. The camera needs to have sufficient resolution and frame rate to ensure accurate capture of the key details of the target object's rapid movements. The motion evaluation method provided in this application can be applied to a variety of different training scenarios. For example, in tennis training, whether on a standard tennis court, a tennis ball machine practice area, or a virtual reality (AR / VR) training platform, a single camera can achieve accurate motion capture and analysis. The tennis service device 160 can identify different scenarios and perform adaptive evaluation in the identified scenarios (such as comparing the standard key element data in the adapted scenario with the actual key element data) to ensure the consistency of the motion evaluation results. The motion evaluation results are fed back in real time through the display device 170, such as through screen, voice, or other feedback methods to provide real-time prompts to the athlete. Athletes can adjust their movements and improve their technical level through this real-time feedback. The motion evaluation method of this application can generate a technical score for each movement of each target object. This technical score includes one or more dimensions, such as movement accuracy, movement coordination, and movement speed.

[0071] Furthermore, the movement assessment method of this application can also track and analyze athletes' training data (including movement data and corresponding movement assessment results) over a long period. Recording the target athlete's movements and assessment results for each training session and generating a technical progress report according to a preset template helps coaches and target athletes evaluate training effectiveness and formulate subsequent training goals.

[0072] In this application, data collection is performed using a single camera, avoiding the high costs associated with multiple cameras and complex hardware, and reducing equipment procurement and maintenance expenses. Through real-time, accurate motion assessment and personalized guidance, the target audience can promptly identify and adjust technical deficiencies, significantly improving training effectiveness. It is applicable to various training scenarios, such as tennis training scenarios including tennis ball machines, standard courts, and AR / VR training platforms, demonstrating broad applicability and scalability. Utilizing artificial intelligence algorithms for motion analysis and assessment, combined with data-driven personalized guidance, ensures more scientific and intelligent training. It can track the target audience's technical progress over the long term, generating detailed technical reports, which helps both the target audience and coaches better plan training paths.

[0073] Please see Figure 3 As shown, Figure 3 This is a flowchart illustrating an action evaluation method provided in an embodiment of the present invention. This method can be applied to both terminals and servers; this embodiment uses server-side application as an example. The action evaluation method includes the following steps:

[0074] S101: Obtain motion data of the target object.

[0075] The target object is the athlete being evaluated, which can be a ball sports athlete, track and field athlete, water sports athlete, combat athlete, etc. The sports data can be data generated by the athlete during training. For example, if the target object is a player performing a ball sports, the sports data could be continuous video frame data of the athlete's movements during ball sports (such as tennis, badminton, table tennis, etc.).

[0076] In one embodiment, the target object is a tennis player, and the sports data is data generated by the tennis player during tennis training. Alternatively, the target object can be a sprinter, and the sports data is data generated by the sprinter during sprint training; the target object can also be a swimmer, and the sports data is data generated by the swimmer during swimming training; or the target object can also be a wrestler, and the sports data is data generated by the wrestler during wrestling training.

[0077] Specifically, a single data acquisition device, such as a camera, can capture motion images of a target object from different angles (e.g., assuming a court includes a left, middle, and right area, different angles could include the angle of a player performing ball movements in the left, middle, and right areas of the court). These motion images can then be used as motion data of the target object. In one embodiment, a video processing library, such as OpenCV, can be used to read motion data from the data acquisition device using a Real-Time Streaming Protocol (RTSP).

[0078] S102: Extract actual key element data from the motion data.

[0079] The actual key element data may include motion data used to evaluate the motion characteristics of the target object.

[0080] Optionally, the actual key element data may include human body key element data and / or ball-hitting equipment key element data. Human body key point data may include the location information of human joints, etc.; ball-hitting equipment key element data may include the location information of characteristic parts of the ball-hitting equipment, etc. For example, if the target object is a tennis player, when the tennis player is playing tennis, the ball-hitting equipment is a tennis racket, and the human body joints may include key points of body parts such as the head, shoulders, elbows, wrists, and knees, while the characteristic parts of the ball-hitting equipment may include key points of parts such as the racket face, racket head, and racket handle. As another example, if the target object is a basketball player, when the basketball player is shooting a basketball, the ball-hitting equipment is the basketball player's arm, and the human body joints may include the head, shoulders, elbows, wrists, knees, ankles, and hips, etc., while the characteristic parts of the ball-hitting equipment may include the palm, fingers, forearm, upper arm, and shoulder joint, etc.

[0081] In one embodiment, the actual key element data includes one or more combinations of the following: actual key point data, actual key line data, and actual key region data.

[0082] The actual keypoint data can be the keypoint data used to evaluate the motion characteristics of the target object in the motion data, and this keypoint data can include the coordinates of each keypoint. The actual keyline data can be the connecting line data used to evaluate the motion characteristics of the target object in the motion data, and this connecting line data can include the skeletal connecting line data of the human body and / or the equipment connecting line data of the hitting device. The skeletal connecting line data includes information such as the length, direction, and angle of each skeletal connecting line, and the equipment connecting line data includes information such as the length, direction, and angle of each skeletal connecting line. The actual key region data can be the key region data used to evaluate the motion characteristics of the target object in the motion data, and this key region data can include the location information of each key region.

[0083] In one embodiment, computer vision techniques, such as image processing algorithms and deep learning algorithms, can be used, and deep learning pose estimation networks, such as transpose networks or high-resolution networks (HRNet), can be used to extract key elements from motion data to obtain actual key element data.

[0084] In one embodiment, the motion data of the target object can be segmented and extracted to obtain multiple motion images; key points of the human body and key points of the hitting device in each motion image can be extracted, and actual key point data can be obtained based on the key points of the human body and key points of the hitting device in each motion image. The multiple motion images are images of a series of continuous actions that occur at different points in time when the target object is moving.

[0085] Specifically, target detection algorithms (such as YOLO and SSD) are used to detect target actions in each frame of the motion data, that is, to identify the player's actions and the hitting equipment (such as a racket), to obtain multiple frames of motion images of the target object during action training (such as from the start of the hit to the end of the hit); then, pose estimation algorithms (such as OpenPose and HRNet) are used to extract key points from the multiple frames of motion images, to obtain human key points and hitting equipment key points in each frame of motion images; finally, the human key points and hitting equipment key points in each frame of motion images are used as actual key point data.

[0086] For example, such as Figure 3As shown, when the target is practicing golf, the identified key points of the human body and the hitting equipment include the head key point a11, neck key point a12, left shoulder key point a13, right shoulder key point a14, left hip key point a15, grip end key point a16, left knee key point a17, right knee key point a18, left ankle key point a19, right ankle key point a20, hose key point a21, sweet spot (optimal hitting position key point) a22, right elbow key point a24, and right wrist key point a25. Among them, a23 is the key area of ​​the golf club. The grip end key point a16, hose key point a21, and sweet spot (optimal hitting position key point) a22 can be identified through the detected key areas of the golf club, i.e., the target key point.

[0087] For example, such as Figure 4 As shown, when the target is training in table tennis, the identified key points of the human body and the hitting equipment include the head key point b11, neck key point b12, left shoulder key point b13, right shoulder key point b14, left wrist key point b15, grip key point b16, racket left shoulder key point b17, racket right shoulder key point b18, racket head key point b19, left hip key point b20, right hip key point b21, left elbow key point b23, and right elbow key point b24. Among them, b22 is the key area of ​​the table tennis racket. The grip key point b16, racket left shoulder key point b17, racket right shoulder key point b18, and racket head key point b19, i.e., the target key point, can be identified by detecting the key areas of the table tennis racket.

[0088] For example, such as Figure 5 As shown, when the target is training in badminton, the identified key points of the human body and the hitting equipment include: head key point c11, neck key point c12, left shoulder key point c13, right shoulder key point c14, left elbow key point c15, left wrist key point c16, left hip key point c17, right hip key point c18, right elbow key point c19, right wrist key point c20, top of racket handle key point c21, top of racket head key point c22, bottom of racket head key point c24, right knee key point c25, left knee key point c26, left ankle key point c27, and right ankle key point c28. Among them, c23 is the key area of ​​the badminton racket. The key points of the top of the racket handle c21, top of the racket head key point c22, and bottom of the racket head key point c24, i.e., the target key points, can be identified by detecting the key areas of the badminton racket.

[0089] In one embodiment, after extracting the actual keypoint data from the motion data, to avoid keypoints being missing due to occlusion at different angles of the target object in the training field, thus affecting the accuracy of motion evaluation, keypoint integrity detection can be performed on the actual keypoint data. If missing keypoints are detected in the actual keypoint data, the missing keypoints can be marked and predicted, and the predicted keypoints can be filled into the actual keypoint data, thereby obtaining more complete and accurate actual keypoint data. For example, deep learning models, such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), or Transformers, can be used to predict the location of missing keypoints in the actual keypoint data. This deep learning model can learn the spatial and temporal relationships between keypoints, thereby making accurate predictions when keypoints are missing. Alternatively, optical flow methods can be used to estimate the motion vectors of pixels between adjacent frames in multi-frame motion images, thereby predicting missing keypoints. In this embodiment, by using keypoint integrity detection and prediction, the error caused by missing keypoints is reduced, and the accuracy of motion evaluation is improved.

[0090] In one embodiment, before determining the evaluation index value based on the actual key point data, the method further includes:

[0091] The actual key point data is normalized.

[0092] The mean of all keypoints in the actual keypoint data can be calculated as the center point; the standard deviation of the distance between all keypoints in the actual keypoint data and the center point can be calculated as the scaling factor; the coordinates of all keypoints in the actual keypoint data are subtracted from the coordinates of the center point to achieve centering; and the coordinates of the center point are divided by the scaling factor to achieve scaling normalization of the actual keypoint data.

[0093] In one application scenario, to eliminate the influence of the target object's height and arm length, the target object's height and arm length can be used as scale factors to normalize the actual keypoint data. Specifically, in one embodiment, the normalization of the actual keypoint data includes:

[0094] Determine the first and second positioning points in the actual key point data;

[0095] Calculate the distance between the first positioning point and the corresponding second positioning point;

[0096] The actual key point data is normalized based on the distance.

[0097] The first and second positioning points are two reference keypoints of the target object, used to normalize the actual keypoint data. For example, assuming the first positioning point is the top of the target object's head, the second positioning point is the bottom of the target object's feet; assuming the first positioning point is the shoulder of the target object, the second positioning point is the wrist of the target object. Optionally, if the actual keypoint data includes keypoints corresponding to multiple frames of motion images, the first and second positioning points can be determined from all keypoints corresponding to any frame of motion image.

[0098] For example, if the target object is a tennis player, the first positioning point is the key point corresponding to the top of the tennis player's head, and the second positioning point is the key point corresponding to the bottom of the tennis player's feet. The Euclidean distance between the key point on the top of the head and the key point on the bottom of the feet can be calculated. The coordinates of all key points in the actual key point data can be subtracted from the coordinates of the first positioning point to achieve centering. The coordinates of all key points after centering are divided by the Euclidean distance to achieve normalization of the actual key point data.

[0099] In this embodiment, the actual key point data is normalized by the distance between the first positioning point and the second positioning point, which reduces the impact of differences in position and scale between different individuals and helps to improve the accuracy of motion assessment.

[0100] S103: Determine the evaluation index value based on the actual key element data.

[0101] The evaluation index value is an indicator used to assess the standard of the target object's movements. The evaluation index value may include one or more of the following combinations: movement accuracy, movement coordination, and movement speed.

[0102] In one embodiment, determining the accuracy of the action based on the actual key element data includes:

[0103] Obtain standard key element data;

[0104] The accuracy of the action is determined based on the actual key element data and the standard key element data.

[0105] The standard key element data is obtained by extracting key elements from the motion data corresponding to the standard movements. The extraction method can refer to the actual key element data acquisition method, and will not be repeated here to avoid duplication. The standard motion data can be continuous video frame data of a reference individual performing standard movements (such as tennis, badminton, table tennis, dancing, swimming, etc.) captured by a single camera, as a standard reference.

[0106] Artificial intelligence algorithms, such as convolutional neural networks (CNN) and recurrent neural networks (LSTM), can be used to compare the features of actual key element data and standard key element data to obtain the accuracy of the action.

[0107] In one embodiment, when the actual key element data includes the actual key point data, determining the evaluation index value based on the actual key element data includes determining the evaluation index value based on the actual key point data; when the actual key element data includes the actual key line data and / or the actual key region data, determining the evaluation index value based on the actual key element data includes: determining at least one target key point data based on the actual key line data and / or the actual key region data; and determining the evaluation index value based on the at least one target key point data.

[0108] The target keypoint data consists of target keypoints determined based on actual keyline data and / or actual key region data. For example, assuming the actual keyline data includes the head-neck skeleton connection line, shoulder-elbow connection line, elbow-wrist connection line, and clapper head-and-clapper tail connection line for each frame of the image, then the head keypoint, neck keypoint, shoulder keypoint, elbow keypoint, wrist keypoint, and clapper head keypoint for each frame of the motion image can be determined based on the head-neck skeleton connection line, shoulder-elbow connection line, and elbow-wrist connection line, thus obtaining the target keypoint data. As another example, when performing target detection on motion data, each frame of the motion image can be marked with markers such as circles, triangles, and squares, and the center point of the marked shape can be determined as the target keypoint, such as the head keypoint, neck keypoint, shoulder keypoint, elbow keypoint, wrist keypoint, and clapper head keypoint, thus obtaining the target keypoint data.

[0109] In one embodiment, determining the accuracy of the action based on the actual key element data and the standard key element data includes:

[0110] The first curvature data and / or the first angle data are used to calculate the actual key element data, and the second curvature data and / or the second angle data are used to calculate the standard key element data;

[0111] A first similarity is calculated based on the first curvature data and the second curvature data, and / or a second similarity is calculated based on the first angle data and the second angle data;

[0112] The accuracy of the action is calculated based on the first similarity and / or the second similarity.

[0113] Among them, motion accuracy is used to evaluate the accuracy of the target object's motion movements.

[0114] When the actual key element data includes actual key point data, and the standard key element data correspondingly includes standard key point data, the first curvature data and / or the first angle data are calculated based on the actual key point data. Similarly, the second curvature data and / or the second angle data are calculated based on the standard key point data. When the actual key element data includes actual key line data and / or actual key region data, and the standard key element data may include standard key line data and / or standard key region data, the actual target key point data can be determined based on the actual key line data and / or the actual key region data. Similarly, the standard target key point data can be determined.

[0115] The first curvature data includes the curvature of the curve between adjacent keypoints on the curve connecting the same keypoint (e.g., the head of a tennis racket) in multiple frames of motion images corresponding to the actual keypoint data. The second curvature data includes the curvature of the curve between adjacent keypoints on the curve connecting the same keypoint (e.g., the head of a tennis racket) in multiple frames of motion images corresponding to the standard keypoint data. The first angle data includes the angle between the slope of the curve connecting adjacent keypoints on the curve connecting the same keypoint (e.g., the head of a tennis racket) in multiple frames of motion images corresponding to the actual keypoint data and a reference line (i.e., the horizontal line). The second angle data includes the angle between the slope of the curve connecting adjacent keypoints on the curve connecting the same keypoint (e.g., the head of a tennis racket) in multiple frames of motion images corresponding to the standard keypoint data and a reference line (i.e., the horizontal line).

[0116] Similarity algorithms (such as cosine similarity) can be used to calculate the similarity value of corresponding curvatures between the first curvature data and the second curvature data, and the average of each similarity value is calculated to obtain the first similarity. Similarly, similarity algorithms (such as cosine similarity) can be used to calculate the similarity value of corresponding angles between the first angle data and the second angle data, and the average of each similarity value is calculated to obtain the second similarity. The first similarity value and / or the second similarity value can be used as the motion accuracy. Alternatively, the first similarity and the second similarity can be weighted and summed, and the calculated sum can be used as the motion accuracy.

[0117] In one embodiment, assuming that the curvature of the curve for the same key point (such as a striking point) in actual key point data or target key point data is [0.1, 0.2, 0.3, 0.4, 0.5], and the curvature of the curve for the same key point in standard key point data is [0.15, 0.25, 0.35, 0.45, 0.55], calculate the similarity value of the curvature between the two:

[0118]

[0119] Suppose that in actual keypoint data, all angles of the curve for the same keypoint (such as the keypoint of a certain hitting body) are [30, 60, 90, 120, 150], and in standard keypoint data, all angles of the curve for the same keypoint are [35, 65, 95, 125, 155]. Calculate the similarity of the angles between the two:

[0120]

[0121] Similarly, the curvature similarity value and angle similarity value of other key points can be calculated; thus, the mean of the curvature similarity values ​​of the same key point can be calculated to obtain the first similarity value, and the angle similarity value of the same key point can be calculated to obtain the second similarity value.

[0122] Furthermore, the action accuracy can be normalized, that is, mapped to a preset range, such as [0, 100]. For example, assuming that the maximum value of action accuracy is determined based on historical experience data, such as 0 and the minimum value is such as 1, and the action accuracy is 0.85, then mapping 0.85 to [0, 100] can be expressed as:

[0123]

[0124] In one embodiment, before determining the accuracy of the action based on the actual key element data and the standard key element data, the method further includes:

[0125] The actual key element data and the standard key element data are frame aligned.

[0126] Optionally, dynamic time warping can be used to perform frame alignment processing between the actual key element data and the standard key element data. Specifically, a similarity algorithm such as Euclidean distance or cosine similarity is used to calculate the distance matrix between the actual key element data and the standard key element data; a cumulative distance matrix is ​​constructed based on this distance matrix to record the optimal alignment path; and a dynamic programming algorithm is used to find the optimal path on the cumulative distance matrix, so that the actual key element data and the standard key element data are aligned in time series, thereby achieving frame alignment.

[0127] Optionally, an optical flow algorithm can be used to perform frame alignment processing between the actual key element data and the standard key element data. Specifically, an optical flow algorithm (such as the Lucas-Kanade method) is used to calculate the motion vector of the key elements in each frame; based on the calculated optical flow (i.e., motion vector), motion compensation is performed on the actual key element data to align it with the standard key element data frame.

[0128] Optionally, a deformable convolutional network can be used to perform frame alignment processing between the actual key feature data and the standard key feature data. Specifically, the deformable convolutional network calculates the offset of key features in each frame of the motion image, and adjusts the key features based on the calculated offset, such as adjusting the coordinates of key points, to achieve frame alignment between the actual key feature data and the standard key feature data.

[0129] In this embodiment, frame alignment processing helps to reduce errors caused by inconsistencies in time series and improves the accuracy of motion evaluation.

[0130] In one embodiment, determining the degree of motion coordination based on the actual key element data includes:

[0131] Based on each set of key element sequences in the actual key element data, multiple velocity sequences of adjacent points are calculated.

[0132] The acceleration peak time point of each set of key element sequences is determined based on the velocity sequences of multiple adjacent points.

[0133] The degree of motion coordination is calculated based on each of the aforementioned acceleration peak time points.

[0134] Among them, motion coordination is used to assess the ability of different parts of the body to coordinate when a target object is moving.

[0135] The actual key element data includes multiple sets of key element sequences. When the actual key element data includes actual key point data, each set of key element sequences can be a vector composed of the same key point in the actual key point data corresponding to the continuous motion image. When the actual key element data includes actual key line data and / or actual key region data, each set of key element sequences can be a vector composed of the same key point in the target key point data corresponding to the continuous motion image.

[0136] It can calculate the displacement of all adjacent keypoints in a key element sequence and the time difference between adjacent keypoints based on the timestamps of adjacent frames. This allows for the calculation of the velocity between adjacent keypoints in the key point sequence. The velocities of adjacent points in each velocity sequence are sorted to determine the acceleration peak time point (i.e., the timestamp corresponding to the maximum velocity of the adjacent point) for each velocity sequence. Based on each acceleration peak time point, the keypoint names corresponding to each set of key element sequences are sorted to obtain a keypoint name sequence. If the keypoint name sequence meets preset basic conditions, the preset value corresponding to meeting the preset basic conditions is obtained as the motion coordination degree. The preset basic conditions are a preset keypoint name sorting, where the acceleration peak time points of the keypoint name sequence are sorted from largest to smallest.

[0137] For example, assuming that key points in the human body include the hip, shoulder, elbow (upper arm), and wrist, and key points in the hitting equipment include the racket head, then we can obtain the key element sequences corresponding to the hip, shoulder, elbow (upper arm), wrist, and racket head in continuous motion images. We can then calculate the velocities of adjacent points in each key element sequence to obtain the adjacent point velocity sequence for each set of key element sequences. Finally, we sort the adjacent point velocities in each adjacent point velocity sequence to determine the velocity sequence for each adjacent point. The corresponding acceleration peak time point is used to sort the key point names corresponding to each set of key element sequences according to each acceleration peak time point, resulting in a key point name sequence. If the key name sequence is [hip key point, shoulder key point, elbow (upper arm) key point, wrist key point, and racket head key point], that is, the peak time point of the hip key point is earlier than the peak time point of the shoulder key point, which is earlier than the peak time point of the elbow (upper arm) key point, and earlier than the peak time point of the wrist key point, which is earlier than the peak time point of the racket head key point, then it means that the target object's movement meets the basic requirements of coordination (i.e., the preset basic conditions). If the preset basic conditions are met, the preset value corresponding to the movement coordination is b, and b is taken as the movement coordination degree.

[0138] In this embodiment, by calculating the velocity of adjacent points and the peak acceleration time, the change in the target object's motion speed at different time points can be analyzed more accurately. This is beneficial for analyzing the smoothness and rhythm of the motion and for more accurately assessing the athlete's motion coordination.

[0139] In one embodiment, calculating the motion coordination degree based on each of the acceleration peak time points includes:

[0140] The peak time points are paired based on a preset sequence of key element names;

[0141] Calculate the time difference for each pair of peak time points;

[0142] Calculate the mean absolute deviation of each of the aforementioned time differences;

[0143] The degree of motion coordination is calculated based on the mean absolute deviation.

[0144] Mean Absolute Deviation (MAD) is used to quantify the variation or dispersion of each time difference. It measures the dispersion of the data by calculating the average of the absolute differences between each time difference and the average of all time differences.

[0145] For example, assuming the keypoint name sequence is [hip keypoint, shoulder keypoint, elbow (upper arm) keypoint, wrist keypoint, and racket head keypoint], the peak time points are paired based on this keypoint name sequence. Then, the time difference between the peak time points of the hip and shoulder keypoints, the peak time difference between the shoulder and elbow (upper arm) keypoints, the peak time difference between the elbow (upper arm) and wrist keypoints, and the peak time difference between the wrist and racket head keypoints are calculated. The mean absolute deviation of each time difference is then calculated, and this mean absolute deviation is used as the degree of motion coordination. The smaller the mean absolute deviation, the higher the degree of motion coordination; conversely, the larger the mean absolute deviation, the lower the degree of motion coordination.

[0146] Optionally, the mean absolute deviation can be mapped to a preset interval, such as [0, 100]. For example, assuming that the maximum value of the movement accuracy is determined based on historical experience data, such as 0 and the minimum value is such as 1, and the movement coordination is 0.85, then mapping 0.85 to [0, 100] can be expressed as:

[0147]

[0148] In one embodiment, determining the motion speed based on the actual key element data includes:

[0149] Based on each set of key element sequences in the actual key element data, multiple velocity sequences of adjacent points are calculated.

[0150] The average velocity of each set of key element sequences is determined based on the velocity sequences of multiple adjacent points.

[0151] The motion speed is determined based on the average speeds described above.

[0152] The actual key element data includes multiple sets of key element sequences. When the actual key element data includes actual key point data, each set of key element sequences can be a vector composed of the same key point in the actual key point data corresponding to a continuous motion image. When the actual key element data includes actual key line data and / or actual key region data, each set of key element sequences can be a vector composed of the same key point in the target key point data corresponding to a continuous motion image. The displacement of all adjacent key points in the key element sequence can be calculated, and the time difference between adjacent key points can be calculated based on the timestamps of adjacent frames. Therefore, the velocity of adjacent points in each set of key element sequences can be calculated, and the average velocity corresponding to each set of key element sequences can be obtained by averaging the velocities of all adjacent points in each set of key element sequences. The motion speed is then determined based on these average velocities.

[0153] For example, assuming that the key points of the human body include the hip key point, shoulder key point, elbow (upper arm) key point, and wrist key point, and the key points of the hitting equipment include the racket head key point, then we can obtain the key element sequence corresponding to the hip key point, the shoulder key point, the elbow (upper arm) key point, the wrist key point, and the racket head key point in the continuous motion image. We can then calculate the velocity of adjacent points between adjacent key points in each set of key element sequences, calculate the average velocity of each set of key element sequences, and obtain the average velocity of each set of key element sequences. We can choose the average velocity corresponding to the racket head key point as the motion velocity, or we can calculate only the average velocity corresponding to the racket head key point as the motion velocity.

[0154] S104: Determine the action evaluation result of the target object based on the evaluation index value.

[0155] Multiple evaluation index values ​​can be used as the evaluation results of the target object's actions.

[0156] Multiple evaluation index values ​​can be weighted and summed, and the sum can be used as the evaluation result of the player's action.

[0157] In one embodiment, such as Figure 7 As shown, multiple evaluation index values ​​can be mapped to the same preset interval, such as [0,100], and then each evaluation index value can be mapped to a preset technical scoring chart and displayed.

[0158] The above is the motion evaluation process for this application.

[0159] As described above, this application provides a motion evaluation method, which involves acquiring motion data of a target object; extracting actual key point data from the motion data; determining evaluation index values ​​based on the actual key point data; and determining the motion evaluation result of the target object based on the evaluation index values. In the motion evaluation scheme provided by this application, by accurately extracting the actual key point data of the target object from the motion data and determining the evaluation index values, the motion evaluation result is determined based on the evaluation index values, reducing the subjective error of manual evaluation and improving the accuracy and reliability of motion evaluation.

[0160] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0161] In one embodiment, a motion evaluation device is provided, which corresponds one-to-one with the motion evaluation methods in the above embodiments. It should be understood that various variations and specific embodiments of the motion evaluation methods provided in the above embodiments are also applicable to the motion evaluation device in this embodiment. Through the detailed description of the aforementioned motion evaluation methods, those skilled in the art can clearly understand the implementation process of the motion evaluation device in this embodiment.

[0162] Please see Figure 8 As shown, the motion evaluation device includes:

[0163] The acquisition module is used to acquire motion data of the target object;

[0164] The extraction module is used to extract actual key element data from the motion data;

[0165] The first determining module is used to determine the evaluation index value based on the actual key element data;

[0166] The second determining module is used to determine the action evaluation result of the target object based on the evaluation index value.

[0167] This application obtains motion data of a target object; extracts actual key point data from the motion data; determines evaluation index values ​​based on the actual key point data; and determines the motion evaluation result of the target object based on the evaluation index values. In the motion evaluation scheme provided by this application, by accurately extracting the actual key point data of the target object from the motion data and determining the evaluation index values, the motion evaluation result is determined based on the evaluation index values. This reduces the subjective error of manual evaluation, improves the accuracy and reliability of motion evaluation, and is beneficial for providing effective personalized guidance for the target object's motion movements.

[0168] Optionally, in some embodiments of this application, the evaluation index values ​​include one or more of the following combinations: motion accuracy, motion coordination, and motion speed.

[0169] Optionally, in some embodiments of this application, the actual key element data includes one or more combinations of the following: actual key point data, actual key line data, and actual key region data.

[0170] Optionally, in some embodiments of this application, when the actual key element data includes the actual key point data, the first determining module is used to determine the evaluation index value based on the actual key point data; when the actual key element data includes the actual key line data and / or the actual key region data, the first determining module is used to determine at least one target key point data based on the actual key line data and / or the actual key region data; and to determine the evaluation index value based on the at least one target key point data.

[0171] Optionally, in some embodiments of this application, the first determining module includes:

[0172] The acquisition submodule is used to acquire standard key element data;

[0173] The first determining submodule is used to determine the accuracy of the action based on the actual key element data and the standard key element data.

[0174] Optionally, in some embodiments of this application, the first determining submodule includes:

[0175] The first calculation unit is used to calculate the first curvature data and / or the first angle data based on the actual key element data, and to calculate the second curvature data and / or the second angle data based on the preset standard key element data.

[0176] The second calculation unit is used to calculate a first similarity based on the first curvature data and the second curvature data and / or to calculate a second similarity based on the first angle data and the second angle data;

[0177] The first determining unit is configured to determine the accuracy of the action based on the first similarity and / or the second similarity.

[0178] Optionally, in some embodiments of this application, the apparatus further includes:

[0179] The first processing submodule is used to perform frame alignment processing on the actual key element data and the standard key element data.

[0180] Optionally, in some embodiments of this application, the first determining module includes:

[0181] The first calculation submodule is used to calculate multiple adjacent point velocity sequences based on each set of key element sequences in the actual key element data.

[0182] The second determining submodule is used to determine the acceleration peak time point of each set of key element sequences based on the velocity sequences of multiple adjacent points;

[0183] The third determining submodule is used to determine the degree of coordination of the action based on each of the acceleration peak time points.

[0184] Optionally, in some embodiments of this application, the third determining submodule includes:

[0185] A pairing unit is used to pair each of the peak time points based on a preset sequence of key element names;

[0186] The third calculation unit is used to calculate the time difference between each pair of peak time points;

[0187] The fourth calculation unit is used to calculate the average absolute deviation of each of the time differences;

[0188] The second determining unit is used to determine the degree of motion coordination based on the mean absolute deviation.

[0189] Optionally, in some embodiments of this application, the first determining module includes:

[0190] The second calculation submodule is used to calculate multiple adjacent point velocity sequences based on each set of key element sequences in the actual key element data.

[0191] The fourth determination submodule is used to determine the average velocity of each set of key element sequences based on the velocity sequences of multiple adjacent points;

[0192] The fifth determining submodule is used to determine the motion speed based on each of the average speeds.

[0193] Optionally, in some embodiments of this application, the apparatus further includes:

[0194] The processing module is used to normalize the actual key element data.

[0195] Optionally, in some embodiments of this application, the processing module includes:

[0196] The sixth determining submodule is used to determine the first positioning point and the second positioning point in the actual key element data;

[0197] The third calculation submodule is used to calculate the distance between the first positioning point and the corresponding second positioning point;

[0198] The second processing submodule is used to normalize the actual key element data according to the distance.

[0199] Optionally, in some embodiments of this application, the actual key element data includes human body key element data and / or hitting device key element data.

[0200] Optionally, in some embodiments of this application, the target object is a tennis player, and the sports data is data generated by the tennis player during tennis training.

[0201] In one embodiment, a computer device is provided, the internal structure of which can be as follows: Figure 9 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When executed by the processor, the computer program implements the functions or steps of an action evaluation method.

[0202] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0203] Acquire motion data of the target object; extract actual key point data from the motion data; determine evaluation index values ​​based on the actual key point data; determine the motion evaluation result of the target object based on the evaluation index values.

[0204] In this embodiment, by accurately extracting the actual key point data of the target object from the motion data and determining the evaluation index value, the motion evaluation result is determined based on the evaluation index value, which reduces the subjective error of manual evaluation and improves the accuracy and reliability of motion evaluation.

[0205] In one embodiment, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, performs the following steps:

[0206] Acquire motion data of the target object; extract actual key point data from the motion data; determine evaluation index values ​​based on the actual key point data; determine the motion evaluation result of the target object based on the evaluation index values.

[0207] In this embodiment, by accurately extracting the actual key point data of the target object from the motion data and determining the evaluation index value, the motion evaluation result is determined based on the evaluation index value, which reduces the subjective error of manual evaluation and improves the accuracy and reliability of motion evaluation.

[0208] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0209] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0210] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0211] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method of motion assessment, characterized by, The method comprises: acquiring motion data of a target object; the motion data comprises a video sequence captured by a single camera; extracting actual key element data from the motion data; the actual key element data comprises actual key point data, actual key line data, and actual key region data; determining evaluation index values according to the actual key element data; the evaluation index values comprise action accuracy, action coordination, and motion speed; the determination of the action coordination comprises: calculating a plurality of adjacent point speed sequences based on each group of key element sequences in the actual key element data; determining an acceleration peak time point of each group of key element sequences based on a plurality of the adjacent point speed sequences; pairing each of the peak time points based on a preset sequence of key element names; calculating a time difference of each pair of peak time points; calculating an average absolute deviation of each of the time differences; and determining the action coordination according to the average absolute deviation; determining an action evaluation result of the target object according to the evaluation index values.

2. The motion evaluation method according to claim 1, characterized by, When the actual key element data comprises the actual key point data, the determining of the evaluation index values according to the actual key element data comprises determining the evaluation index values according to the actual key point data. When the actual key element data comprises the actual key line data and / or the actual key region data, the determining of the evaluation index values according to the actual key element data comprises: determining at least one target key point data according to the actual key line data and / or the actual key region data; and determining the evaluation index values according to the at least one target key point data. The determining of the action accuracy according to the actual key element data comprises:

3. The motion evaluation method according to claim 2, characterized in that, acquiring standard key element data; determining the action accuracy according to the actual key element data and the standard key element data. The determining of the action accuracy according to the actual key element data and the standard key element data comprises:

4. The motion evaluation method according to claim 3, characterized by, calculating first curvature data and / or first angle data according to the actual key element data, and calculating second curvature data and / or second angle data according to preset standard key element data; calculating a first similarity according to the first curvature data and the second curvature data, and / or calculating a second similarity according to the first angle data and the second angle data; determining the action accuracy according to the first similarity and / or the second similarity. Before the determining of the action accuracy according to the actual key element data and the standard key element data, the method further comprises:

5. The motion evaluation method according to claim 3, characterized by, performing frame alignment processing on the actual key element data and the standard key element data. The determining of the motion speed according to the actual key element data comprises:

6. The motion evaluation method according to claim 2, characterized by, calculating a plurality of adjacent point speed sequences based on each group of key element sequences in the actual key element data; determining an average speed of each group of key element sequences based on a plurality of the adjacent point speed sequences; determining the motion speed based on each of the average speeds. Before the determining of the evaluation index values according to the actual key element data, the method further comprises:

7. The motion evaluation method according to claim 1, characterized by, ​ Normalizing the actual key element data.

8. The motion assessment method of claim 7, wherein, The normalizing the actual key element data comprises: determining a first positioning point and a second positioning point in the actual key element data; calculating a distance between the first positioning point and the corresponding second positioning point; normalizing the actual key element data according to the distance.

9. The motion assessment method according to any one of claims 1 to 8, characterized in that, The actual key element data comprises human body key element data and / or hitting equipment key element data.

10. The motion assessment method according to any one of claims 1 to 8, characterized in that, The target object is a tennis player, and the motion data is data generated when the tennis player performs tennis training.

11. An action evaluation device characterized by comprising: Comprise: an acquisition module configured to acquire motion data of a target object; the motion data comprises a video sequence collected by a single camera; an extraction module configured to extract actual key element data from the motion data; The actual key element data comprises actual key point data, actual key line data, and actual key region data; a first determination module configured to determine evaluation index values according to the actual key element data; the evaluation index values comprise action accuracy, action coordination, and motion speed; the determination of the action coordination comprises: calculating a plurality of adjacent point speed sequences based on each group of key element sequences in the actual key element data; determining an acceleration peak time point of each group of key element sequences based on a plurality of the adjacent point speed sequences; pairing each of the peak time points based on a preset key element name sequence; calculating a time difference of each pair of peak time points; calculating an average absolute deviation of each of the time differences; and determining the action coordination according to the average absolute deviation; a second determination module configured to determine an action evaluation result of the target object according to the evaluation index values.

12. A motion assessment system, characterized by Comprise a data acquisition device, a tennis serving device, and a display device; The data acquisition device is configured to acquire motion data of a target object and send the motion data to the tennis serving device; the motion data comprises a video sequence collected by a single camera; The tennis serving device is configured to extract actual key element data from the motion data; determine evaluation index values according to the actual key element data; and determine an action evaluation result of the target object according to the evaluation index values; The actual key element data comprises actual key point data, actual key line data, and actual key region data; the evaluation index values comprise action accuracy, action coordination, and motion speed; the determination of the action coordination comprises: calculating a plurality of adjacent point speed sequences based on each group of key element sequences in the actual key element data; determining an acceleration peak time point of each group of key element sequences based on a plurality of the adjacent point speed sequences; pairing each of the peak time points based on a preset key element name sequence; calculating a time difference of each pair of peak time points; calculating an average absolute deviation of each of the time differences; and determining the action coordination according to the average absolute deviation; The display device is configured to feed back the action evaluation result in real time.

13. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The computer program is executed by the processor to implement the steps of the action evaluation method according to any one of claims 1 to 10.

14. A computer-readable storage medium, the computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the action evaluation method according to any one of claims 1 to 10.

Images