A method for evaluating sports teaching actions

Abstract

The present application relates to the technical field of image recognition action, especially to a kind of motion teaching action evaluation method.The action frame is collected, after being cropped and normalized, to generate motion enhancement graph in three channels of adjacent frame difference, cross interval difference and window mean difference.Sharing network extracts features, and adaptively selects static or dynamic branch according to action type and fusion intensity of difference-optical flow, outputs results, and marks review when low confidence or occlusion.Grade deviation from template and individualized threshold;Sensitive band energy is calculated by frequency domain transformation for continuous deviation of same joint, and risk level is determined by combining risk rules.Frame rate and strategy are fed back and controlled by deviation level, risk level and confidence.Evaluation report is generated, only standard segment updates individualized threshold, and the rest is frozen for review, which is used as candidate for next cycle after teacher confirmation.The present application can reduce the pollution of individualized joint threshold of students by credible sample gating, version reservation and review release.

Description

Technical Field

[0001] This invention relates to the field of image recognition motion technology, and in particular to a method for evaluating motion in sports teaching. Background Technology

[0002] The physical education curriculum includes both static postures such as yoga, stretching, and balance, and dynamic movements such as aerobics, rhythm training, and continuous movement combinations. In multi-student classes, teachers need to observe each student's joint angles, limb extension range, rhythm, symmetry, and stability, and promptly provide feedback when movements may cause overstretching or repetitive loads. Relying solely on manual observation makes it difficult to continuously record subtle deviations in each student and to create verifiable teaching records.

[0003] Current visual motion assessment methods typically extract key human features or motion categories from videos and then compare them with standard movements. Motion history maps, optical flow, pose estimation, joint angles, stability, symmetry, and motion templates can all be used for motion recognition or quality assessment. However, continuously processing complete video sequences or running multiple independent networks simultaneously on ordinary teaching terminals increases computational overhead; fixed assessment thresholds are also difficult to adapt to students' body proportions, prior abilities, and training progress.

[0004] Furthermore, if high-risk actions, severely occluded segments, or low-confidence identification results are directly used to update individual action thresholds, erroneous actions may be gradually absorbed into individual standards. Simply providing risk scores without writing them back into the acquisition and identification strategy also fails to reduce computational load at low risk levels or restore analytical accuracy when risk increases. Therefore, a method is needed that connects temporal information compression, static and dynamic action branches, risk feedback control, and reliable sample threshold updates within the same technology chain. Summary of the Invention

[0005] This invention aims to at least solve one of the technical problems existing in related technologies. To this end, this invention provides a method for assessing movement in sports teaching, thereby reducing the contamination of students' personalized joint thresholds by incorrect movements.

[0006] This invention provides a method for assessing movement in sports teaching, comprising: S1: Collect student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, and write them into three color channels through the same normalization reference to form a motion enhancement image; S2: Input the motion-enhanced image into the shared feature network to obtain shared human features. Based on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, select the static pose recognition branch, dynamic action recognition branch, or both. S3: Output joint angles, symmetry, stability, or movement speed, trajectory, and rhythm from the selected branch. If the confidence level or occlusion does not meet the threshold, it is marked as a verification segment. S4: Generate multiple types of deviations by comparing with course templates and personalized joint thresholds, and classify them by amplitude, duration and number of repetitions; S5: Arrange the continuous deviations of the same joint into a time sequence, calculate the spectrum energy of the sensitive frequency band through frequency domain transformation, and output the teaching risk level in combination with risk rules; S6: Control the frame rate and recognition strategy using deviation level, risk level and confidence level feedback; S7: Generate teaching evaluation reports; only update personalized joint thresholds for segments that meet the standards in terms of completion, risk, and confidence, and freeze the rest for review. Once the teacher confirms that the segments are qualified, they will be used as candidate samples for the next cycle.

[0007] According to the present invention, a method for evaluating movement in sports teaching includes step S1, which includes: the difference between adjacent frames represents short-term limb changes, the difference across preset frame intervals represents cross-frame movement changes, and the difference between the current frame and the window mean frame represents slow posture drift within the window; the preset frame interval and preset window are determined by conversion based on the movement cycle of the course movement calibration segment and the current acquisition frame rate, and are recalculated according to the actual time length after frame rate switching; the three difference images are written to the red, green, and blue color channels respectively, using the same size and normalized upper and lower bounds.

[0008] According to a method for evaluating movement in sports teaching provided by the present invention, step S2 includes: the shared feature network includes a basic convolutional layer, a multi-scale attention module, a horizontal strip convolution, a vertical strip convolution, and a channel reweighting structure; the horizontal strip convolution is used to enhance the shoulder, hip, and arm extension edges, and the vertical strip convolution is used to enhance the directional contours of the trunk, lower limbs, and spine; the channel reweighting structure adjusts the shared features using the local response of the human body key point heatmap; when yoga, stretching, or balance-type movements have a normalized temporal change intensity lower than the branch threshold, a static posture recognition branch is selected; when aerobics, rhythm training, or continuous movement combinations, or when the normalized temporal change intensity is not lower than the branch threshold, a dynamic movement recognition branch is selected; when the course movement template simultaneously includes a posture holding phase and a rhythmic movement phase, both recognition branches are selected simultaneously; the branch threshold is determined by the normalized temporal change intensity distribution of the static holding phase and the dynamic phase in the course movement calibration segment, and when it is within the threshold tolerance band, both branches are invoked simultaneously.

[0009] According to the method for evaluating movement in sports teaching provided by the present invention, step S3 includes: the shared feature network, the static posture recognition branch and the dynamic movement recognition branch are trained by teaching movement samples labeled with key points, movement types and course stages, and the recognition confidence threshold and occlusion threshold are determined by independent calibration samples; when the two recognition branches output conflicting outputs for the movement type or movement stage of the same course stage, the course stage is marked as a verification segment, and no deterministic movement deviation is output and it is not used to update the personalized joint threshold.

[0010] According to the method for evaluating movement in sports teaching provided by the present invention, step S4 includes: personalized joint thresholds for students, including the maximum safe angle, minimum extension angle, bilateral symmetry tolerance, movement holding time tolerance, and rhythm tolerance for each joint; during the first training session, the thresholds are jointly established by the course movement template and the baseline movement segments confirmed by the teacher and whose identification confidence meets the requirements; if the baseline segments are insufficient, the conservative boundary of the course template is used and the thresholds are marked as pending confirmation; during subsequent training sessions, the previous valid version is read.

[0011] According to the method for evaluating movement in sports teaching provided by the present invention, step S5 includes: the sensitive frequency band includes a repeatability deviation frequency band determined based on the repetition frequency of movement calibration segments in the course, a rapid change frequency band determined based on the allowable time scale of rapid stretching movements, and a low-frequency frequency band used to characterize posture maintenance drift; the upper limit of each sensitive frequency band does not exceed half of the corresponding sampling frequency; the frequency band energy, together with the risk threshold, deviation amplitude, and duration conditions of the corresponding joint, jointly determine the risk level.

[0012] According to the method for evaluating movement in sports teaching provided by the present invention, step S6 includes: the energy-saving frame rate, the standard frame rate, and the full frame rate are determined by the frame rate level supported by the acquisition device; when switching the frame rate, the differential window is reset, and the sampling interval of the preset frame interval, the preset window, and the joint deviation timing is recalculated according to the actual time length before and after the switch; when the frame timestamp is interrupted, the current window is cleared and re-accumulated.

[0013] According to the method for evaluating movement in sports teaching provided by the present invention, step S7 includes: updating the student's personalized joint threshold using a restricted index, wherein the update coefficient is greater than zero and not greater than the upper limit of the preset update coefficient determined by the repeated measurement stability of the baseline movement segment; the updated threshold is constrained by the safety boundary of the course template and the maximum safety angle or risk limit shall not be relaxed due to the individual's historical movement; when the continuous review results show that the student's movement ability changes beyond the effective range of the individual baseline, the incremental update is stopped, the previous effective version is retained, and the baseline movement segment confirmed by the teacher is re-collected.

[0014] This invention also provides a motion assessment system for sports teaching, comprising: Enhanced image generation module: used to capture student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, which are written into three color channels through the same normalization reference to form motion enhancement images; Shared feature branch module: used to input motion-enhanced images into the shared feature network to obtain shared human features. Depending on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, it selects the static pose recognition branch, dynamic action recognition branch, or both. Posture and motion recognition module: used to output joint angles, symmetry, stability or motion speed, trajectory and rhythm from the selected branch, and to mark the segment as a verification segment when the confidence or occlusion does not meet the threshold; The motion deviation quantification module is used to generate multiple types of deviations by comparing them with course templates and personalized joint thresholds, and classifying them by amplitude, duration and number of repetitions. The sports risk assessment module is used to arrange continuous deviations of the same joint into a time sequence, calculate the spectral energy of sensitive frequency bands through frequency domain transformation, and output the teaching risk level in combination with risk rules. The graded teaching control module is used to control the frame rate and recognition strategy based on deviation level, risk level, and confidence level feedback. Teaching report update module: used to generate teaching evaluation reports. Only segments that meet the completion, risk and confidence standards are updated with personalized joint thresholds, while the rest are frozen for review. Once the teacher confirms that the segment is qualified, it will be used as a candidate sample for the next cycle.

[0015] The present invention also provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor is used to execute a computer program to implement the steps of a motion assessment method for teaching sports as described above.

[0016] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: This invention provides a method for evaluating movement in sports teaching. It compresses movement changes at different time scales into a single image input using three-channel motion enhancement images, and the two recognition branches share front-end human features, which helps reduce repetitive calculations on ordinary teaching terminals. By adjusting the acquisition and recognition strategy in reverse according to the risk results, it enables the use of lower resource allocation in the low-risk stage and restores the analysis capability in the high-risk stage. By using trusted sample gating, version retention, and release after review, it can reduce the contamination of students' personalized joint thresholds by erroneous movements.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for evaluating movement in sports teaching provided by the present invention.

[0020] Figure 2 This is a schematic diagram of the structure of a sports teaching movement assessment system provided by the present invention.

[0021] Figure 3 This is a schematic diagram of the electronic device provided by the present invention.

[0022] Figure label: 101. Enhanced Image Generation Module; 102. Shared Feature Branch Module; 103. Posture and Action Recognition Module; 104. Action Deviation Quantification Module; 105. Motion Risk Assessment Module; 106. Tiered Teaching Control Module; 107. Teaching Report Update Module; 810. Processor; 820. Communication Interface; 830. Memory; 840. Communication Bus. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.

[0024] In the description of the embodiments of the present invention, it should be noted that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention. In addition, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0025] In the description of the embodiments of the present invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of the present invention based on the specific circumstances.

[0026] In embodiments of the present invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0027] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0028] The following is combined Figures 1 to 3 This invention is described.

[0029] Example Figure 1 A flowchart illustrating a method for assessing movement in sports teaching provided by the present invention includes: S1: Collect student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, and write them into three color channels through the same normalization reference to form a motion enhancement image; S2: Input the motion-enhanced image into the shared feature network to obtain shared human features. Based on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, select the static pose recognition branch, dynamic action recognition branch, or both. S3: Output joint angles, symmetry, stability, or movement speed, trajectory, and rhythm from the selected branch. If the confidence level or occlusion does not meet the threshold, it is marked as a verification segment. S4: Generate multiple types of deviations by comparing with course templates and personalized joint thresholds, and classify them by amplitude, duration and number of repetitions; S5: Arrange the continuous deviations of the same joint into a time sequence, calculate the spectrum energy of the sensitive frequency band through frequency domain transformation, and output the teaching risk level in combination with risk rules; S6: Control the frame rate and recognition strategy using deviation level, risk level and confidence level feedback; S7: Generate teaching evaluation reports; only update personalized joint thresholds for segments that meet the standards in terms of completion, risk, and confidence, and freeze the rest for review. Once the teacher confirms that the segments are qualified, they will be used as candidate samples for the next cycle.

[0030] Before the initial formal calculation, the terminal establishes three types of reference bases. The first type is the course movement template, formed from standard movement videos confirmed by the teacher. The template saves standard joint angles, range of motion, torso tilt, symmetry tolerance, rhythm, and hold time according to the course movement, movement stage, and key joints. The second type is the student's personal baseline. During the initial training, at least one baseline movement segment confirmed by the teacher and meeting the required confidence level is collected. The key point coordinates are normalized according to body proportions to form the student's personal safety angle and tolerance. If the baseline segment is insufficient, severely occluded, or the movement is unqualified, only the conservative boundaries of the course template are used, and the personal threshold is marked as pending confirmation. The third type is the equipment and course configuration, including the frame rate levels supported by the acquisition equipment, course movement calibration segments, preset frame intervals, preset windows, branch thresholds, confidence thresholds, occlusion thresholds, risk rules, and update coefficient upper limits. The preset frame interval and window are converted from the movement cycle of the course movement calibration segments to the number of frames at the current frame rate. The conversion is recalculated after frame rate changes; old frame numbers are not directly reused.

[0031] Specifically, step S1 includes: the adjacent frame difference represents short-term limb changes, the cross-preset frame interval difference represents cross-frame action changes, and the difference between the current frame and the window mean frame represents slow posture drift within the window; the preset frame interval and preset window are determined based on the action cycle of the course action calibration segment and the current acquisition frame rate, and are recalculated according to the actual time length after frame rate switching; the three difference images use the same size and normalized upper and lower bounds, and are written into the red, green, and blue color channels respectively, using the following formula: in, For the current three-channel motion enhancement image, To normalize the difference image between adjacent frames, To normalize the difference image across a preset frame interval, For a normalized window mean difference image, , and The values ​​are obtained by taking the absolute values ​​of the differences between consecutive image frames at adjacent frames, across preset frame intervals, and between the current frame and the preset window mean frame, and then normalizing them using the same camera quantization range.

[0032] Specifically, step S2 includes: the shared feature network includes a base convolutional layer, a multi-scale attention module, a horizontal strip convolution, a vertical strip convolution, and a channel reweighting structure; the horizontal strip convolution is used to enhance the shoulder, hip, and arm extension edges, and the vertical strip convolution is used to enhance the trunk, lower limb, and spinal contours; the channel reweighting structure uses the local response of the human key point heatmap to adjust the shared features, selecting the static posture recognition branch when yoga, stretching, or balance movements have a normalized temporal change intensity lower than the branch threshold; selecting the dynamic movement recognition branch when aerobics, rhythm training, or continuous movement combinations, or when the normalized temporal change intensity is not lower than the branch threshold; when the course movement template contains both a posture holding phase and a rhythmic movement phase, both recognition branches are selected simultaneously; the branch threshold is determined by the normalized temporal change intensity distribution of the static holding phase and the dynamic phase in the course movement calibration segment, and both branches are called simultaneously when it is within the threshold tolerance band.

[0033] Specifically, step S3 includes: the shared feature network, static pose recognition branch and dynamic action recognition branch are trained by teaching action samples labeled with key points, action types and course stages, and the recognition confidence threshold and occlusion threshold are determined by independent calibration samples; when the two recognition branches output conflicting action types or action stages for the same course stage, the course stage is marked as a verification segment, no deterministic action deviation is output and it is not used to update the personalized joint threshold.

[0034] Specifically, step S4 includes: personalized joint thresholds for students, including the maximum safe angle, minimum extension angle, bilateral symmetry tolerance, movement holding time tolerance, and rhythm tolerance for each joint; during the first training session, the thresholds are jointly established by the course movement template and the baseline movement segments confirmed by the teacher and meeting the recognition confidence requirements. If the baseline segments are insufficient, the conservative boundary of the course template is used and the thresholds are marked as pending confirmation; during subsequent training sessions, the previous valid version is read.

[0035] Specifically, step S5 includes: the sensitive frequency bands include a repeatability deviation frequency band determined based on the repetition frequency of the course movement calibration segment, a rapid change frequency band determined based on the allowable time scale of the rapid pulling movement, and a low-frequency frequency band used to characterize posture maintenance drift; the upper limit of each sensitive frequency band does not exceed half of the corresponding sampling frequency; the frequency band energy, the risk threshold of the corresponding joint, the deviation amplitude, and the duration condition jointly determine the risk level; the teaching risk level and risk warning are only used for teaching safety reminders and movement adjustment suggestions, and are not considered as medical diagnostic conclusions.

[0036] Specifically, step S6 includes: the energy-saving frame rate, standard frame rate, and full frame rate below the standard frame rate are determined by the frame rate level supported by the acquisition device; when switching the frame rate, the differential window is reset, and the sampling interval of the preset frame interval, preset window, and joint deviation timing is recalculated according to the actual time length before and after the switch; when the frame timestamp is interrupted, the current window is cleared and re-accumulated.

[0037] Specifically, step S7 includes: updating the student's personalized joint threshold using a restricted index, with the update coefficient being greater than zero and not greater than the upper limit of the preset update coefficient calibrated by the repeated measurement stability of the baseline movement segment; the updated threshold is constrained by the safety boundary of the course template and the maximum safety angle or risk limit shall not be relaxed due to the individual's historical movement; when the continuous review results show that the student's movement ability changes beyond the effective range of the individual baseline, the incremental update is stopped, the previous effective version is retained, and the baseline movement segment confirmed by the teacher is re-collected.

[0038] Specifically, the calculation formula is as follows: in, For the j-th updated student-specific joint threshold; This is the previous valid threshold for this item; The candidate thresholds are obtained from the statistics of reliable samples, and all three use the same unit for the corresponding joint parameters; These are dimensionless update coefficients.

[0039] The dispersion is calibrated based on repeated measurements of the same baseline action, and is greater than zero but not greater than the preset upper limit of the update coefficient; The lower bound is the minimum permissible boundary specified by the course template, and the upper bound is the maximum safe boundary specified by the course template. The upper and lower bounds are determined by the course action template and course safety rules confirmed by the teacher.

[0040] like Figure 2 As shown below, a motion assessment system for sports teaching provided by the present invention is described, including the following modules: Enhanced image generation module 101: used to acquire student action frames, crop and normalize the human body region to generate adjacent frame differences, cross-preset frame interval differences and current frame and window mean frame differences, and write them into three color channels through the same normalization reference to form motion enhancement images; Shared feature branch module 102: used to input motion-enhanced images into the shared feature network to obtain shared human features, and select static pose recognition branch, dynamic action recognition branch or both according to the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude. Posture and motion recognition module 103: used to output joint angles, symmetry, stability or motion speed, trajectory and rhythm from the selected branch, and to mark the segment as a verification segment when the confidence or occlusion does not meet the threshold. Motion Deviation Quantification Module 104: Used to generate multiple types of deviations by comparing with course templates and personalized joint thresholds, and classifying them by amplitude, duration and number of repetitions; Sports Risk Assessment Module 105: This module arranges continuous deviations of the same joint into a time sequence, calculates the spectral energy of sensitive frequency bands through frequency domain transformation, and outputs the teaching risk level in combination with risk rules. Tiered teaching control module 106: used to control frame rate and recognition strategy with feedback based on deviation level, risk level and confidence level; Teaching report update module 107: Used to generate teaching evaluation reports. Only segments that meet the completion, risk, and confidence standards are updated with personalized joint thresholds. The rest are frozen for review. Once the teacher confirms that the segment is qualified, it will be used as a candidate sample for the next cycle.

[0041] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a motion teaching action evaluation method, which includes: S1: Collect student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, and write them into three color channels through the same normalization reference to form a motion enhancement image; S2: Input the motion-enhanced image into the shared feature network to obtain shared human features. Based on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, select the static pose recognition branch, dynamic action recognition branch, or both. S3: Output joint angles, symmetry, stability, or movement speed, trajectory, and rhythm from the selected branch. If the confidence level or occlusion does not meet the threshold, it is marked as a verification segment. S4: Generate multiple types of deviations by comparing with course templates and personalized joint thresholds, and classify them by amplitude, duration and number of repetitions; S5: Arrange the continuous deviations of the same joint into a time sequence, calculate the spectrum energy of the sensitive frequency band through frequency domain transformation, and output the teaching risk level in combination with risk rules; S6: Control the frame rate and recognition strategy using deviation level, risk level and confidence level feedback; S7: Generate teaching evaluation reports; only update personalized joint thresholds for segments that meet the standards in terms of completion, risk, and confidence, and freeze the rest for review. Once the teacher confirms that the segments are qualified, they will be used as candidate samples for the next cycle.

[0042] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0043] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0044] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0045] 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 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; and these 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.

[0046] It should be noted that the embodiments of this disclosure can be implemented using hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a programmable memory or a data carrier such as an optical or electronic signal carrier.

[0047] While this disclosure has been described with reference to several specific embodiments, it should be understood that this disclosure is not limited to the specific embodiments disclosed. This disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for evaluating movement in sports teaching, characterized in that, include: S1: Collect student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, and write them into three color channels through the same normalization reference to form a motion enhancement image; S2: Input the motion-enhanced image into the shared feature network to obtain shared human features. Based on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, select the static pose recognition branch, dynamic action recognition branch, or both. S3: Output joint angles, symmetry, stability, or movement speed, trajectory, and rhythm from the selected branch. If the confidence level or occlusion does not meet the threshold, it is marked as a verification segment. S4: Generate multiple types of deviations by comparing with course templates and personalized joint thresholds, and classify them by amplitude, duration and number of repetitions; S5: Arrange the continuous deviations of the same joint into a time sequence, calculate the spectrum energy of the sensitive frequency band through frequency domain transformation, and output the teaching risk level in combination with risk rules; S6: Control the frame rate and recognition strategy using deviation level, risk level and confidence level feedback; S7: Generate teaching evaluation reports; only update personalized joint thresholds for segments that meet the standards in terms of completion, risk, and confidence, and freeze the rest for review. Once the teacher confirms that the segments are qualified, they will be used as candidate samples for the next cycle.

2. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S1 includes: the adjacent frame difference represents short-term limb changes, the cross-preset frame interval difference represents cross-frame action changes, and the difference between the current frame and the window mean frame represents slow posture drift within the window; the preset frame interval and preset window are determined based on the action cycle of the course action calibration segment and the current acquisition frame rate, and are recalculated according to the actual time length after the frame rate is switched; the three difference images use the same size and normalized upper and lower bounds, and are written into the red, green, and blue color channels respectively.

3. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S2 includes: the shared feature network includes a base convolutional layer, a multi-scale attention module, a horizontal strip convolution, a vertical strip convolution, and a channel reweighting structure; the horizontal strip convolution is used to enhance the shoulder, hip, and arm extension edges, and the vertical strip convolution is used to enhance the directional contours of the trunk, lower limbs, and spine; the channel reweighting structure uses the local response of the human key point heatmap to adjust the shared features. When yoga, stretching, or balance-type movements have a normalized temporal change intensity lower than the branch threshold, a static posture recognition branch is selected; when aerobics, rhythm training, or continuous movement combinations have a normalized temporal change intensity not lower than the branch threshold, a dynamic movement recognition branch is selected; when the course movement template contains both a posture holding phase and a rhythmic movement phase, both recognition branches are selected simultaneously; the branch threshold is determined by the normalized temporal change intensity distribution of the static holding phase and the dynamic phase in the course movement calibration segment, and both branches are called simultaneously when the threshold is within the threshold tolerance band.

4. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S3 includes: the shared feature network, static pose recognition branch and dynamic action recognition branch are trained by teaching action samples labeled with key points, action types and course stages, and the recognition confidence threshold and occlusion threshold are determined by independent calibration samples; when the two recognition branches output conflicting action types or action stages for the same course stage, the course stage is marked as a verification segment, no deterministic action deviation is output and it is not used to update the personalized joint threshold.

5. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S4 includes: personalized joint thresholds for students, including the maximum safe angle, minimum extension angle, bilateral symmetry tolerance, movement holding time tolerance, and rhythm tolerance for each joint; during the first training session, the thresholds are jointly established by the course movement template and the baseline movement segments confirmed by the teacher and meeting the recognition confidence requirements. If the baseline segments are insufficient, the conservative boundary of the course template is used and the thresholds are marked as pending confirmation; during subsequent training sessions, the previous valid version is read.

6. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S5 includes: the sensitive frequency bands include a repeatability deviation frequency band determined based on the repetition frequency of the course action calibration segment, a rapid change frequency band determined based on the allowable time scale of the rapid traction action, and a low-frequency frequency band used to characterize attitude maintenance drift; the upper limit of each sensitive frequency band does not exceed half of the corresponding sampling frequency; the frequency band energy, the risk threshold of the corresponding joint, the deviation amplitude, and the duration condition jointly determine the risk level.

7. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S6 includes: the energy-saving frame rate, standard frame rate, and full frame rate are determined by the frame rate level supported by the acquisition device; when switching frame rates, the differential window is reset, and the sampling interval of the preset frame interval, preset window, and joint deviation timing is recalculated according to the actual time length before and after the switch; when the frame timestamp is interrupted, the current window is cleared and re-accumulated.

8. The method for evaluating movement in sports teaching according to claim 1, characterized in that, Step S7 includes: updating the student's personalized joint threshold using a restricted index, with the update coefficient being greater than zero and not greater than the upper limit of the preset update coefficient calibrated by the repeated measurement stability of the baseline movement segment; the updated threshold is constrained by the safety boundary of the course template and the maximum safety angle or risk limit shall not be relaxed due to the individual's historical movement; when the continuous review results show that the student's movement ability changes beyond the effective range of the individual baseline, the incremental update is stopped, the previous effective version is retained, and the baseline movement segment confirmed by the teacher is re-collected.

9. A motion assessment system for sports teaching, characterized in that, include: Enhanced image generation module: used to capture student action frames, crop and normalize the human body region, and generate adjacent frame differences, cross-preset frame interval differences, and current frame and window mean frame differences, which are written into three color channels through the same normalization reference to form motion enhancement images; Shared feature branch module: used to input motion-enhanced images into the shared feature network to obtain shared human features. Depending on the course action type and the normalized temporal change intensity determined by the fusion of differential amplitude and optical flow amplitude, it selects the static pose recognition branch, dynamic action recognition branch, or both. Posture and motion recognition module: used to output joint angles, symmetry, stability or motion speed, trajectory and rhythm from the selected branch, and to mark the segment as a verification segment when the confidence or occlusion does not meet the threshold; The motion deviation quantification module is used to generate multiple types of deviations by comparing them with course templates and personalized joint thresholds, and classifying them by amplitude, duration and number of repetitions. The sports risk assessment module is used to arrange continuous deviations of the same joint into a time sequence, calculate the spectral energy of sensitive frequency bands through frequency domain transformation, and output the teaching risk level in combination with risk rules. The graded teaching control module is used to control the frame rate and recognition strategy based on deviation level, risk level, and confidence level feedback. Teaching report update module: used to generate teaching evaluation reports. Only segments that meet the completion, risk and confidence standards are updated with personalized joint thresholds, while the rest are frozen for review. Once the teacher confirms that the segment is qualified, it will be used as a candidate sample for the next cycle.

10. An electronic device comprising a processor, a communication interface, a memory, and a communication bus, characterized in that, When the processor executes the computer program, it implements the steps of the motion assessment method for sports teaching as described in any one of claims 1 to 8.

Images