An interactive sports teaching system and method based on video image processing

Abstract

The present invention belongs to the field of teaching video processing technology, and specifically relates to an interactive physical education teaching system and method based on video image processing. The invention can capture the user's training movements in real time through an interactive physical education teaching method based on video image processing, and compare and analyze with standard movements, so as to accurately identify correct movements and deviation movements. Through the analysis of key nodes, continuous correct periods and continuous deviation periods can be determined, and then the deviation movements can be analyzed in an orderly manner, and independent deviations and progressive deviations can be distinguished. According to the type of deviation movement, targeted correction suggestions can be provided to help users improve their movements and improve training effects. In addition, real-time feedback can be provided on the user's training movements according to the improvement amount, so that the user can understand his or her training effect in time and make adjustments according to the feedback. Through continuous correction analysis and feedback, the user can gradually improve the quality of the movement and achieve better training effects.

Description

Technical Field

[0001] The invention belongs to the technical field of teaching video processing, and in particular relates to an interactive sports teaching system and method based on video image processing. Background Art

[0002] With the rapid development of computer technology and artificial intelligence, traditional sports teaching methods can no longer meet the needs of modern society, especially in personalized teaching and real-time feedback. Traditional methods have many limitations. Therefore, an interactive sports teaching system based on video image processing came into being, aiming to provide users with more efficient and personalized sports teaching services through advanced image recognition and analysis technology.

[0003] However, existing interactive physical education teaching systems based on video image processing mostly focus on the accuracy of user's body movements and provide feedback through this accuracy, but often ignore the continuity and fluency of the movements, and are unable to find the causes of low-accuracy movements, which requires users to spend a lot of time exploring and correcting errors, which obviously limits the improvement of teaching effectiveness. Based on this, this solution provides an interactive physical education teaching method based on video image processing to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide an interactive physical education teaching system and method based on video image processing, which can carefully classify movements with training deviations, thereby providing users with more accurate correction suggestions and real-time feedback to improve teaching effects.

[0005] The technical solution adopted by the present invention is as follows:

[0006] An interactive sports teaching method based on video image processing, comprising:

[0007] Perform feature extraction on the acquired physical education teaching video to obtain multiple standard action frames;

[0008] Capture the user's training action frame in real time, and compare and analyze the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame;

[0009] Collecting the intersection nodes between the correct action frame and the deviation action frame, marking the intersection nodes as key nodes, and determining the continuous correct time period and the continuous deviation time period according to the key nodes;

[0010] Performing orderliness analysis on the correct action frames within the continuous correct period to obtain the deviation type of the deviation action frame after the end node of the continuous correct period;

[0011] Performing correction analysis on the deviation action frames within the continuous deviation period, and outputting correction suggestions for each of the deviation action frames according to the results of the correction analysis;

[0012] The improvement amount of each of the deviation action frames under the correction suggestion is collected, and real-time feedback is given to the user's training action based on the improvement amount, and the correction suggestion of each of the deviation action frames is updated.

[0013] In a preferred embodiment, the step of extracting features from the acquired physical education teaching video to obtain a plurality of standard action frames includes:

[0014] Acquire a physical education teaching video, and preprocess the physical education teaching video, wherein the preprocessing includes denoising, contrast enhancement, and brightness adjustment;

[0015] Decode the preprocessed physical education teaching video and extract the video frame sequence;

[0016] Key action features are obtained, and key frames of a video frame sequence are screened according to the key action features to obtain a plurality of standard action frames.

[0017] In a preferred embodiment, the step of comparing and analyzing the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame includes:

[0018] Capturing the user's current training action sequence and the training action frames under the training action sequence;

[0019] Adding time stamps and space marks to the training action frames and standard action frames;

[0020] Performing time synchronization and spatial alignment on the training action frame and the standard action frame, and performing comparative analysis on the training action frame and the standard action frame to obtain an action matching degree between the training action frame and the standard action frame;

[0021] The training action frames are screened according to the action matching degree, and the screened training action frames are classified into correct action frames and deviation action frames.

[0022] In a preferred embodiment, the step of comparing and analyzing the training action frame with the standard action frame to obtain the action matching degree between the training action frame and the standard action frame includes:

[0023] Acquire feature points of the training action frame and the standard action frame, and summarize them into a training action feature vector corresponding to the training action frame and a standard action feature vector corresponding to the standard action frame;

[0024] Obtaining a matching degree calculation function, and inputting the training action feature vector and the standard action feature vector into the matching degree calculation function, and calibrating the output result of the matching degree calculation function as the action matching degree;

[0025] After the action matching degree is output, the screening parameters are obtained;

[0026] comparing the screening parameters with the action matching degree;

[0027] When the screening parameter is greater than or equal to the action matching degree, the corresponding training action frame is marked as a correct action frame;

[0028] When the screening parameter is less than the action matching degree, the corresponding training action frame is marked as a deviation action frame.

[0029] In a preferred solution, the step of determining the continuous correct period and the continuous deviation period according to the key node includes:

[0030] Acquire time series information of the key nodes, and sort the key nodes according to the time series information;

[0031] Counting the time lengths between adjacent key nodes and recording them as the interval length to be evaluated;

[0032] Obtaining a standard interval duration, and comparing the standard interval duration with the interval duration to be evaluated;

[0033] If the duration of the interval to be evaluated is less than or equal to the standard interval duration, the duration of the interval to be evaluated is recorded as an occasional period;

[0034] If the duration of the interval to be evaluated is greater than the standard duration, the duration of the interval to be evaluated in which the training action frame is a correct action frame is recorded as a continuous correct period; otherwise, it is recorded as a continuous deviation period.

[0035] In a preferred solution, the step of performing orderliness analysis on the correct action frames within the continuous correct period to obtain the deviation type of the deviation action frame after the end node of the continuous correct period includes:

[0036] Acquire the correct action frames and standard action frames within the continuous correct time period, and arrange them according to the occurrence sequence;

[0037] Collecting the action differences between the correct action frame and the standard action frame one by one, and recording them as the differences to be evaluated;

[0038] Obtaining a verification function, inputting the difference to be evaluated into the verification function, and calibrating the output result of the verification function as a first verification condition parameter;

[0039] Obtaining the action difference between the deviation action frame after the end node of the continuous correct period and the corresponding standard action frame, and recording it as the second verification condition parameter;

[0040] Performing a difference processing on the first verification condition parameter and the second verification condition parameter, and recording the difference result of the first verification condition parameter and the second verification condition parameter as a classification condition parameter;

[0041] Obtain a classification condition threshold, and compare the classification condition threshold with a classification condition parameter, and when the classification condition parameter is greater than or equal to the classification condition threshold, output the deviation type of the deviation action frame as an independent deviation, otherwise, output the deviation type of the deviation action frame as a progressive deviation.

[0042] In a preferred solution, under the progressive deviation, the deviation action frame is traced to determine the starting node of the deviation action frame, and according to the standard action frame, the correct action frame after the starting node in the continuous correct period is corrected and analyzed, and correction suggestions for each of the deviation action frames are output according to the results of the correction analysis;

[0043] Under the independent deviation, correction analysis is directly performed on the deviation action frames within the continuous deviation period, and correction suggestions for each of the deviation action frames are output based on the results of the correction analysis.

[0044] In a preferred embodiment, the step of providing real-time feedback on the user's training action according to the improvement amount includes:

[0045] After obtaining the correction suggestion output, the user's correction action frame;

[0046] Performing time synchronization and spatial alignment on the correction action frame and the deviation action frame, and performing comparative analysis on the correction action frame and the deviation action frame to obtain an improvement amount of the deviation action frame;

[0047] Obtaining a standard correction amount under a correction suggestion corresponding to each of the deviation action frames, and performing an offset process on the standard correction amount to obtain a valid correction interval;

[0048] comparing the effective correction interval with the improvement amount;

[0049] When the improvement amount is within the effective correction interval, outputting real-time feedback that the user's training action is improved;

[0050] When the improvement amount is not within the effective correction interval, real-time feedback is outputted indicating that the user's training action has not been improved.

[0051] The present invention also provides an interactive physical education teaching system based on video image processing, using the above-mentioned interactive physical education teaching method based on video image processing, comprising:

[0052] A feature extraction module, wherein the feature extraction module is used to extract features from the acquired physical education teaching video to obtain a plurality of standard action frames;

[0053] An action comparison module, which is used to capture the user's training action frame in real time, and compare and analyze the captured training action frame with the standard action frame to obtain a correct action frame and a deviation action frame;

[0054] A segmentation module, the segmentation module is used to collect the intersection nodes between the correct action frame and the deviation action frame, mark the intersection nodes as key nodes, and determine the continuous correct time period and the continuous deviation time period according to the key nodes;

[0055] A deviation type output module, the deviation type output module is used to perform orderliness analysis on the correct action frames within the continuous correct period, and obtain the deviation type of the deviation action frame after the end node of the continuous correct period;

[0056] A correction analysis module, the correction analysis module is used to perform correction analysis on the deviation action frames within the continuous deviation period, and output correction suggestions for each of the deviation action frames according to the results of the correction analysis;

[0057] A correction feedback module, the correction feedback module is used to collect the improvement amount of each of the deviation action frames under the correction suggestion, and provide real-time feedback on the user's training action based on the improvement amount, and update the correction suggestion of each of the deviation action frames.

[0058] And, an electronic device, the electronic device comprising:

[0059] at least one processor;

[0060] and a memory communicatively coupled to the at least one processor;

[0061] Wherein, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned interactive sports teaching method based on video image processing.

[0062] The technical effects achieved by the present invention are:

[0063] The present invention uses an interactive sports teaching method based on video image processing to capture the user's training movements in real time, and compare and analyze them with standard movements, so as to accurately identify correct movements and deviant movements. Through the analysis of key nodes, continuous correct time periods and continuous deviant time periods can be determined, and then the deviant movements can be analyzed in an orderly manner to distinguish independent deviations from progressive deviations. According to the type of deviant movement, targeted correction suggestions can be provided to help users improve their movements and improve training effects. In addition, real-time feedback can be provided on the user's training movements according to the improvement amount, so that the user can understand the training effect in time and make adjustments based on the feedback. Through continuous correction analysis and feedback, the user can gradually improve the quality of the movement and achieve better training effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic flow chart of the method of the present invention;

[0065] Figure 2 It is a schematic diagram of the system module of the present invention;

[0066] Figure 3 It is a schematic diagram of the structure of an electronic device of the present invention. DETAILED DESCRIPTION

[0067] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0068] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0069] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" that appears in different places in this specification does not refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0070] See also Figure 1 As shown, the present invention provides an interactive sports teaching method based on video image processing, comprising:

[0071] S1, extracting features from the acquired physical education teaching video to obtain multiple standard action frames;

[0072] S2, capturing the user's training action frame in real time, and comparing and analyzing the captured training action frame with the standard action frame to obtain a correct action frame and a deviation action frame;

[0073] S3, collecting the intersection nodes between the correct action frame and the deviation action frame, marking the intersection nodes as key nodes, and determining the continuous correct period and the continuous deviation period based on the key nodes;

[0074] S4, performing orderliness analysis on the correct action frames within the continuous correct period, and obtaining the deviation type of the deviation action frame after the end node of the continuous correct period;

[0075] S5, performing correction analysis on the deviation action frames within the continuous deviation period, and outputting correction suggestions for each deviation action frame according to the results of the correction analysis;

[0076] S6. Collect the improvement amount of each deviation action frame under the correction suggestion, and provide real-time feedback on the user's training action based on the improvement amount, and update the correction suggestion of each deviation action frame.

[0077] As described in the above steps S1-S6, with the continuous advancement of image processing technology, its application scope is also expanding. In the field of physical education, the interactive teaching method based on video image processing can provide users with more intuitive and efficient training feedback, thereby improving the accuracy of users' movements in physical training. In this embodiment, it is first necessary to collect physical education teaching videos and extract standard action frames from them. The standard action frames will be used as a reference for subsequent comparison and analysis with the user's training movements. Specifically, by capturing the user's training action frames in real time, the captured training action frames can be compared and analyzed with the standard action frames to identify whether the action is correct or not. The correct action frame will be marked as a correct action frame, otherwise it will be recorded as a deviation action frame that requires further correction processing. By collecting the intersection nodes between the correct action frame and the deviation action frame, the key nodes between the correct action and the incorrect action can be determined. At the same time, the determination of the key nodes is also helpful to identify the connection. There are continuous correct periods and continuous deviation periods, and then the correct action frames in the continuous correct periods will be analyzed in an orderly manner, so as to obtain the deviation type of the deviation action frame after the end node of the continuous correct period, which is helpful to identify the specific errors of the user in the process of action execution, so as to provide more targeted correction suggestions. For the deviation action frames in the continuous deviation period, corresponding correction analysis will be carried out, and correction suggestions for each deviation action frame will be output based on the results of the correction analysis. The correction suggestions include but are not limited to specific information such as the adjustment direction, amplitude, speed, etc. of the action, so as to help users improve their actions. Finally, the improvement amount of each deviation action frame under the correction suggestion is collected, and the user's training action is fed back in real time based on the improvement amount, so as to ensure that the user can understand the improvement of the deviation action in time. At the same time, according to the user's improvement, the correction suggestions for each deviation action frame can be updated to form a continuously optimized training process, so as to gradually improve the user's action accuracy in sports training.

[0078] In a preferred embodiment, the step of extracting features from the acquired physical education teaching video to obtain a plurality of standard action frames includes:

[0079] S101, obtaining a physical education teaching video, and preprocessing the physical education teaching video, wherein the preprocessing includes denoising, contrast enhancement, and brightness adjustment;

[0080] S102, decoding the pre-processed physical education teaching video to extract a video frame sequence;

[0081] S103, acquiring key action features, and screening key frames of the video frame sequence according to the key action features to obtain a plurality of standard action frames.

[0082] As described in the above steps S101-S103, when extracting standard action frames from physical education teaching videos, it is first necessary to obtain the required physical education teaching video resources to ensure that there are sufficient materials for subsequent processing and analysis, and then pre-process the physical education teaching video to improve the accuracy and efficiency of subsequent processing. The pre-processing step includes removing noise from the video, which can be achieved through various filtering algorithms to ensure that the image in the video is clearer. In addition, it is necessary to enhance the contrast of the video to make the details of the action more obvious and easy to identify and analyze. Finally, adjust the brightness of the video to ensure that the characteristics of the action in the sports video can be clearly captured under different lighting conditions to complete the physical education teaching video. After preprocessing the teaching video, the video will be decoded and the video frame sequence will be extracted. Specifically, the compressed data in the video file will be decompressed and converted into a series of continuous image frames. The image data of each frame can be obtained through decoding, providing rich materials for further feature extraction. Then the key action features need to be identified and extracted. The key action features need to be set according to the user's needs, and based on the extracted key action features, the video frame sequence will be screened for key frames, that is, the most representative standard action frames will be screened from a large number of video frames. With the reference of standard action frames, learners can more intuitively understand the correct action postures and techniques, thereby improving learning efficiency and sports performance.

[0083] In a preferred embodiment, the steps of comparing and analyzing the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame include:

[0084] S201, capturing the user's current training action sequence and the training action frames under the training action sequence;

[0085] S202, adding time stamps and space marks to the training action frames and the standard action frames;

[0086] S203, performing time synchronization and spatial alignment on the training action frame and the standard action frame, and performing comparative analysis on the training action frame and the standard action frame to obtain the action matching degree between the training action frame and the standard action frame;

[0087] S204 , screening the training action frames according to the action matching degree, and classifying the screened training action frames into correct action frames and deviation action frames.

[0088] As described in the above steps S201-S204, in order to ensure the accuracy and effectiveness of the training action, it is necessary to perform corresponding comparative analysis on the captured training action frames and the standard action frames. First, the user's current training action sequence is captured, and each training action frame is extracted therefrom. Then, the captured training action frames and the standard action frames are marked, and timestamps and space marks are added. The timestamp is used to record the specific time point when each action frame occurs, and the space mark is used to record the position and direction of the action in space, so as to facilitate comparative analysis. Then, the training action frames and the standard action frames are synchronized in time and aligned in space. Time synchronization ensures that the actions at the same moment are compared, and space alignment ensures that the actions are in the same place. The position and direction in space are consistent, so that the difference between the training action frame and the standard action frame can be accurately compared and analyzed. After completing the time synchronization and space alignment, the training action frame and the standard action frame will be compared and analyzed in detail to obtain the action matching degree between the training action frame and the standard action frame. The action matching degree reflects the similarity between the training action and the standard action. Then the training action frames are screened according to the action matching degree. According to the screening results, the screened training action frames are classified into correct action frames and deviation action frames. The correct action frame indicates that the user's action is consistent or highly consistent with the standard action, while the deviation action frame indicates that the user's action has a certain deviation, which requires further correction and improvement.

[0089] In a preferred embodiment, the step of comparing and analyzing the training action frame and the standard action frame to obtain the action matching degree between the training action frame and the standard action frame includes:

[0090] Acquire feature points of the training action frame and the standard action frame, and summarize them into a training action feature vector corresponding to the training action frame and a standard action feature vector corresponding to the standard action frame;

[0091] Obtain a matching degree calculation function, input the training action feature vector and the standard action feature vector into the matching degree calculation function, and calibrate the output result of the matching degree calculation function as the action matching degree;

[0092] Among them, after the action matching degree is output, the screening parameters are obtained;

[0093] Compare the screening parameters with the action match;

[0094] When the screening parameter is greater than or equal to the action matching degree, the corresponding training action frame is marked as the correct action frame;

[0095] When the screening parameter is less than the action matching degree, the corresponding training action frame is marked as a deviation action frame.

[0096] In this embodiment, when calculating the action matching degree between the training action frame and the standard action frame, the feature points of the training action frame and the standard action frame are first obtained, and then the feature points are summarized to form a training action feature vector corresponding to the training action frame and a standard action feature vector corresponding to the standard action frame, so as to convert the action data into a digital representation, and then introduce a preset matching degree measurement function to quantitatively evaluate the matching degree between the two feature vectors, wherein the expression of the matching degree measurement function is: , where Indicates the action matching degree, Represents the number of feature points of training action frames and standard action frames, represents the training action feature vector, Represents the standard action feature vector. By inputting the training action feature vector and the standard action feature vector into the matching degree measurement function together, the action matching degree between the training action frame and the standard action frame can be calculated. After the action matching degree is calculated, a preset screening parameter will be introduced. The screening parameter is set according to specific requirements and standards, and is used to further screen and evaluate the quality of the action frame. By comparing the screening parameter with the action matching degree, it can be determined whether the training action frame has reached the expected standard. Specifically, if the screening parameter is greater than or equal to the action matching degree, it means that the similarity between the training action frame and the standard action frame is high, and the training action is performed relatively accurately. In this case, the corresponding training action frame will be calibrated as a correct action frame, indicating that the action is performed in line with expectations. On the contrary, if the screening parameter is less than the action matching degree, it means that there is a certain deviation between the training action frame and the standard action frame, and the training action is not performed accurately enough. In this case, the corresponding training action frame will be calibrated as a deviation action frame, indicating that the training action needs further improvement and practice to meet the requirements of the standard action.

[0097] In a preferred embodiment, the step of determining the continuous correct period and the continuous deviation period according to the key nodes includes:

[0098] Obtain the time series information of key nodes and sort the key nodes according to the time series information;

[0099] Count the time lengths between adjacent key nodes and record them as the interval length to be evaluated;

[0100] Obtain the standard interval duration, and compare the standard interval duration with the interval duration to be evaluated;

[0101] If the duration of the interval to be evaluated is less than or equal to the standard interval, the duration of the interval to be evaluated is recorded as an occasional period;

[0102] If the duration of the interval to be evaluated is greater than the standard interval duration, the duration of the interval to be evaluated when the training action frame is a correct action frame is recorded as a continuous correct period, otherwise, it is recorded as a continuous deviation period.

[0103] In this implementation, when determining the continuous correct time period and the continuous deviation time period, it is necessary to clarify the time series information of the key nodes, that is, the specific time points at which the key nodes appear. At the same time, the key nodes will be sorted according to these time series information. After the sorting is completed, the time lengths between adjacent key nodes are counted and recorded as the interval length to be evaluated. In order to determine whether the time length of the interval length to be evaluated is required, a preset standard interval length is introduced as a reference. The standard interval length is determined based on historical data or preset standards, and represents the time interval that should appear between adjacent key nodes under normal circumstances. Then the interval length to be evaluated is recorded. Compared with the standard interval duration, if the interval duration to be evaluated is less than or equal to the standard interval duration, it means that the time interval between adjacent key nodes is too short, and the interval duration to be evaluated is recorded as an occasional period. The correct action frames in the occasional period are directly saved, and the deviation action frames in the occasional period are directly recorded as independent deviations. If the interval duration to be evaluated is greater than the standard interval duration, the interval duration to be evaluated when the training action frame is a correct action frame will be recorded as a continuous correct period. On the contrary, if the training action frame is an incorrect action frame, it will be recorded as a continuous deviation period, so as to achieve a more detailed evaluation and classification of the training actions.

[0104] In a preferred embodiment, the step of performing orderliness analysis on the correct action frames within the continuous correct period to obtain the deviation type of the deviation action frame after the end node of the continuous correct period includes:

[0105] S401, obtaining correct action frames and standard action frames within a continuous correct period, and arranging them according to the occurrence sequence;

[0106] S402, collecting the action differences between the correct action frame and the standard action frame one by one, and recording them as the difference to be evaluated;

[0107] S403, obtaining a verification function, inputting the difference to be evaluated into the verification function, and calibrating the output result of the verification function as a first verification condition parameter;

[0108] S404, obtaining the action difference between the deviation action frame and the corresponding standard action frame after the end node of the continuous correct period, and recording it as the second verification condition parameter;

[0109] S405, performing a difference process on the first verification condition parameter and the second verification condition parameter, and recording the difference result of the first verification condition parameter and the second verification condition parameter as a classification condition parameter;

[0110] S406. Obtain a classification condition threshold, and compare the classification condition threshold with the classification condition parameter. When the classification condition parameter is greater than or equal to the classification condition threshold, output the deviation type of the deviation action frame as an independent deviation. Otherwise, output the deviation type of the deviation action frame as a progressive deviation.

[0111] As described in the above steps S401-S406, when analyzing the orderliness of the correct action frames in the continuous correct period, it is first necessary to collect all the correct action frames that appear in the continuous correct period, and the corresponding standardized action frames, and the correct action frames and the standard action frames will be arranged in the order of occurrence, and then the action differences between the correct action frames and the standard action frames are collected one by one to reflect the allowable deviation of the action execution, and recorded as the difference to be evaluated, and then the verification function is introduced, and the difference to be evaluated is used as input for calculation, wherein the expression of the verification function is: , where represents the first verification condition parameter, Indicates the action difference between the correct action frame and the standard action frame at the end node of the continuous correct period. Indicates the length of time for consecutive correct periods, Indicates the time length between the end node of the continuous correct period and the node where the adjacent deviation action frame occurs. represents the number of differences to be evaluated, It represents the difference to be evaluated. After the first verification condition parameter is output, the action difference between the deviation action frame and the corresponding standard action frame after the end node of the continuous correct time period will be collected and recorded as the second verification condition parameter. Then, the first verification condition parameter and the second verification condition parameter will be differenced, that is, the first verification condition parameter will be subtracted from the second verification condition parameter. The result will be recorded as the classification condition parameter, and then the classification condition threshold will be introduced. The classification condition threshold will be compared with the classification condition parameter. If the classification condition parameter is greater than or equal to the classification condition threshold, the deviation type of the deviation action frame will be determined to be an independent deviation. Otherwise, if the classification condition parameter is less than the classification condition threshold, the deviation type of the deviation action frame will be determined to be a progressive deviation.

[0112] In a preferred embodiment, under progressive deviation, the deviation action frame is traced to determine the starting node of the deviation action frame, and the correct action frame after the starting node in the continuous correct period is corrected and analyzed according to the standard action frame, and correction suggestions for each deviation action frame are output according to the results of the correction analysis;

[0113] Under independent deviation, the deviation action frames within the continuous deviation period are directly corrected and analyzed, and correction suggestions for each deviation action frame are output based on the results of the correction analysis.

[0114] As mentioned above, in the case of progressive deviation, it is necessary to perform detailed tracing of the deviation action frame, that is, it is necessary to track and determine the starting node of the deviation action frame, and find out the starting position that causes the deviation. Once the starting node is located, the correct action frames in the continuous correct time period after the starting node can be corrected and analyzed based on the standard action frame. The positioning of the starting node can be achieved by finding the first occurrence time point of the slight deviation within the corresponding continuous correct time period. Based on the results of the correction analysis, specific correction suggestions can be provided for each deviation action frame to help correct these deviations and ensure the accuracy of the action. In the case of independent deviation, the deviation action frames in the continuous deviation period can be directly corrected and analyzed without tracing back to the starting node.

[0115] In a preferred embodiment, the step of providing real-time feedback on the user's training action according to the improvement amount includes:

[0116] After obtaining the correction suggestion output, the user's correction action frame;

[0117] The correction action frame is synchronized in time and aligned in space with the deviation action frame, and the correction action frame and the deviation action frame are compared and analyzed to obtain the improvement amount of the deviation action frame;

[0118] Obtain the standard correction amount under the correction suggestion corresponding to each deviation action frame, and perform offset processing on the standard correction amount to obtain the effective correction interval;

[0119] Compare the effective correction interval with the improvement amount;

[0120] When the improvement amount is within the effective correction interval, real-time feedback that the user's training action has been improved is output;

[0121] When the improvement amount is not within the effective correction interval, real-time feedback is output that the user's training action has not been improved.

[0122] In this embodiment, in order to achieve real-time feedback on the user's training action, after the correction suggestion is output, the correction action frame of the user's continued training will be synchronously recorded, and then the correction action frame and the deviation action frame will be synchronized and spatially aligned, so that the improvement amount of the deviation action frame can be obtained. Then, the standard correction amount under the correction suggestion corresponding to each deviation action frame is obtained, and the standard correction amount is offset to obtain an effective correction interval. The effective correction interval refers to a reasonable range of user action improvement under a given correction suggestion. After that, the effective correction interval is compared with the improvement amount to determine whether the user's improvement amount is within the effective correction interval, thereby determining whether the user's training action has been effectively improved. When the improvement amount is within the effective correction interval, real-time feedback that the user's training action has been improved will be output, which means that the action improvement performed by the user according to the correction suggestion is effective. On the contrary, when the improvement amount is not within the effective correction interval, real-time feedback that the user's training action has not been improved will be output, which indicates that the user may not have fully performed the action according to the correction suggestion in some aspects, so as to remind the user to further adjust and improve the corresponding training action.

[0123] See also Figure 2 , an interactive physical education teaching system based on video image processing, using the above-mentioned interactive physical education teaching method based on video image processing, comprising:

[0124] A feature extraction module is used to extract features from the acquired physical education teaching video to obtain multiple standard action frames;

[0125] The action comparison module is used to capture the user's training action frame in real time, and compare and analyze the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame;

[0126] The segmentation module is used to collect the cross nodes between the correct action frame and the deviation action frame, mark the cross nodes as key nodes, and determine the continuous correct time period and the continuous deviation time period based on the key nodes;

[0127] Deviation type output module, the deviation type output module is used to analyze the orderliness of the correct action frames within the continuous correct period, and obtain the deviation type of the deviation action frame after the end node of the continuous correct period;

[0128] A correction analysis module, which is used to perform correction analysis on the deviation action frames within the continuous deviation period, and output correction suggestions for each deviation action frame based on the correction analysis results;

[0129] The correction feedback module is used to collect the improvement amount of each deviation action frame under the correction suggestion, and provide real-time feedback on the user's training action based on the improvement amount, and update the correction suggestion of each deviation action frame.

[0130] In the above, the system includes a feature extraction module, an action comparison module, a segmentation module, a deviation type output module, a correction analysis module and a correction feedback module. The feature extraction module is responsible for analyzing the collected physical education teaching videos and extracting multiple key standard action frames therefrom. The standard action frames will be used as references for subsequent action comparison and analysis. The action comparison module captures the user's training action frames in real time, and compares and analyzes the captured training action frames in detail with the previously extracted standard action frames. Through the comparison results, it is possible to identify which actions are correct and which actions have deviations, and mark them as correct action frames and deviation action frames respectively. The segmentation module further analyzes the cross nodes between the correct action frames and the deviation action frames. These cross nodes are marked as key nodes for determining the continuous correct period and the continuous deviation period in the user's action. The deviation type output module focuses on the continuous correct period. The correct action frames are analyzed in an orderly manner. Through this analysis, the system can identify the specific deviation type of the deviation action frame after the end node of the continuous correct period, thereby providing users with more specific improvement directions. The correction analysis module performs corresponding correction analysis on the deviation action frames within the continuous deviation period. Based on the analysis results, the system can output specific correction suggestions for each deviation action frame to help users correct incorrect actions more effectively. The correction suggestions include how to adjust key elements such as body posture, movement amplitude, and force distribution to ensure the accuracy and effectiveness of the action. The correction feedback module is responsible for real-time collection of the improvement amount of each deviation action frame under the correction suggestion. Based on the improvement amount, the system can provide users with real-time training action feedback, and dynamically update the correction suggestions for each deviation action frame based on the user's actual improvement, so that users can continue to obtain personalized guidance and improve their sports skills faster.

[0131] See also Figure 3 , an electronic device, the electronic device comprising:

[0132] at least one processor;

[0133] and a memory communicatively coupled to the at least one processor;

[0134] Among them, the memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the above-mentioned interactive sports teaching method based on video image processing.

[0135] The processor of the electronic device mentioned above can be any general or dedicated central processing unit (CPU), such as an ARM architecture processor, an x86 architecture processor, or a dedicated image processing unit (GPU). The memory of the electronic device may include random access memory (RAM), read-only memory (ROM), flash memory, or other forms of non-volatile memory. Computer programs can be stored in the memory in any form, such as executable files, scripts, or firmware. The electronic device may also include an input and output (I / O) interface for connecting external devices such as a camera, a display, a keyboard, a mouse, etc. Through these interfaces, the user can interact with the electronic device, such as capturing training movements through a camera and viewing real-time feedback and standard action frames through a display.

[0136] In addition, the electronic device may also include an operator and a network communication module, wherein the operator is used to perform mathematical operations and data processing tasks to support the efficient operation of video image processing and motion analysis algorithms, and the network communication module enables the electronic device to connect to the Internet or other networks to achieve remote teaching and data sharing.

[0137] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0138] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. An interactive sports teaching method based on video image processing, characterized in that: include: Perform feature extraction on the acquired physical education teaching video to obtain multiple standard action frames; Capture the user's training action frame in real time, and compare and analyze the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame; Collecting the intersection nodes between the correct action frame and the deviation action frame, marking the intersection nodes as key nodes, and determining the continuous correct time period and the continuous deviation time period according to the key nodes; Performing orderliness analysis on the correct action frames within the continuous correct period to obtain the deviation type of the deviation action frame after the end node of the continuous correct period; Performing correction analysis on the deviation action frames within the continuous deviation period, and outputting correction suggestions for each of the deviation action frames according to the results of the correction analysis; The improvement amount of each of the deviation action frames under the correction suggestion is collected, and real-time feedback is given to the user's training action based on the improvement amount, and the correction suggestion of each of the deviation action frames is updated.

2. The interactive sports teaching method based on video image processing according to claim 1, characterized in that: The step of extracting features from the acquired physical education teaching video to obtain a plurality of standard action frames includes: Acquire a physical education teaching video, and preprocess the physical education teaching video, wherein the preprocessing includes denoising, contrast enhancement, and brightness adjustment; Decode the preprocessed physical education teaching video and extract the video frame sequence; Key action features are obtained, and key frames of a video frame sequence are screened according to the key action features to obtain a plurality of standard action frames.

3. The interactive sports teaching method based on video image processing according to claim 1, characterized in that: The step of comparing and analyzing the captured training action frame with the standard action frame to obtain the correct action frame and the deviation action frame includes: Capturing the user's current training action sequence and the training action frames under the training action sequence; Adding time stamps and space marks to the training action frames and standard action frames; Performing time synchronization and spatial alignment on the training action frame and the standard action frame, and performing comparative analysis on the training action frame and the standard action frame to obtain an action matching degree between the training action frame and the standard action frame; The training action frames are screened according to the action matching degree, and the screened training action frames are classified into correct action frames and deviation action frames.

4. The interactive sports teaching method based on video image processing according to claim 3 is characterized in that: The step of comparing and analyzing the training action frame and the standard action frame to obtain the action matching degree between the training action frame and the standard action frame includes: Acquire feature points of the training action frame and the standard action frame, and summarize them into a training action feature vector corresponding to the training action frame and a standard action feature vector corresponding to the standard action frame; Obtaining a matching degree calculation function, and inputting the training action feature vector and the standard action feature vector into the matching degree calculation function, and calibrating the output result of the matching degree calculation function as the action matching degree; After the action matching degree is output, the screening parameters are obtained; comparing the screening parameters with the action matching degree; When the screening parameter is greater than or equal to the action matching degree, the corresponding training action frame is marked as a correct action frame; When the screening parameter is less than the action matching degree, the corresponding training action frame is marked as a deviation action frame.

5. The interactive sports teaching method based on video image processing according to claim 1 is characterized in that: The step of determining the continuous correct time period and the continuous deviation time period according to the key nodes includes: Acquire time series information of the key nodes, and sort the key nodes according to the time series information; Counting the time lengths between adjacent key nodes and recording them as the interval length to be evaluated; Obtaining a standard interval duration, and comparing the standard interval duration with the interval duration to be evaluated; If the duration of the interval to be evaluated is less than or equal to the standard interval duration, the duration of the interval to be evaluated is recorded as an occasional period; If the duration of the interval to be evaluated is greater than the standard duration, the duration of the interval to be evaluated in which the training action frame is a correct action frame is recorded as a continuous correct period; otherwise, it is recorded as a continuous deviation period.

6. The interactive sports teaching method based on video image processing according to claim 1, characterized in that: The step of performing orderliness analysis on the correct action frames within the continuous correct period to obtain the deviation type of the deviation action frame after the end node of the continuous correct period includes: Acquire the correct action frames and standard action frames within the continuous correct time period, and arrange them according to the occurrence sequence; Collecting the action differences between the correct action frame and the standard action frame one by one, and recording them as the differences to be evaluated; Obtaining a verification function, inputting the difference to be evaluated into the verification function, and calibrating the output result of the verification function as a first verification condition parameter; Obtaining the action difference between the deviation action frame after the end node of the continuous correct period and the corresponding standard action frame, and recording it as the second verification condition parameter; Performing a difference processing on the first verification condition parameter and the second verification condition parameter, and recording the difference result of the first verification condition parameter and the second verification condition parameter as a classification condition parameter; Obtain a classification condition threshold, and compare the classification condition threshold with a classification condition parameter, and when the classification condition parameter is greater than or equal to the classification condition threshold, output the deviation type of the deviation action frame as an independent deviation, otherwise, output the deviation type of the deviation action frame as a progressive deviation.

7. The interactive sports teaching method based on video image processing according to claim 6 is characterized by: Under the progressive deviation, the deviation action frame is traced to determine the starting node of the deviation action frame, and according to the standard action frame, the correct action frame after the starting node in the continuous correct period is corrected and analyzed, and correction suggestions for each of the deviation action frames are output according to the results of the correction analysis; Under the independent deviation, correction analysis is directly performed on the deviation action frames within the continuous deviation period, and correction suggestions for each of the deviation action frames are output based on the results of the correction analysis.

8. The interactive sports teaching method based on video image processing according to claim 1 is characterized by: The step of providing real-time feedback on the user's training action according to the improvement amount comprises: After obtaining the correction suggestion output, the user's correction action frame; Performing time synchronization and spatial alignment on the correction action frame and the deviation action frame, and performing comparative analysis on the correction action frame and the deviation action frame to obtain an improvement amount of the deviation action frame; Obtaining a standard correction amount under a correction suggestion corresponding to each of the deviation action frames, and performing an offset process on the standard correction amount to obtain a valid correction interval; comparing the effective correction interval with the improvement amount; When the improvement amount is within the effective correction interval, outputting real-time feedback that the user's training action is improved; When the improvement amount is not within the effective correction interval, real-time feedback is outputted indicating that the user's training action has not been improved.

9. An interactive physical education teaching system based on video image processing, characterized in that: The interactive sports teaching method based on video image processing according to any one of claims 1 to 8 comprises: A feature extraction module, wherein the feature extraction module is used to extract features from the acquired physical education teaching video to obtain a plurality of standard action frames; An action comparison module, which is used to capture the user's training action frame in real time, and compare and analyze the captured training action frame with the standard action frame to obtain a correct action frame and a deviation action frame; A segmentation module, the segmentation module is used to collect the intersection nodes between the correct action frame and the deviation action frame, mark the intersection nodes as key nodes, and determine the continuous correct time period and the continuous deviation time period according to the key nodes; A deviation type output module, the deviation type output module is used to perform orderliness analysis on the correct action frames within the continuous correct period, and obtain the deviation type of the deviation action frame after the end node of the continuous correct period; A correction analysis module, the correction analysis module is used to perform correction analysis on the deviation action frames within the continuous deviation period, and output correction suggestions for each of the deviation action frames according to the results of the correction analysis; A correction feedback module, the correction feedback module is used to collect the improvement amount of each of the deviation action frames under the correction suggestion, and provide real-time feedback on the user's training action based on the improvement amount, and update the correction suggestion of each of the deviation action frames.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the interactive sports teaching method based on video image processing as described in any one of claims 1 to 8.

Images