Sports performance assessment method and system capable of identifying portrait features

Abstract

The invention discloses a sports performance assessment method and system capable of recognizing portrait features, and relates to the technical field of image recognition, and the method comprises the steps: deploying a signal transmitter and a signal receiver in an examination room, and forming a wireless signal grid architecture; constructing a learning model to train the examinee action characteristic parameters; and transmitting the parameter score to an assessment system through an image transmission method, and completing identification of sports performance assessment through image display. According to the method, the influence of subjective factors in traditional manual judgment is eliminated, the objectivity and the accuracy of sports performance assessment are improved, the assessment efficiency is greatly improved, the labor cost is reduced, meanwhile, personalized training suggestions can be provided according to specific expressions of examinees, and the training experience of the examinees is improved. And examinees are helped to pertinently improve and enhance sports skills, so that the fairness and scientificity of physical education are better promoted.

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and in particular to a sports performance assessment method and system for recognizing portrait features. Background Art

[0002] With the development of science and technology and the progress of society, people's requirements for education quality are increasing, especially in the field of physical education. Traditional assessment methods can no longer meet the needs of modern society. Traditional sports performance assessment mostly relies on manual observation and manual recording. This method is not only inefficient, but also easily affected by subjective factors, resulting in evaluation results that are not objective and fair. In addition, with the improvement of national fitness awareness, more and more people are participating in sports activities, and the demand for scientific evaluation of sports skills is also increasing. Therefore, it is particularly important to develop a method that can accurately and efficiently evaluate sports performance.

[0003] At present, the existing sports performance assessment methods mostly rely on manual judgment, which has obvious limitations, such as strong subjectivity, time-consuming and labor-intensive, and difficulty in ensuring fairness. Especially in large-scale examinations or competitions, due to the large number of candidates, the inefficiency of manual judgment is particularly prominent. In addition, it is difficult for traditional methods to accurately analyze the details of the candidates' movements and provide personalized training suggestions, which limits the further improvement of the candidates. The present invention proposes a sports performance assessment method for identifying portrait features based on a wireless signal grid architecture. By automatically collecting and analyzing the candidate's movement feature parameters, objective evaluation of sports performance and provision of personalized training suggestions are achieved. Summary of the invention

[0004] In view of the problems existing in the existing multi-source power grid information fusion and system based on the Internet of Things, the present invention is proposed.

[0005] Therefore, the purpose of the present invention is to provide a sports performance assessment method and system for identifying portrait features. In view of the problems of strong subjectivity, low efficiency and difficulty in providing personalized training suggestions in existing sports performance assessment methods, the present invention adopts a wireless signal grid architecture to automatically collect and analyze the technical means of the examinee's movement feature parameters, thereby realizing the objective evaluation of sports performance and the automatic provision of personalized training suggestions.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for assessing sports performance by identifying human portrait features, which includes deploying a signal transmitter and a signal receiver in an examination room, collecting the actions performed by examinees within the examination room, and forming a wireless signal grid architecture; using the wireless signal grid architecture to extract the examinee's action feature parameters from the wireless signal of the receiver, and constructing a learning model to train the examinee's action feature parameters; Interactive training suggestions are added to the assessment system, and the scoring of the examinee's movement characteristic parameters is completed based on the training suggestions. The parameter scores are transmitted to the assessment system through image transmission, and the recognition of sports performance assessment is completed through image display.

[0007] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, the signal transmitter includes N transmitters deployed at the corners and center of the examination room, and N receivers deployed around the transmitter. At this time, the position range of the transmitter and the receiver cannot exceed the distance M, and the transmitted signal is monitored according to the distance M; When the test system shows that the transmission signal is less than or equal to the distance M, the transmission signal is in normal transmission; When the test system shows that the transmission signal is higher than the distance M, the transmission signal received by the receiver does not meet the distance M. At this time, the transmission signal is in abnormal transmission; When the assessment system displays a transmission signal, there is no abnormality in the assessment system. At this time, the transmission signal is collected normally, and the transmission signal of normal transmission and the transmission signal of abnormal transmission are isolated. The isolation includes area A and area B. Area A includes the transmission signal of normal transmission, and area B includes the transmission signal of abnormal transmission.

[0008] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, the collection of the actions performed by the examinee within the examination room includes extracting the examinee's action feature parameters from the transmission signal normally transmitted in area A, and the examinee's action feature parameters include intensity change parameters, phase difference parameters and arrival time difference parameters. The examinee's action feature parameters are extracted by calculation based on the intensity change parameters, phase difference parameters and arrival time difference parameters. The specific calculation formula is: ;in, Represents the characteristic parameters of the examinee's movements, represents the arrival time difference parameter, represents the intensity variation parameter, represents the phase difference parameter; According to the candidate's action characteristic parameters , completing the wireless signal grid architecture, when the assessment system displays a transmission signal of abnormal transmission, the wireless signal grid architecture is adjusted, the adjustment includes preprocessing the transmission signal of abnormal transmission, the preprocessing includes enhancing, detecting and feature standardizing the transmission signal of abnormal transmission; The enhancement includes enhancing the transmission signal of abnormal transmission using an adaptive filter based on the characteristic parameters of the examinee's movements in area A. The specific calculation formula of the enhancement is: ;in, represents the enhancement value, represents the weight vector of the adaptive filter, Indicates abnormal transmission signal of input, The error signal representing the transmission signal of normal transmission and the transmission signal of abnormal transmission, represents the step size parameter; The detection includes removing abnormal values ​​of the transmission signal of the enhanced abnormal transmission, and extracting the characteristic parameters of the examinee's movements for a second time. The specific calculation formula for removing abnormal values ​​is: ;in, It represents the secondary extraction of the examinee's action feature parameters. represents an outlier data point, represents the mean of outlier data, Indicates the standard deviation of outlier data; when When , the data point is considered an outlier and is removed; otherwise, when When , the data point is considered normal, and the transmission signal is normally transmitted in area A; The feature standardization process includes scaling the secondary extracted examinee's motion feature parameters to interval R, where interval R refers to a transmission signal interval of normal transmission in the same range as area A.

[0009] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, wherein: extracting the candidate's action feature parameters from the wireless signal includes extracting the candidate's action feature parameters and re-extracting the candidate's action feature parameters, and training the candidate's action feature parameters includes introducing sample data of the candidate's action feature parameters to adjust the learning model, adopting the LSTM layer strategy of the learning model, and verifying the performance of the learning model; Adding a regularization term based on the learning model, the regularization term includes L1 and L2, the L1 includes an upper portion of the LSTM layer strategy, and the L2 includes a lower portion of the LSTM layer strategy; When the number of iterations of the regularization term increases, the candidate's action feature parameters are secondary identification; When the number of iterations of the regularization term decreases, the candidate's action feature parameters are first-level identification; The first-level recognition refers to the fact that the candidate's action feature parameters for recognizing portrait features are rare, and they are saved and displayed in the assessment system.

[0010] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, the training suggestion includes generating training suggestions according to the score and recognition level of the examinee's action feature parameters; When the motion characteristic parameters of the examinee for identifying the human figure are rare, it is necessary to score based on the transmission signal of the normal transmission in area A and the motion characteristic parameters of the examinee in area B extracted twice; When the action characteristic parameters of the examinee for identifying the portrait feature are normal, the score is directly given; The interactive mode includes text suggestions and image suggestions, and the text suggestions include generating text descriptions based on the scoring results and recognition levels to explain the scoring results and training suggestions; The image suggestions include using dynamic images or videos to show the correct action demonstration to help candidates understand and imitate intuitively; The scoring results include a score distribution graph display and a time series graph display, wherein the score distribution graph display includes using a histogram to display the distribution of scores and displaying the rare action characteristic parameters of examinees of the scores; The time series graph display includes using a line graph to display the changing trend of the score over time and marking the time period of the score.

[0011] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, wherein: the method of transmitting the parameter score to the assessment system through image transmission includes using the score distribution graph to transmit the displayed result in PNG format, according to the requirements of the API document, constructing an HTTP or POST request with the help of a programming language, transmitting the PNG file, and when the receiver of the assessment system receives the PNG file, a database record of the assessment system is created, and the results of the score distribution graph display and the time series graph display are associated; When the receiver of the assessment system creates new resources, the transmission of the PNG file is successful. When the receiver of the assessment system does not create new resources, the transmission of the PNG file fails.

[0012] As a preferred solution of the sports performance assessment method for identifying portrait features of the present invention, wherein: the recognition of sports performance assessment completed by image display includes checking whether the API document requirements are correct in the case of failure to transmit PNG files; Set the API document request header. When the status code of the request header is 200 or 201, parse the programming language and confirm the file success information returned by the server. When the status code of the request header is not 200 or 201, the programming language is parsed, the file failure information returned by the server is confirmed, the cause of the failure is processed in a timely manner, and the identification of the sports performance assessment is completed.

[0013] In a second aspect, the present invention provides a sports performance assessment system for recognizing portrait features, which comprises: The collection module deploys signal transmitters and signal receivers in the examination room to collect the actions performed by candidates in the examination room to form a wireless signal grid architecture; A training module, which uses a wireless signal grid architecture to extract the examinee's motion feature parameters from the receiver's wireless signal and constructs a learning model to train the examinee's motion feature parameters; The recognition module adds interactive training suggestions to the assessment system, completes the scoring of the examinee's movement feature parameters based on the training suggestions, transmits the parameter scores to the assessment system through image transmission, and completes the recognition of sports performance assessment through image display.

[0014] In a third aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, it implements any step of the above-mentioned sports performance assessment method for identifying portrait features.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the above-mentioned sports performance assessment method for identifying portrait features is implemented.

[0016] The beneficial effects of the present invention are as follows: the present invention automatically collects and analyzes the motion feature parameters of examinees by adopting a wireless signal grid architecture, which not only eliminates the influence of subjective factors in traditional manual judgment, improves the objectivity and accuracy of sports performance assessment, but also greatly improves the assessment efficiency and reduces labor costs. At the same time, the present invention can provide personalized training suggestions based on the specific performance of the examinees, which helps the examinees to improve and enhance their sports skills in a targeted manner, thereby better promoting the fairness and scientific nature of physical education. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them: Figure 1 A specific flow chart of a sports performance assessment method and system for identifying portrait features provided by an embodiment of the present invention.

[0018] Figure 2 A schematic diagram of a sports performance assessment method and system for recognizing portrait features provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned purposes, 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 drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0020] 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.

[0021] 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 term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0022] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0023] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0024] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0025] Example 1: Reference Figure 1 and Figure 2, which is the first embodiment of the present invention, provides a sports performance assessment method for identifying portrait features, comprising: S1: Deploy signal transmitters and signal receivers in the examination room to collect the actions performed by candidates within the examination room to form a wireless signal grid architecture.

[0026] The signal transmitter includes deploying N transmitters in the corners and the center of the examination room, and deploying N receivers around the transmitters. At this time, the position range of the transmitters and the receivers cannot exceed the distance M, and the transmitted signal is monitored according to the distance M; When the test system shows that the transmission signal is less than or equal to the distance M, the transmission signal is in normal transmission; When the test system shows that the transmission signal is higher than the distance M, the transmission signal received by the receiver does not meet the distance M. At this time, the transmission signal is in abnormal transmission; When the assessment system displays a transmission signal, there is no abnormality in the assessment system. At this time, the transmission signal is collected normally, and the transmission signal of normal transmission and the transmission signal of abnormal transmission are isolated. The isolation includes area A and area B. Area A includes the transmission signal of normal transmission, and area B includes the transmission signal of abnormal transmission.

[0027] S1.1: Collecting the actions performed by the examinee within the examination room includes extracting the examinee's action characteristic parameters from the transmission signal transmitted normally in area A. The examinee's action characteristic parameters include intensity change parameters, phase difference parameters and arrival time difference parameters. The examinee's action characteristic parameters are extracted by calculation based on the intensity change parameters, phase difference parameters and arrival time difference parameters. The specific calculation formula is: ;in, Represents the characteristic parameters of the examinee's movements, represents the arrival time difference parameter, represents the intensity variation parameter, represents the phase difference parameter; According to the candidate's action characteristic parameters , completing the wireless signal grid architecture, when the assessment system shows the transmission signal of abnormal transmission, the wireless signal grid architecture is adjusted, the adjustment includes preprocessing the transmission signal of abnormal transmission, the preprocessing includes enhancing, detecting and feature standardizing the transmission signal of abnormal transmission; The enhancement includes using an adaptive filter to enhance the abnormal transmission signal based on the characteristic parameters of the examinee's movements in area A. The specific calculation formula for the enhancement is: ;in, represents the enhancement value, represents the weight vector of the adaptive filter, Indicates abnormal transmission signal of input, The error signal representing the transmission signal of normal transmission and the transmission signal of abnormal transmission, represents the step size parameter; The detection includes removing abnormal values ​​of the enhanced abnormal transmission signal and extracting the candidate's action feature parameters for the second time. The specific calculation formula for removing abnormal values ​​is: ;in, It represents the secondary extraction of the examinee's action feature parameters. represents an outlier data point, represents the mean of outlier data, Indicates the standard deviation of outlier data; when When , the data point is considered an outlier and is removed; otherwise, when When , the data point is considered normal, and the transmission signal is normally transmitted in area A; The feature standardization process includes scaling the secondary extracted candidate's motion feature parameters to interval R, where interval R refers to the transmission signal interval of normal transmission in the same range as area A.

[0028] Furthermore, in the physical education performance assessment process of the present invention, the wireless signals generated when the examinees perform actions are collected by signal transmitters and receivers deployed in the examination room, and the signals are divided into two areas: Area A represents normally transmitted signals, and Area B represents abnormally transmitted signals. In one experiment, 1,000 sets of wireless signal data were collected, of which about 90% of the data belonged to Area A, i.e., normally transmitted signals, and the remaining 10% of the data belonged to Area B, i.e., abnormally transmitted signals.

[0029] For the normal transmission signal in area A, the examinee's motion characteristic parameters were extracted, mainly including intensity change parameter, phase difference parameter and arrival time difference parameter. Through analysis, it was found that the average value of the intensity change parameter was about -5dBm, and the standard deviation was 2dBm; the average value of the phase difference parameter was about 15°, and the standard deviation was 5°; the average value of the arrival time difference parameter was about 20ns, and the standard deviation was 5ns. The data provided a basis for the subsequent characteristic parameter calculation.

[0030] When the assessment system detects an abnormally transmitted signal, the system will automatically start the preprocessing process to improve the signal quality. In the enhancement stage, an adaptive filter is used to enhance the abnormally transmitted signal. Experimental results show that after the enhancement process, the intensity of the abnormally transmitted signal is increased by an average of about 10dBm, the phase difference is corrected by about 10°, and the arrival time difference is corrected by about 10ns, which significantly improves the signal quality.

[0031] The enhanced signal was subjected to outlier removal. Through calculation, it was found that the average value of the outlier data points was about -10dBm and the standard deviation was 4dBm. According to the set standard, when the data point deviates from the average value by more than 3 times the standard deviation, it is considered an outlier and is removed. In actual operation, about 5% of the data points are identified as outliers and effectively removed.

[0032] In the feature standardization processing stage, the second extracted candidate action feature parameters are scaled to a normal transmission signal interval in the same range as area A. Specifically, the intensity change parameter, phase difference parameter and arrival time difference parameter are scaled to the intervals of -10dBm to -1dBm, 10° to 20° and 15ns to 25ns respectively to ensure that all feature parameters are within a reasonable range, which is convenient for subsequent model training and performance evaluation. The candidate action feature parameter data are shown in Table 1 below: Table 1 Data table of candidate action characteristic parameters Data Types Parameter name average value Standard Deviation unit Normal transmission signal (Area A) Intensity variation parameters -5dBm 2dBm dBm Normal transmission signal (Area A) Phase difference parameters 15° 5° ° Normal transmission signal (Area A) Time difference of arrival parameter 20ns 5ns ns Abnormal transmission signal ratio Area A 90% - - Abnormal transmission signal ratio Area B 10% - - After abnormal transmission signal enhancement Intensity variation parameters +10dBm - dBm After abnormal transmission signal enhancement Phase difference parameters 10° - ° After abnormal transmission signal enhancement Time difference of arrival parameter 10ns - ns Outlier Removal Average value of outlier data points -10dBm 4dBm dBm Outlier Removal Elimination ratio 5% - - Feature normalization Intensity variation parameter range -10dBm -1dBm dBm Feature normalization Phase difference parameter range 10° 20° ° Feature normalization Arrival time difference parameter interval 15ns 25ns ns Table 1 shows the key data of examinees' movement characteristic parameters collected and processed through the wireless signal grid architecture during the physical education performance assessment, including the ratio of normal transmission signals (area A) and abnormal transmission signals (area B), as well as the mean and standard deviation of various characteristic parameters (intensity change parameters, phase difference parameters and arrival time difference parameters). In addition, the table also lists the effect of abnormal transmission signal enhancement, the ratio of outliers eliminated, and the parameter range after feature standardization. These data provide important support for improving the accuracy and reliability of physical education performance assessment.

[0033] S2: Use the wireless signal grid architecture to extract the candidate's motion feature parameters from the receiver's wireless signal, and build a learning model to train the candidate's motion feature parameters.

[0034] The extraction of the candidate's motion characteristic parameters from the wireless signal includes extracting the candidate's motion characteristic parameters and re-extracting the candidate's motion characteristic parameters. The training of the candidate's motion characteristic parameters includes introducing sample data of the candidate's motion characteristic parameters to adjust the learning model, adopting the LSTM layer strategy of the learning model, and verifying the performance of the learning model. Regularization items are added based on the learning model. The regularization items include L1 and L2. L1 includes the upper part of the LSTM layer strategy, and L2 includes the lower part of the LSTM layer strategy. When the number of iterations of the regularization term increases, the candidate's action feature parameters are secondary identification; When the number of iterations of the regularization term decreases, the candidate's action feature parameters are first-level identification; Level 1 recognition means that the action characteristic parameters of the examinee for recognizing portrait features are rare, and they are saved and displayed in the assessment system.

[0035] Furthermore, in the sports performance assessment method of the present invention, the motion feature parameters of the examinee are first extracted from the wireless signal of the receiver through the wireless signal grid architecture, including intensity change parameters, phase difference parameters and arrival time difference parameters. After preprocessing the abnormally transmitted signal, the feature parameters are extracted for the second time. The constructed learning model adopts the LSTM layer strategy, and uses training data and verification data for model training, and the model achieves a high accuracy rate. L1 and L2 regularization terms are added. When the number of iterations of the regularization terms increases, most of the examinee's motion feature parameters are identified as secondary recognition; when the number of iterations decreases, some of the examinee's motion feature parameters are identified as primary recognition. Finally, interactive training suggestions are generated through the assessment system, and the parameter scores are transmitted to the assessment system through image transmission, and the recognition of the sports performance assessment is completed through image display.

[0036] S3: Add interactive training suggestions to the assessment system, complete the scoring of the examinee's movement characteristic parameters based on the training suggestions, transmit the parameter scores to the assessment system through image transmission, and complete the recognition of sports performance assessment through image display.

[0037] Among them, the training suggestions include generating training suggestions based on the examinee's action feature parameter scores and recognition levels; When the motion characteristic parameters of the examinee for identifying the human figure are rare, it is necessary to score based on the transmission signal of the normal transmission in area A and the motion characteristic parameters of the examinee in area B extracted twice; When the action characteristic parameters of the examinee for identifying the portrait feature are normal, the score is directly given; The interactive mode includes text suggestions and image suggestions. The text suggestions include generating text descriptions based on the scoring results and recognition levels to explain the scoring results and training suggestions. Graphic suggestions include using dynamic images or videos to show correct action demonstrations to help candidates intuitively understand and imitate; The scoring results include a score distribution graph and a time series graph. The score distribution graph uses a histogram to display the score distribution and displays the rare action characteristic parameters of the examinees. The time series graph display includes using a line graph to show the trend of score changes over time and marking the time period of the score.

[0038] S3.1: Transmitting the parameter scores to the assessment system through the image transmission method includes using the score distribution graph to transmit the displayed results in PNG format. According to the requirements of the API document, an HTTP or POST request is constructed with the help of a programming language to transmit the PNG file. When the receiver of the assessment system receives the PNG file, a database record of the assessment system is created to associate the results displayed by the score distribution graph with those displayed by the time series graph; When the receiver of the assessment system creates new resources, the transmission of the PNG file is successful. When the receiver of the assessment system does not create new resources, the transmission of the PNG file fails.

[0039] S3.2: Complete the identification of sports performance assessment through image display, including checking whether the API document requirements are correct for the failure of transmitting PNG files; Set the API document request header. When the status code of the request header is 200 or 201, parse the programming language and confirm the file success information returned by the server. When the status code of the request header is not 200 or 201, the programming language is parsed, the file failure information returned by the server is confirmed, the cause of the failure is processed in a timely manner, and the identification of the sports performance assessment is completed.

[0040] In a preferred embodiment, a sports performance assessment system for identifying human portrait features includes a collection module, which deploys a signal transmitter and a signal receiver in an examination room to collect the actions performed by the examinee within the examination room to form a wireless signal grid architecture; a training module, which uses the wireless signal grid architecture to extract the examinee's action feature parameters from the receiver's wireless signal, and constructs a learning model to train the examinee's action feature parameters; an identification module, which adds interactive training suggestions to the assessment system, completes the score of the examinee's action feature parameters based on the training suggestions, transmits the parameter score to the assessment system through an image transmission method, and completes the recognition of the sports performance assessment through image display.

[0041] The above-mentioned unit modules may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above-mentioned modules.

[0042] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0043] In summary, the present invention adopts a wireless signal grid architecture to automatically collect and analyze the motion feature parameters of examinees, which not only eliminates the influence of subjective factors in traditional manual judgment, improves the objectivity and accuracy of sports performance assessment, but also greatly improves the assessment efficiency and reduces labor costs. At the same time, the present invention can provide personalized training suggestions based on the specific performance of the examinees, which helps the examinees to improve and enhance their sports skills in a targeted manner, thereby better promoting the fairness and scientific nature of physical education. Example

[0044] Reference Figure 1 and Figure 2 , which is the second embodiment of the present invention, and this embodiment provides a sports performance assessment method for identifying portrait features. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0045] In the sports performance assessment method of the present invention, the action feature parameters of the examinee are first extracted from the wireless signal of the receiver through the wireless signal grid architecture, including intensity change parameters, average value -6dBm, standard deviation 2.5dBm; phase difference parameters, average value 16°, standard deviation 4°; arrival time difference parameters, average value 22ns, standard deviation 6ns. The abnormally transmitted signal is preprocessed, including signal enhancement, outlier removal and feature standardization processing, and the secondary extracted feature parameters are: intensity change parameters, average value -4.5dBm, standard deviation 1.8dBm; phase difference parameters, average value 14.5°, standard deviation 3.5°; arrival time difference parameters, average value 19ns, standard deviation 4.5ns. The constructed learning model adopts the LSTM layer strategy, uses 1200 sets of training data and 300 sets of verification data, and the model accuracy rate reaches 96%. When the L1 and L2 regularization terms are added, when the number of iterations of the regularization terms increases to 120, about 85% of the candidate's action feature parameters are identified as secondary recognition; when the number of iterations is reduced to 30, about 15% of the candidate's action feature parameters are identified as primary recognition. Finally, the assessment system generates interactive training suggestions, and the parameter scores are transmitted to the assessment system through image transmission. The recognition of sports performance assessment is completed through image display. The candidate's action feature parameters and model training data are shown in Table 2 below: Table 2. Table of candidate action feature parameters and model training data Data Types Parameter name average value Standard Deviation unit Data volume Accuracy Iterations Identification level Initial extraction feature parameters Intensity variation parameters -6dBm 2.5dBm dBm 0 0 0 0 Initial extraction feature parameters Phase difference parameters 16° 4° ° 0 0 0 0 Initial extraction feature parameters Time difference of arrival parameter 22ns 6ns ns 0 0 0 0 Secondary extraction of feature parameters Intensity variation parameters -4.5dBm 1.8dBm dBm 0 0 0 0 Secondary extraction of feature parameters Phase difference parameters 14.5° 3.5° ° 0 0 0 0 Secondary extraction of feature parameters Time difference of arrival parameter 19ns 4.5ns ns 0 0 0 0 Model Training Amount of training data 0 0 0 1200 96% 0 0 Model Training Verify data volume 0 0 0 300 96% 0 0 Regularization term Iterations: 120 0 0 0 0 0 120 Secondary Identification Regularization term Iterations: 30 0 0 0 0 0 30 First level identification Table 2 shows the relevant data of the candidate's action feature parameters extracted from the wireless signal and its model training in the sports performance assessment method. The initially extracted feature parameters include intensity change parameters, with an average value of -6dBm and a standard deviation of 2.5dBm; phase difference parameters, with an average value of 16° and a standard deviation of 4°; and arrival time difference parameters, with an average value of 22ns and a standard deviation of 6ns. After preprocessing, the secondary extracted feature parameters are intensity change parameters, with an average value of -4.5dBm and a standard deviation of 1.8dBm; phase difference parameters, with an average value of 14.5° and a standard deviation of 3.5°; and arrival time difference parameters, with an average value of 19ns and a standard deviation of 4.5ns. The model training used 1,200 sets of training data and 300 sets of verification data, and the model accuracy reached 96%. When the number of iterations of the regularization term is 120, about 85% of the candidate's action feature parameters are identified as secondary recognition; when the number of iterations is 30, about 15% of the candidate's action feature parameters are identified as primary recognition. The comparison between the present invention and the prior art is shown in Table 3 below: Table 3 Comparison between the present invention and the prior art Compare Projects Prior art The present invention Advantages Data collection method Manual observation and recording or simple sensors Wireless signal mesh architecture Automated data collection reduces human errors and improves data accuracy Feature parameter extraction Manual measurement or simple algorithm Multi-dimensional feature parameter extraction (intensity change, phase difference, arrival time difference) Extract feature parameters of more dimensions to improve recognition accuracy Outlier handling Reliance on human judgment Automatic preprocessing (signal enhancement, outlier removal, feature standardization) Automatically process outliers to improve data quality and model training results Model Training Simple statistical models or empirical rules LSTM Neural Network More powerful learning ability, higher model accuracy (96%) Regularization No or simple regularization L1 and L2 Regularization Prevent overfitting and improve model generalization ability Recognition accuracy Low, easily affected by subjective factors High (96%), strong objectivity Significantly improve recognition accuracy and reduce the impact of subjective factors Amount of training data Small-scale datasets Large-scale dataset (1200 sets of training data, 300 sets of validation data) More data support improves the robustness and generalization ability of the model Dynamic adjustment capability Limited, difficult to adjust in real time Dynamically adjust the number of regularization iterations Adjust model parameters in real time to meet recognition needs in different scenarios Training suggestion generation Manually generated or None Automatically generate interactive training suggestions Provide personalized training suggestions to help candidates improve and enhance their sports skills in a targeted manner Results Simple text or graphics Image transmission and display More intuitive and detailed display of results for easier understanding and analysis The present invention automatically collects the motion feature parameters of examinees through a wireless signal grid architecture, including intensity changes, phase differences and arrival time differences. Compared with the manual observation records or simple sensors in the prior art, it reduces human errors and improves data accuracy. The present invention adopts multi-dimensional feature parameter extraction and automatic preprocessing to significantly improve recognition accuracy. The model training uses LSTM neural networks and large-scale data sets with an accuracy rate of 96%, and prevents overfitting through L1 and L2 regularization. In addition, the present invention can dynamically adjust the number of iterations of the regularization term, provide personalized training suggestions, and transmit and display results through images, so that examinees can more easily understand and improve their movements.

[0046] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A sports performance assessment method for identifying portrait features, characterized in that: include, Signal transmitters and receivers are deployed in the examination room to collect the actions performed by candidates in the examination room and form a wireless signal grid architecture; The wireless signal grid architecture is used to extract the candidate's motion feature parameters from the receiver's wireless signal, and a learning model is constructed to train the candidate's motion feature parameters; Interactive training suggestions are added to the assessment system, and the scoring of the examinee's movement characteristic parameters is completed based on the training suggestions. The parameter scores are transmitted to the assessment system through image transmission, and the recognition of sports performance assessment is completed through image display.

2. The method for assessing sports performance by identifying portrait features as claimed in claim 1, characterized in that: The signal transmitter includes deploying N transmitters at the corners and the center of the examination room, and deploying N receivers around the transmitters. At this time, the position range of the transmitters and the receivers cannot exceed the distance M, and the transmitted signal is monitored according to the distance M; When the test system shows that the transmission signal is less than or equal to the distance M, the transmission signal is in normal transmission; When the test system shows that the transmission signal is higher than the distance M, the transmission signal received by the receiver does not meet the distance M. At this time, the transmission signal is in abnormal transmission; When the assessment system displays a transmission signal, there is no abnormality in the assessment system. At this time, the transmission signal is collected normally, and the transmission signal of normal transmission and the transmission signal of abnormal transmission are isolated. The isolation includes area A and area B. Area A includes the transmission signal of normal transmission, and area B includes the transmission signal of abnormal transmission.

3. The method for assessing sports performance by identifying portrait features as claimed in claim 2, characterized in that: The collecting of the actions performed by the examinee within the examination room includes extracting the examinee's action characteristic parameters from the transmission signal normally transmitted in area A, wherein the examinee's action characteristic parameters include intensity change parameters, phase difference parameters, and arrival time difference parameters, and calculating according to the intensity change parameters, phase difference parameters, and arrival time difference parameters to extract the examinee's action characteristic parameters, and the specific calculation formula is: ;in, Represents the characteristic parameters of the examinee's movements, represents the arrival time difference parameter, represents the intensity variation parameter, represents the phase difference parameter; According to the candidate's action characteristic parameters , completing the wireless signal grid architecture, when the assessment system displays a transmission signal of abnormal transmission, the wireless signal grid architecture is adjusted, the adjustment includes preprocessing the transmission signal of abnormal transmission, the preprocessing includes enhancing, detecting and feature standardizing the transmission signal of abnormal transmission; The enhancement includes enhancing the transmission signal of abnormal transmission using an adaptive filter based on the characteristic parameters of the examinee's movements in area A. The specific calculation formula of the enhancement is: ;in, represents the enhancement value, represents the weight vector of the adaptive filter, Indicates abnormal transmission signal of input, The error signal representing the transmission signal of normal transmission and the transmission signal of abnormal transmission, represents the step size parameter; The detection includes removing abnormal values ​​of the transmission signal of the enhanced abnormal transmission, and extracting the characteristic parameters of the examinee's movements for a second time. The specific calculation formula for removing abnormal values ​​is: ;in, It represents the secondary extraction of the examinee's action feature parameters. represents an outlier data point, represents the mean of outlier data, Indicates the standard deviation of outlier data; when When , the data point is considered an outlier and is removed; otherwise, when When , the data point is considered normal, and the transmission signal is normally transmitted in area A; The feature standardization process includes scaling the secondary extracted examinee's motion feature parameters to interval R, where interval R refers to a transmission signal interval of normal transmission in the same range as area A.

4. The method for assessing sports performance by identifying portrait features as claimed in claim 3, characterized in that: Extracting the candidate's motion feature parameters from the wireless signal includes extracting the candidate's motion feature parameters and re-extracting the candidate's motion feature parameters, and training the candidate's motion feature parameters includes introducing sample data of the candidate's motion feature parameters to adjust the learning model, adopting the LSTM layer strategy of the learning model, and verifying the performance of the learning model; Adding a regularization term based on the learning model, the regularization term includes L1 and L2, the L1 includes an upper portion of the LSTM layer strategy, and the L2 includes a lower portion of the LSTM layer strategy; When the number of iterations of the regularization term increases, the candidate's action feature parameters are secondary identification; When the number of iterations of the regularization term decreases, the candidate's action feature parameters are first-level identification; The first-level recognition refers to the fact that the candidate's action feature parameters for recognizing portrait features are rare, and they are saved and displayed in the assessment system.

5. The method for assessing sports performance by identifying portrait features as claimed in claim 4, characterized in that: The training suggestion includes generating a training suggestion based on the examinee's action feature parameter score and recognition level; When the motion characteristic parameters of the examinee for identifying the human figure are rare, it is necessary to score based on the transmission signal of the normal transmission in area A and the motion characteristic parameters of the examinee in area B extracted twice; When the action characteristic parameters of the examinee for identifying the portrait feature are normal, the score is directly given; The interactive mode includes text suggestions and image suggestions, and the text suggestions include generating text descriptions based on the scoring results and recognition levels to explain the scoring results and training suggestions; The image suggestions include using dynamic images or videos to show the correct action demonstration to help candidates understand and imitate intuitively; The scoring results include a score distribution graph display and a time series graph display, wherein the score distribution graph display includes using a histogram to display the distribution of scores and displaying the rare action characteristic parameters of examinees of the scores; The time series graph display includes using a line graph to display the changing trend of the score over time and marking the time period of the score.

6. The method for assessing sports performance by identifying portrait features as claimed in claim 5, characterized in that: The method of transmitting the parameter score to the assessment system through image transmission includes using the score distribution graph to transmit the displayed result in PNG format, constructing an HTTP or POST request with the help of a programming language according to the requirements of the API document, transmitting the PNG file, and when the receiver of the assessment system receives the PNG file, creating a database record of the assessment system, and associating the results displayed by the score distribution graph and the time series graph; When the receiver of the assessment system creates new resources, the transmission of the PNG file is successful. When the receiver of the assessment system does not create new resources, the transmission of the PNG file fails.

7. The method for assessing sports performance by identifying portrait features as claimed in claim 6, characterized in that: The identification of completing the sports performance assessment through image display includes checking whether the API document requirements are correct in the case of failure to transmit the PNG file; Set the API document request header. When the status code of the request header is 200 or 201, parse the programming language and confirm the file success information returned by the server. When the status code of the request header is not 200 or 201, the programming language is parsed, the file failure information returned by the server is confirmed, the cause of the failure is processed in a timely manner, and the identification of the sports performance assessment is completed.

8. A sports performance assessment system for identifying human portrait features, based on the sports performance assessment method for identifying human portrait features as described in any one of claims 1 to 7, characterized in that: include, The collection module deploys signal transmitters and signal receivers in the examination room to collect the actions performed by candidates in the examination room to form a wireless signal grid architecture; A training module, which uses a wireless signal grid architecture to extract the examinee's motion feature parameters from the receiver's wireless signal and constructs a learning model to train the examinee's motion feature parameters; The recognition module adds interactive training suggestions to the assessment system, completes the scoring of the examinee's movement feature parameters based on the training suggestions, transmits the parameter scores to the assessment system through image transmission, and completes the recognition of sports performance assessment through image display.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the sports performance assessment method for identifying portrait features described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the sports performance assessment method for identifying portrait features as described in any one of claims 1 to 7 are implemented.

Images