Intelligent acquisition method and equipment for teaching behaviors, and storage medium

Abstract

The invention provides an intelligent collection method and device for teaching behaviors and a storage medium, and is applied to the technical field of data processing. The method comprises the following steps: processing target user information to generate physiological parameter information of a target user; processing the target class information to generate preset course information; processing the preset course information and the physiological parameter information of the target user based on a teaching behavior early warning model to generate course adjustment information; processing the teaching purpose information matched with the target class information and the physiological parameter information of the target user based on a teaching behavior early warning model to generate a course matching influence factor; processing the target user information based on the course adjustment information and the course matching influence factor to generate course information of the target user; and processing the course information of the target user to generate teaching behavior information.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an intelligent collection method, device and storage medium for teaching behavior. Background Art

[0002] In the traditional teaching model, teachers usually teach according to the established teaching plan. The fixed teaching plan is often accompanied by a single evaluation system, which may not be able to fully evaluate the comprehensive ability of students. Under the fixed teaching model, it may be difficult for teachers to carry out personalized resource allocation and support according to the needs of different students. Taking the enlightenment education of young children as an example, in order to better provide enlightenment education for young children, parents or teachers usually observe the children's response to external things, discover the children's interests, and develop learning plans for them.

[0003] However, due to the different ways of thinking between adults and children, it is difficult for parents and teachers to accurately guess the interests of children. Moreover, if teachers are responsible for too many children, the accuracy of discovering the children's interests will often be affected due to insufficient observation of each child's behavior, resulting in unreasonable learning plans for children, which affects the children's learning efficiency.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of this application is to provide an intelligent collection method, device, storage medium and system for teaching behavior, which at least overcomes the problems existing in the prior art to a certain extent, by obtaining multiple types of information and teaching behavior early warning models, generating physiological parameter information from target user information, and generating preset course information from target class information. Then, with the help of the early warning model, the course adjustment information and course matching influencing factors are obtained, and then the course information containing environment, emotion and progress values ​​is generated. Finally, by processing the course information, the teaching resources and activity type information, the user's multi-faceted status information, and the learning behavior pattern information are obtained in turn, and finally the teaching behavior information is generated, which helps the training class adjust the teaching according to the students' situation, realize personalized teaching, and improve the teaching quality and effect.

[0006] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.

[0007] According to one aspect of the present application, there is provided a method for intelligently collecting teaching behaviors, comprising: obtaining target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information, wherein the target user information comprises user information to be monitored, and the users to be monitored are respectively in different classes at different time periods; processing the target user information to generate physiological parameter information of the target user, wherein the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes; processing the target class information to generate preset course information, wherein the preset course The information is generated based on the interests and hobbies of the target user; the preset course information and the physiological parameter information of the target user are processed based on the teaching behavior warning model to generate course adjustment information; the teaching purpose information matching the target class information and the physiological parameter information of the target user are processed based on the teaching behavior warning model to generate a course matching influence factor; the target user information is processed based on the course adjustment information and the course matching influence factor to generate the course information of the target user; the course information of the target user is processed to generate teaching behavior information, wherein the teaching behavior information is used to characterize the learning progress of the target user and the learning time information of the target user.

[0008] Another aspect of the present application is an intelligent collection device for teaching behavior, characterized in that it includes: an acquisition module, used to acquire target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information, wherein the target user information includes user information to be monitored, and the users to be monitored are respectively in different classes at different time periods; a processing module, used to process the target user information and generate physiological parameter information of the target user, wherein the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes; the target class information is processed to generate preset course information, wherein In the method, the preset course information is generated based on the interests and hobbies of the target user; the preset course information and the physiological parameter information of the target user are processed based on the teaching behavior warning model to generate course adjustment information; the teaching purpose information matching the target class information and the physiological parameter information of the target user are processed based on the teaching behavior warning model to generate a course matching influence factor; the target user information is processed based on the course adjustment information and the course matching influence factor to generate the course information of the target user; the course information of the target user is processed to generate teaching behavior information, wherein the teaching behavior information is used to characterize the learning progress of the target user and the learning time information of the target user.

[0009] According to another aspect of the present application, an electronic device is provided, characterized in that it comprises: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the intelligent collection method for implementing the above-mentioned teaching behavior by executing the executable instructions.

[0010] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a second processor, the intelligent collection method of teaching behaviors described above is implemented.

[0011] According to another aspect of the present application, a computer program product is provided, including a computer program, characterized in that when the computer program is executed by a third processor, the computer program implements the above-mentioned intelligent collection method of teaching behavior.

[0012] The present application provides an intelligent collection method, device and storage medium for teaching behavior, which obtains target user information, target class information, teaching purpose information and corresponding teaching behavior early warning model, wherein the target user may be in different classes at different time periods. Then, the target user information is processed to obtain physiological parameter information, including generating expression feature influencing factors from the initial facial image information, and then deriving emotional information, and obtaining real-time physical sign information from identity information, and combining the two to obtain physiological parameter information. At the same time, the target class information is processed to generate preset course information based on user interests and hobbies. Then, the preset course information and physiological parameter information are processed using the teaching behavior early warning model to obtain course adjustment information, which is specifically achieved by generating a synergistic value of courses and environmental factors; the teaching purpose information and physiological parameter information can also be processed to obtain course matching influencing factors. The target user information is processed based on the course adjustment information and the course matching influencing factors to generate course information including environmental factor adjustment factors, emotional values ​​and learning progress values.

[0013] Finally, the target user's course information is processed to obtain teaching resource information and teaching activity type information, and then the user's emotional information, learning progress value and learning time are generated based on this information, and then the learning behavior pattern information is obtained, and finally the teaching behavior information representing the learning progress and time is generated. This method helps interest training classes adjust teaching according to students' situations, improve teaching quality and effect, and realize personalized teaching.

[0014] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flow chart showing an intelligent collection method of teaching behavior provided by an embodiment of the present application;

[0016] Figure 2 A schematic diagram of the structure of an intelligent collection device for teaching behavior provided in one embodiment of the present application is shown. DETAILED DESCRIPTION

[0017] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0018] Combine the following Figure 1 To describe the intelligent collection method of teaching behavior according to the exemplary embodiment of the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principle of the present application, and the embodiments of the present application are not limited in this respect. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0019] In one implementation, the present application also proposes an intelligent collection method, device and storage medium for teaching behavior. Figure 1 The following schematically shows a flow chart of an intelligent collection method of teaching behavior according to an embodiment of the present application. Figure 1 As shown, the method is applied to a server, comprising:

[0020] S101, obtaining target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information.

[0021] In one embodiment, the target user information includes user information to be monitored, and the users to be monitored are in different classes at different time periods. The user information to be monitored includes the student's basic identity information (such as name, age, gender), past learning experience (whether there is a foundation for relevant interest courses), interest preferences (determined through a questionnaire or interview before admission, for example, those who are sensitive to colors may prefer painting, and those who have a sense of rhythm may be more suitable for music) and learning goals (whether to cultivate as a hobby or hope to have expertise in this field), etc. For example, student Xiao Ming, 10 years old, boy, has participated in some children's creative painting activities before, has a strong interest in painting, and hopes to further improve his painting skills. He signed up for the school's advanced painting interest class. At the same time, he also showed a certain curiosity about programming. He participated in a simple introductory programming course in the school's holiday experience course, and he also wanted to study programming in depth in the future, so he might participate in painting classes and programming classes at different time periods.

[0022] Each interest training class is divided into different classes according to course content, difficulty level, teaching methods and other factors. For example, the painting interest class is divided into elementary graffiti class, intermediate sketching class, advanced oil painting class, etc.; the music interest class is divided into basic music theory class, elementary instrument performance class, chorus class, etc. The target class information includes the class's syllabus (covering the specific knowledge and skills to be taught, such as line application, color matching, composition methods, etc. in the painting class; note recognition, beat mastery, repertoire performance skills, etc. in the music class), teaching plan (weekly course schedule, distribution of teaching key points and difficulties, ratio of practical exercises to theoretical explanations, etc.), teacher allocation (teacher's professional background, teaching experience, teaching style, etc.) and class size (limited number of students). Taking the advanced painting interest class that Xiao Ming signed up for as an example, the class's syllabus clearly states that this semester will focus on learning the proportion structure, light and shadow performance and other skills of figure sketching. The teaching plan is two classes per week. Each class will first conduct theoretical explanations and demonstrations, and then arrange a lot of time for students to practice. The teacher is a teacher with many years of painting teaching experience and is good at realistic style. The class size is 20 people.

[0023] The teaching purposes of different interest training classes are different. The teaching purpose of the painting class may be to cultivate students' aesthetic ability, creativity and artistic expression, and to express emotions and ideas through painting; the music class aims to improve students' musical literacy, sense of rhythm and performance and singing ability, and cultivate their understanding and appreciation of music; the dance class focuses on the training of body coordination, flexibility and dance skills, while cultivating stage performance and self-confidence; the programming class is to exercise logical thinking ability and problem-solving ability, and cultivate students' ability to use programming to realize creativity and solve practical problems. For Xiao Ming's advanced painting interest class, its teaching purpose is to further improve students' realistic painting skills based on their existing painting skills, guide students to use different painting tools and materials to accurately express the shape, texture and spatial relationship of objects, and encourage students to incorporate their own creativity and emotions into their works to cultivate independent creative ability.

[0024] The teaching behavior early warning model is used to predict and evaluate the risk factors that may affect the teaching effect and student learning experience during the teaching process, and provide timely adjustment suggestions. The teaching behavior early warning model consists of an input layer, a hidden layer, and an output layer. Specifically, the input layer is responsible for receiving various feature data extracted from the training samples. These data types are diverse and cover many aspects related to teaching behavior and student learning. For example, for a painting interest class, the data received includes students' basic information (age, gender), painting skill-related data (basic painting level, previous work scores, etc.), course-related information (course content difficulty), teaching environment data (temperature, light intensity, noise decibel value), teacher teaching characteristics (demonstration frequency, guidance method), student learning behavior and status data (classroom participation, number of questions, emotional state obtained by expression recognition, weekly practice time), and evaluation data (stage test scores, student self-evaluation, teacher evaluation), etc.

[0025] The hidden layer includes the first hidden layer for feature learning and pattern recognition and the second hidden layer for risk factor mining. Feature learning and pattern recognition (first hidden layer), this hidden layer uses the powerful learning ability of neural networks to conduct in-depth learning and analysis of multiple input feature data. It can identify complex relationships and potential patterns between different features. For example, it can find the relationship between students' age and the speed of understanding color theory, or the connection between the basic level of painting and the difficulty of learning specific painting skills. Through the study of a large amount of data, a cognitive model of the interaction of various factors in the teaching process of painting interest classes is gradually constructed. At the same time, it can model the changing trends of students' learning status under different teaching environments and teaching methods. For example, in the two cases of high-frequency demonstration and low-frequency demonstration by teachers, the different modes of students' painting skills improvement speed, and the comprehensive relationship between these modes and factors such as students' own foundation and course difficulty.

[0026] Risk factor mining (second hidden layer) is based on the feature relationships and patterns learned in the first hidden layer to further explore potential risk factors that affect teaching behavior. For example, through analysis, it is found that when the difficulty of the course content suddenly increases and the teacher's demonstration frequency is not adjusted accordingly, students are prone to the risk of inattention and difficulty in mastering painting skills. Or when the noise in the classroom environment exceeds a certain decibel and the light is dim, students tend to become irritable, which in turn affects the quality and efficiency of painting creation. This hidden layer can also analyze the mutual influence between different risk factors to determine which risk factors combined will have a greater negative impact on teaching effectiveness. For example, when the teacher's guidance method is inappropriate and the course content is too difficult, students may feel more frustrated, and may even cause some students to give up learning painting.

[0027] The output layer generates specific risk warning information based on the risk factors and analysis results mined by the hidden layer. For example, when it is detected that the temperature of the current teaching environment is too high, the teacher's demonstration frequency is low, and the course content is difficult, the output warning reminds the teacher that there is a high risk in the current teaching situation, and students may have problems such as inattention and decreased learning enthusiasm. At the same time, combined with past successful teaching cases and data patterns, targeted teaching suggestions are provided to teachers. For example, it is recommended that teachers appropriately reduce the difficulty of the course, increase the number of demonstrations, improve the classroom environment (such as adjusting the temperature, enhancing the light), or adopt more flexible guidance methods (such as increasing heuristic guidance and reducing direct correction) to improve students' learning effects and learning experience.

[0028] In addition to the previously mentioned student classroom performance, homework completion, test scores, evaluation information, and environmental factor data, further expand data acquisition channels. Data on students' extracurricular painting practices can be collected, such as performance in painting competitions, style and technique application of daily independent painting works, etc. These data can more comprehensively reflect the development trend of students' painting abilities and interests, and enrich training samples. For painting interest classes, data on students' preferences for different painting subjects (such as figures, landscapes, still lifes, etc.) can also be collected, because students' interest in the subject matter may affect their involvement and learning effects when learning related course content. In addition, by introducing feedback information from parents, parents may observe changes in students' enthusiasm for painting in their spare time, problems encountered when practicing at home, etc. This information can be included in the training sample set as supplementary data to provide more references for the model from the perspective of the family learning environment.

[0029] During the feature extraction process, the weights of the features are dynamically adjusted according to different painting stages or teaching focuses. For example, in the basic painting teaching stage, the weights of features such as students' basic painting level and teacher demonstration frequency can be relatively high, because these factors are crucial for students to initially master painting skills. In the creative cultivation stage, the weights of features such as students' emotional state and autonomous learning time can be appropriately increased, because positive emotions and sufficient autonomous exploration time help stimulate students' creativity. For data set partitioning, a fixed 70% training set and 30% validation set ratio is no longer used. Instead, an adaptive partitioning method is used based on the distribution characteristics of the data and the needs of model training. For example, if the source of a batch of data is relatively single (such as all from the same teacher's courses), the proportion of the validation set can be appropriately increased to ensure that the model has better generalization ability and can adapt to the teaching styles of different teachers.

[0030] It is no longer limited to a single classifier (such as support vector machine or decision tree), but an integrated learning method is used to combine multiple different types of classifiers (such as random forest, neural network, etc.) for model training and prediction. Each classifier learns different features and patterns of the data during the training process. By integrating their prediction results, the accuracy and stability of the model can be improved. For example, the random forest classifier can handle nonlinear relationships in the data, and the neural network classifier is good at learning complex feature interactions. The combination of the two can more comprehensively capture the rules in the teaching data of the painting interest class, so as to more accurately predict the learning status and risk factors of students in different teaching situations. In the training process, cross-validation technology is introduced to further optimize the model parameters. The training set is divided into multiple subsets, and each subset is used as a validation set in turn. The remaining subsets are used as training sets for multiple training and validation. Finally, the average value is taken as the evaluation indicator of the model. This can more effectively avoid overfitting problems and improve the generalization ability of the model.

[0031] When determining the target feature data, we should not only focus on simple correlations (such as the correlation between the frequency of teacher demonstrations and the difficulty of students mastering painting skills), but also deeply analyze the causal and temporal relationships between risk factors. For example, through time series analysis, we can study how environmental factors (such as temperature changes) affect students' emotional states over time during the course, and then affect their painting performance. At the same time, we can use causal inference algorithms to determine which factors are the root causes of students' learning difficulties or poor teaching results, and which are indirect factors, so as to formulate teaching adjustment strategies more accurately. After generating the target teaching behavior warning model, establish a continuous mechanism for model evaluation and updating. Collect new teaching data regularly to evaluate the model's prediction accuracy and warning effectiveness. If it is found that the model performs poorly in some new teaching situations or student groups, retrain and optimize the model in a timely manner to ensure that the model can always adapt to various changes and developments in the teaching process of the painting interest class.

[0032] Through the above detailed explanation of the teaching behavior warning model structure and the optimization and improvement of the process, the model can more accurately and comprehensively predict teaching risks, provide more targeted and effective teaching decision support for painting interest classes and other interest training classes (music classes, dance classes, programming classes, etc.), and achieve personalized and precise teaching process optimization.

[0033] S102: Process the target user information to generate physiological parameter information of the target user.

[0034] In one implementation, the target user information is processed to generate initial facial image information and target user identity information, wherein the initial facial image information is the facial information of the target user within a preset time period. Assume that in a comprehensive interest training center, there are multiple interest training classes of different types, such as painting classes, music classes, dance classes, etc. Here, taking the painting class as an example, the student Xiaohong is selected as the target user. A high-definition camera is installed in the painting class classroom, and the camera's shooting range covers the entire classroom to ensure that the facial expressions of each student can be clearly captured. The preset time period is set to the entire course of each painting course, that is, from the ringing of the class bell to the ringing of the get out of class bell. When Xiaohong enters the classroom and starts the class, the camera starts to record her facial image information, and these images are stored in chronological order to form an initial facial image information sequence. When Xiaohong signed up for the painting class, she had registered her personal identity information in the training center's system, including name, age, gender, contact information, and a unique student number in the training center. Through face recognition technology, the facial image of Xiaohong captured by the camera is associated and matched with her identity information to ensure that the facial image data of Xiaohong is accurately identified.

[0035] Process the initial facial image information to generate the influencing factors of the target user's expression characteristics. Use advanced computer vision algorithms to process the initial facial image information. These algorithms can identify key facial feature points, such as the position and morphological changes of eyes, eyebrows, mouths, etc. The algorithms include but are not limited to FasterR-CNN, YOLO (YouOnlyLookOnce) series algorithms (YOLOv5, YOLOv8, etc.), methods based on SIFT (Scale-InvariantFeatureTransform) features and SURF (Speeded-UpRobustFeatures) features, and deep learning-based methods such as AlexNet and ResNet. For example, by calculating the characteristic values ​​of the degree of opening and closing of the eyes, the angle of the eyebrows rising or falling, and the degree of curvature of the mouth, the expression characteristics are quantified. When Xiaohong encounters difficulties in the painting process, she may frown unconsciously. At this time, the algorithm can accurately capture the changes in the eyebrow feature points and calculate the corresponding frowning degree characteristic values.

[0036] Combined with the context of the painting course, analyze the factors that may affect these facial expressions. In the painting class, factors that affect facial expressions may include the degree of interest in the painting theme, the difficulty of mastering painting techniques, and the degree of satisfaction with color matching. For example, when the course theme is landscape painting, which Xiaohong is very interested in, she may be more focused and happy during the painting process, with bright eyes and slightly raised corners of the mouth; when the teacher asks her to use a painting technique that she is not familiar with, she may show a confused expression, such as frowning and wandering eyes.

[0037] The influencing factors of the target user's expression characteristics are processed to generate the target user's emotional information, wherein the target user's emotional information is used to characterize the target user's expression change information. The emotion classification model is trained based on a large amount of expression data of students in the painting class. The model takes the influencing factors of the expression characteristics as input and outputs the corresponding emotion categories. Common emotion categories may include happiness, excitement, concentration, confusion, frustration, boredom, etc. For example, when Xiaohong is painting, her expression characteristics are her eyes staring at the picture intently, her mouth slightly pursed, and her facial muscles relaxed. Combined with the previous analysis of the influencing factors of the expression characteristics (such as interest in the painting theme and the painting progress is smooth), the emotion classification model can judge that Xiaohong's emotion at this time is concentration. Emotional information is stored in a structured manner, including dimensions such as emotion category, emotion intensity, and emotion duration. If Xiaohong has a bright smile on her face, her eyes are narrowed into a slit, and her mouth is wide open when she completes a painting that she is satisfied with, according to the expression characteristics analysis and the emotion classification model, her emotion at this time is happy, the emotion intensity is high, and the time from when she starts to smile to when the smile gradually fades is recorded as the emotion duration.

[0038] The target user's identity information is processed to generate the target user's real-time vital sign information. The trainees are equipped with wearable vital sign monitoring devices, such as smart bracelets or smart watches. These devices are connected to the training center's system and transmit data in real time. The smart bracelet worn by Xiaohong will continuously monitor her heart rate, blood pressure, skin conductivity and other physiological indicators, and bind these data to her identity information. When Xiaohong is in the painting class, the system can obtain her vital sign information in real time. The collected real-time vital sign information is preprocessed to remove outliers and noise interference. Then, according to the characteristics of different painting classes and the general physiological characteristics of the student groups, an analysis model for vital sign data is established. For example, in a painting class, in general, the heart rate of students will remain in a relatively stable range when they are quietly painting. If Xiaohong suddenly encounters a problem during the painting process, it may cause a slight increase in heart rate and a small fluctuation in blood pressure. By analyzing the difference between her real-time vital sign information and the vital sign data under normal painting conditions, we can more accurately understand her physiological state changes during the course.

[0039] Based on the real-time vital sign information of the target user and the influencing factors of the target user's expression characteristics, the physiological parameter information of the target user is generated, and the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes. A physiological parameter fusion model is constructed, and Xiaohong's real-time vital sign information and expression characteristic influencing factors are used as input. The model comprehensively considers the relationship between the two and generates physiological parameter information that can accurately characterize Xiaohong's physiological state in the painting class.

[0040] The physiological parameter fusion model includes an input layer, a hidden layer, and an output layer. Specifically, the input layer is used to input real-time vital sign information and receive real-time vital sign data collected from wearable devices, such as heart rate (HR), blood pressure (BP, including systolic and diastolic pressure), skin conductivity (SC), etc. These data can be directly used as input nodes, and each vital sign data corresponds to an input neuron. For example, the heart rate value can be the number of heart beats per minute, the blood pressure value is in millimeters of mercury (mmHg), and the skin conductivity is in micro-Siemens (μS). Input the data of the influencing factors of the expression feature processed by the expression feature extraction algorithm. These data may include facial expression feature vectors, such as vectors obtained by encoding the position changes of feature points of key parts such as eyes, eyebrows, and mouth. Assume that a 10-dimensional expression feature vector is used, and each dimension represents a quantified value of an expression feature, such as the normalized value of the degree of eye opening and closing, the angle of eyebrow raising, and the degree of mouth curvature.

[0041] The hidden layer includes the first fully connected layer, the feature fusion layer and the second fully connected layer. The first fully connected layer is used to perform preliminary integration and feature transformation on the data of the input layer. This layer contains a certain number of neurons (for example, 50), each of which is connected to all nodes of the input layer. The input data is linearly transformed through the weight matrix and the bias term, and the activation function (such as ReLU, i.e., RectifiedLinearUnit) is used to introduce nonlinear factors to enhance the expression ability of the model. The feature fusion layer is used to specifically fuse the vital sign information and the influencing factors of the expression features. This layer can be implemented in many ways. A simple method is to directly concatenate the output vectors of the fully connected layer 1 to form a longer vector, such as concatenating a 50-dimensional vital sign information-related vector and a 50-dimensional expression feature-related vector into a 100-dimensional vector. Another way is to fuse by weighted summation, assigning different weights to the vital sign information and the influencing factors of the expression features (these weights can be obtained through training and learning), and then summing them to obtain the fused feature representation. The second fully connected layer further abstracts and transforms the fused features. This layer also contains a certain number of neurons (for example, 30) and is fully connected to the feature fusion layer. Continue to use activation functions (such as ReLU or other suitable functions) for nonlinear processing.

[0042] The output layer outputs the final physiological parameter information to characterize the physiological state of the target user. The output form can be defined according to specific needs, such as outputting a comprehensive physiological state score (the value range can be 0 to 100, indicating that the physiological state ranges from bad to good), or outputting multiple physiological state indicators, such as physiological excitement, fatigue, emotional stability, etc. (each indicator can be a normalized value). If a comprehensive score is output, the output layer can be a single neuron, using a suitable activation function (such as the Sigmoid function to map the output value to between 0 and 1, and then multiply it by 100 to get the score). If multiple indicators are output, there can be multiple neurons, each corresponding to one indicator. For example, when Xiaohong's expression shows excitement during the painting process (such as eyes wide open and a big smile), and her heart rate is slightly faster but still within the normal range, and her skin conductivity also increases moderately, the fusion model can judge that Xiaohong is in a physiological state of active involvement and full of creative passion at this time, and can calculate a physiological parameter value representing her physiological excitement based on these data. When Xiaohong participates in different types of interest training classes (such as music classes or dance classes), the same process will be repeated. By comparing her physiological parameter information in different classes such as painting class, music class and dance class, we can find the differences in her physiological state in different course situations. For example, in the dance class, due to the high intensity of the course, Xiaohong's heart rate may rise more than in the painting class, and her facial features will show more fatigue or excitement (depending on the difficulty of the dance moves and her mastery). These physiological state comparison data of different classes can provide valuable reference for the training center to optimize the course settings and adjust the teaching rhythm.

[0043] S103: Process the target class information to generate preset course information.

[0044] In one embodiment, the preset course information is generated based on the interests and hobbies of the target user. Xiao Ming filled out a detailed interest questionnaire when he enrolled. The questionnaire included scores for the degree of interest in different fields (1-5 points, 5 points means very interested), such as 4 points for painting, 3 points for music, 4 points for programming, 2 points for basketball, etc. At the same time, there are open-ended questions asking about his favorite painting style (such as he said he likes the impressionist style and likes its color and light and shadow performance), the application direction of programming, and the areas of interest (such as he mentioned that he is curious about game development). The school organized a series of interest test activities. In the painting interest test, Xiao Ming was asked to freely create a painting based on a given theme. From his works, the evaluators found that he used colors boldly and had certain ideas about composition, but there was still room for improvement in detail description. This shows that he has a certain talent and interest in painting, but needs to further strengthen his skills. In the programming interest test, Xiao Ming completed some simple logic puzzles and small programming tasks. The test results show that he has strong logical thinking ability and can quickly understand the basic concepts of programming, but he still needs to learn about code optimization and algorithm design. Based on the results of the admission questionnaire and interest test activities, Xiao Ming's interests and hobbies are prioritized. Programming and painting are his most interesting and potential areas, followed by music, and he is relatively weak in sports. Therefore, in terms of course arrangement, we will focus on courses related to programming and painting, and also arrange some music courses appropriately to cultivate his comprehensive qualities.

[0045] S104: Process the preset course information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate course adjustment information.

[0046] In one implementation, the preset course information and the physiological parameter information of the target user are processed based on the teaching behavior warning model to generate the target user's emotional information for different courses and the real-time environmental factors of the target class. Assume that in a multifunctional interest training center, there are various interest training classes such as painting, music, and programming. Each class is equipped with corresponding teaching facilities and environmental monitoring equipment, such as light sensors, noise monitors, etc., for real-time collection of environmental factor data.

[0047] Painting course, the course content includes sketching basics, color theory, painting skills practice, etc. The main teaching methods include teacher demonstration, student copying, group discussion and mutual evaluation, etc. The course goal is to enable students to master basic painting skills and cultivate artistic creativity and aesthetic ability. The course has high requirements for light, and the appropriate light intensity range is 500-800 lux to ensure that students can accurately distinguish colors and lines; the requirements for quietness are moderate, and the environmental noise level should be less than 50 decibels to avoid interfering with students' creative ideas; the requirements for space size are relatively flexible, but the per capita usable area should not be less than 2 square meters, so that students can freely stretch out their painting tools and works.

[0048] Music courses cover music theory learning, instrument playing skills training (taking piano as an example), chorus rehearsal and other contents. Teaching methods include teacher explanation, individual guidance, group performance and so on. The course aims to improve students' musical literacy and performance level, and cultivate a sense of rhythm and teamwork ability. Music courses have relatively low requirements for light, 300-500 lux can meet basic teaching needs; but the requirements for quietness are extremely high, and the environmental noise level must be controlled below 30 decibels, otherwise it will affect students' listening to music details and the accuracy of performance; in terms of space size, it needs to be large enough to accommodate musical instruments and student activities. Generally, the per capita usable area is not less than 3 square meters. At the same time, the sound effect must be considered, and the reverberation time of the classroom should be within a certain range (such as 0.8-1.2 seconds).

[0049] Programming courses cover basic programming grammar, algorithm design, project practice, etc. The teaching methods mainly include theoretical explanation, case analysis, and student independent programming practice. The goal is to cultivate students' logical thinking ability and programming skills so that they can complete simple programming projects independently. Programming courses have general requirements for light, 400-600 lux can ensure that students can see the screen and teaching materials clearly; there are certain requirements for quietness, and the ambient noise level should be lower than 40 decibels to help students concentrate on thinking; the size of the space requires that each student has an independent computer operation space, and the per capita usage area is not less than 1.5 square meters, and the classroom layout should be convenient for teachers to patrol and guide.

[0050] Student A (taking a painting course) collected his physiological parameter information through a wearable device during the painting course. His heart rate remained at 70-80 beats / minute during normal painting. When he encountered painting difficulties, his heart rate rose to 85-95 beats / minute. At the same time, his skin conductivity increased slightly, indicating that his emotions fluctuated to a certain extent. The facial expression recognition system detected that when the creation was smooth, Student A's expression was relaxed, his eyes were focused on the picture, and the corners of his mouth were slightly raised; when he encountered difficulties, he frowned and his eyes became a little anxious. Student B (taking a music course) had a relatively stable heart rate of 65-75 beats / minute during the music course, because music has a certain effect of soothing emotions. However, when performing the performance assessment, the heart rate will rise to 80-90 beats / minute, and the breathing rate will increase. In terms of expression, when listening to music, the expression is calm and enjoyable, and when playing, the expression will frown slightly due to concentration, especially when there is a mistake in the performance, the expression will become nervous and the lips will be tightly closed. Student C (taking a programming course) usually has a heart rate of 75-85 beats per minute when thinking about algorithm problems. Once he finds a solution and starts writing code, his heart rate drops to 70-80 beats per minute. Skin conductivity increases significantly when he encounters code errors that are difficult to debug. In terms of facial expressions, he has a firm look when concentrating on programming, but he blinks frequently when he encounters difficulties, and sometimes bites his lips unconsciously.

[0051] The target user's emotional information for different courses and the real-time environmental factors of the target class are processed to generate the synergy value of the target course and the environmental factors; the teaching behavior warning model receives the preset course information and the target user's physiological parameter information and performs analysis and processing. For student A in the painting course, the model judges his emotional state during the painting process based on his heart rate, skin conductivity and expression changes. When the heart rate and skin conductivity increase and the expression is anxious, the model determines that student A has encountered difficulties and the emotion is anxious; when the heart rate is stable and the expression is relaxed, the emotion is calm and happy. At the same time, through the light sensor and noise monitor in the classroom, real-time environmental factor data is obtained, such as the light intensity of 650 lux (within the appropriate range) and the noise level of 45 decibels (slightly above the appropriate range). For student B in the music course, the model analyzes the emotion based on his physiological parameters and expression changes. When listening to music, the stable heart rate and relaxed expression indicate that the emotion is relaxed and enjoyable; the changes in heart rate and breathing rate and the tense expression during the performance assessment indicate that the emotion is nervous. In terms of real-time environmental factors, the light intensity is 400 lux (meets the requirements) and the noise level is 28 decibels (within the appropriate range), but the classroom reverberation time is 1.3 seconds (slightly beyond the optimal range). Student C in the programming course, when student C is programming, the model judges his emotions based on his heart rate and expression. When thinking about the algorithm is difficult, the heart rate increases and the expression changes, and the emotion is confusion; when writing code smoothly, the heart rate state and the focused expression, the emotion is concentration. In terms of environmental factors, the light intensity is 500 lux (appropriate) and the noise level is 38 decibels (good).

[0052] Student A (painting course) was calm and happy in the first 30 minutes after the course started. Then he encountered a problem with the color matching of the painting, and his mood turned to anxiety, which lasted for about 15 minutes. After solving the problem, he returned to calm and happy. During the entire course, the light intensity fluctuated between 600-700 lux, and the average noise level was 45 decibels. Occasionally, the noise would briefly rise to 50 decibels due to the movement of people outside the classroom. Student B (music course) was relaxed and enjoyed most of the course. He was nervous during the performance assessment (about 10 minutes), and gradually returned to a relaxed state after the assessment. The light intensity was basically stable at 380-420 lux, and the noise level was maintained at 25-30 decibels. However, during the assessment, due to the sound of other students playing, the noise level would briefly rise to about 35 decibels. The reverberation time in the classroom felt slightly extended during the performance, which had a certain impact on the clarity of the music. Student C (programming course) was relatively calm during the theoretical explanation at the beginning of the course. After entering the programming practice, he was confused when he encountered algorithm problems (about 20 minutes). After the teacher's guidance, he was focused and focused until the end of the course. The light intensity was maintained at 450-550 lux, and the noise level was 35 dB on average, which was relatively stable.

[0053] The synergy value between the target course and the environmental factors is processed to generate course adjustment information. Consider the relationship between the course and the actual environmental factors such as light, quietness, and space size, as well as the emotional state of student A. Light matching function Because the light intensity is within the appropriate range, Quietness matching function Since the noise level is slightly higher than the appropriate range, according to the formula (Assume the function value calculation method corresponding to the gap). Assume that the space size meets the requirements, and the weight coefficient is 1 (Assume weight distribution). Emotional information will also affect the synergy value. Anxiety may reduce the synergy value. Assume that the synergy value reduction coefficient corresponding to anxiety is 0.8 (can be set according to the model). Assume that the weight coefficients of light, quietness and space size are ω 光线 =0.4,ω 安静 =0.4 and ω 空间 =0.2. According to the synergy value calculation formula Calculated Among them, S represents the synergy value between the target course and the environmental factors, and its value range can be set between 0 and 1, where 0 represents extremely poor synergy and 1 represents excellent synergy. n is the number of dimensions associated with the considered course and environmental factors. For example, it can include the correlation between the course's demand for light and the actual light intensity, the correlation between the course's demand for quietness and the environmental noise level, the correlation between the course's demand for space size and the actual space of the classroom, etc. i is the weight coefficient of the i-th dimension, indicating the relative importance of the dimension in the overall synergy evaluation. The weight coefficient can be determined based on empirical data, expert evaluation, or machine learning training through a large amount of sample data. For example, if it is a painting course, the light dimension may have a higher weight; while for a theoretical lecture course, the weight of quietness may be relatively more prominent. Is a course feature C i Environmental factors i For example, if the matching function between the course's light requirements and the actual light intensity is It can be set as follows: When the actual light intensity is within the appropriate light intensity range for the course, When the light intensity is slightly outside the appropriate range but still acceptable, When the light intensity deviates seriously from the appropriate range, Matching function between the course's quietness requirement and the ambient noise level The function value can be determined based on the difference between the noise decibel value and the quiet decibel threshold required for the course. The smaller the difference, the higher the function value, and the larger the difference, the lower the function value.

[0054] Student B in music class, ray matching function (The light intensity basically meets the requirement but is slightly below the lower limit). Quietness matching function (Noise level is good but reverberation time is slightly exceeded). The space size meets the requirements, and the weight coefficient is 1. Emotional tension has an impact on the synergy value during performance assessment. Assume that the synergy value reduction coefficient corresponding to tension is 0.7. Assume that the weight coefficients of light, quietness and space size are ω 光线 =0.3,ω 安静 =0.5 and ω 空间 = 0.2. Calculate the synergy value (During performance assessment), during non-assessment periods, the synergy value will be higher, assumed to be 0.9 (because of better environmental factors and emotional state).

[0055] Student C in a programming course, ray matching function Quietness matching function The space size meets the requirements, and the weight coefficient is 1. Emotional confusion has a certain impact on the synergy value. Assume that the synergy value reduction coefficient corresponding to the confused emotion is 0.9. Assume that the weight coefficients of light, quietness and space size are ω 光线 =0.3,ω 安静 =0.4 and ω 空间 =0.3. Calculated (period of confusion), the synergy value is assumed to be 1 during normal programming (because the environment and emotional state are good).

[0056] Student A in the painting course has a synergy value of 0.76, indicating that the course and environmental factors have good synergy, but there is still room for improvement. The lighting conditions are good, but the noise level has a certain impact on the students' painting mood and effect. It is necessary to appropriately reduce noise interference to further improve synergy and optimize students' learning experience and painting results. Student B in the music course has a synergy value of 0.83 during the performance assessment, indicating that the matching degree between the environment and the course is acceptable, but the reverberation time and the tension caused by the performance assessment have a certain impact on the overall synergy. The synergy value of 0.9 is high during the non-assessment period, indicating that the course and environmental factors are well matched under normal teaching conditions. Good quiet environment and lighting conditions should be maintained as much as possible. At the same time, some measures can be taken during the assessment to relieve students' nervousness, such as adapting to the examination environment in advance. Student C in the programming course has a high synergy value of 0.97 during the confusion period, indicating that even if the student's mood fluctuates, the environmental factors have little impact on it, and the course and environment match well. A synergy value of 1 during normal programming means that under ideal conditions, the environment setting of the programming course is very suitable for students to learn, and good light, quiet and space conditions should continue to be maintained to support students' efficient programming.

[0057] For the painting course (student A), according to the synergy value calculation results, since the noise level has a great impact on the synergy value, it is recommended to install sound insulation equipment in the painting classroom, such as soundproof curtains or sound-absorbing panels, to reduce noise interference. At the same time, regularly check and maintain the lighting equipment to ensure that the light intensity is always stable within an appropriate range. In view of the situation where student A encounters color matching difficulties and becomes anxious during the painting process, the teacher can add special explanations and exercises on color matching skills to the course, and provide more color matching cases for students to refer to. And when students encounter difficulties, teachers should provide guidance and encouragement in a more timely manner to help students relieve anxiety and improve their confidence in painting.

[0058] For the music course (student B), for the problem of slightly exceeded reverberation time in the music classroom, consider adjusting the interior decoration of the classroom or using acoustic treatment materials to optimize the acoustic effect to make it more in line with the requirements of music teaching. During the performance assessment, some temporary sound insulation measures can be taken to reduce the interference of other students' performance sounds on the students being assessed. At the same time, further stabilize the light intensity to ensure that the light conditions are always appropriate during music appreciation and performance. In order to relieve students' nervousness during the performance assessment, teachers can add psychological counseling sessions before the assessment and teach students some relaxation techniques, such as deep breathing and meditation. In the teaching process, arrange more small concerts or performance opportunities to allow students to gradually adapt to the performance environment and improve their psychological quality. At the same time, adjust the difficulty and progress of the teaching content according to the different emotional states of students when listening to music and playing, and ensure that students learn music in a comfortable state.

[0059] For the programming course (student C), continue to maintain good lighting, quietness and space conditions in the programming classroom, regularly check equipment and facilities, ensure smooth computer operation and stable network, and provide students with a stable programming environment. In response to the confusion of student C when encountering algorithmic problems, teachers can optimize teaching methods and use a more gradual and progressive approach to explain algorithmic knowledge. Provide more online programming resources and learning tools to facilitate students to find solutions independently when they encounter problems. At the same time, encourage cooperative learning among students, establish programming learning groups, jointly discuss and solve problems, and improve students' programming ability and confidence in dealing with difficulties.

[0060] Through the above examples, we can see how to calculate the synergy value and generate course adjustment information based on the teaching behavior early warning model, taking into account the preset course information, target user physiological parameter information, emotional information and real-time environmental factors, so as to optimize the interest training course and improve the teaching quality. In practical applications, these processes can be carried out continuously, and the course and teaching environment can be dynamically adjusted according to the changing situation to better meet the learning needs of students.

[0061] S105, processing the teaching purpose information matched with the target class information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate a course matching influencing factor.

[0062] In one implementation, the teaching purpose information matching the target class information is processed to generate preset course information, wherein the preset course information is used to characterize the course information to be learned by the target user. The teaching purpose of the painting interest class is mainly to cultivate students' artistic literacy, aesthetic ability, creativity and painting skills. Through systematic course learning, students can master the use of different painting tools and materials, understand basic knowledge such as color theory and composition principles, and use the learned skills to express their creativity and emotions, and finally independently create paintings with a certain artistic level. Xiaolin signed up for the intermediate painting class. The course feature of this class is to further study painting skills on the basis of the primary painting class, such as character structure shaping in sketching, advanced color mixing and matching in color painting, and brushstroke application in oil painting. The teaching method pays more attention to practical operation and personalized guidance, arranges fixed classroom practice time and teacher comment sessions every week, and encourages students to communicate and learn from each other. The class size is moderate, generally 15-20 people, so that teachers can pay attention to the learning progress of each student.

[0063] According to the teaching purpose of the painting interest class and the target class information of the intermediate painting class, the preset course information generated for Xiaolin is as follows: Sketching course: In-depth study of human anatomy knowledge, through sketching and copying master works, improve the expressive ability of figure sketching, focus on the use of lines and the shaping of light and dark contrast, the course duration accounts for about 30% of the total course. Color course: Learn more complex color theory, including the emotional expression of color, the harmony and contrast of tones, etc., conduct a large number of color sketching exercises, and try different styles of color expression. The course duration accounts for about 30%. Oil painting course: Master the material characteristics and special techniques of oil painting, such as thick painting method, thin painting method, etc., create oil paintings, express personal creativity and emotions, and the course duration accounts for about 25%. Creative course: Regularly carry out creative theme activities to encourage Xiaolin to use his imagination, use the painting skills he has learned to paint creatively, and cultivate a unique artistic style. The course duration accounts for about 15%. In order to implement the preset courses, the training center has prepared rich teaching resources. The sketching course is equipped with professional sketching paper, pencils (of different hardness), erasers, drawing boards and other tools; the color course provides watercolor paints, oil pastels, palettes, brushes, etc.; the oil painting course has oil paints, canvases, turpentine, oil painting knives and other materials. At the same time, the classroom also displays a large number of painting albums and art books for Xiaolin to appreciate and learn in his spare time. In addition, the training center also invited professional painting teachers who have rich teaching experience and artistic creation practice experience, and can provide high-quality teaching guidance for Xiaolin.

[0064] The preset course information is processed based on the real-time physiological parameter information of the target user to generate course relevance information; the course relevance information is processed to generate a course matching influence factor, wherein the course matching influence factor is used to characterize the physiological parameter information matched by different learning course information of the target user; the method includes a calculation formula for obtaining the course relevance information, and the calculation formula is:

[0065] Among them, X i and Y i are the observed values ​​of the two variables, and is their average value.

[0066] During class, Xiaolin wears wearable devices, such as smart bracelets, which can collect real-time information on Xiaolin's physiological parameters, including heart rate, skin conductivity, body temperature, etc. For example, during normal painting, Xiaolin's heart rate remains at 70-80 beats / minute, skin conductivity is at a relatively stable level, and body temperature is normal. An expression recognition camera is also installed in the classroom to monitor Xiaolin's expression changes through facial expression recognition technology. When Xiaolin encounters difficulties in the painting process, he may frown unconsciously and his expression becomes serious; when he is satisfied with his work or makes progress, he will smile and his eyes will be bright and energetic. These expression changes are also recorded and analyzed as part of the real-time physiological parameter information.

[0067] Define variable X i is the learning difficulty level of different course content modules (such as sketching course, color course, oil painting course, creative course, etc.) in the preset course information (which can be expressed as 1-5, 1 is the easiest and 5 is the most difficult), Y i is a comprehensive index of Xiaolin's real-time physiological parameter information when learning the corresponding course content (it can be obtained by weighted calculation of parameters such as heart rate and skin conductivity, for example, heart rate accounts for 0.6 and skin conductivity accounts for 0.4). In the study of the sketching course, assuming that the learning difficulty level of the course module X1=3, Xiaolin's real-time physiological parameter comprehensive index Y1 in the learning process is calculated based on the collected data. For example, the heart rate is 75 times / minute and the skin conductivity is 10 microSiemens (assuming the normal range is 8-12 microSiemens), then Y1=0.6×(75-70)+0.4×(10-10)=3. Similarly, when learning the color course, X2=4, assuming that Xiaolin's real-time physiological parameter comprehensive index Y2 in this course is 4 (the calculation process is similar).

[0068] Assume that in one stage, Xiaolin studied three modules: sketching, color, and creativity (N=3). ( Average value of (Assuming the Creative Course difficulty level is 2). Calculation ( Average value of (Assume that the comprehensive index of real-time physiological parameters during creative course learning is 3).

[0069] Calculating molecules

[0070] For the sketching course: (3-3)×(3-3.33)=0;

[0071] For the color course: (4-3)×(4-3.33)=0.67;

[0072] For creative courses: (2-3)×(3-3.33)=0.33;

[0073] Then the sum of the numerators is 0+0.67+0.33=1.

[0074] Calculate the denominator: calculate:

[0075] For Sketching Course: (3-3) 2 =0; For color courses: (4-3) 2 =1;

[0076] For creative courses: (2-3) 2 =1; then

[0077] calculate:

[0078] For Sketching Courses: (3-3.33) 2 =0.11; for color courses: (4-3.33) 2 =0.44; for creative courses: (3-3.33) 2 =0.11; then

[0079] The denominator is Final course relevance information

[0080] The calculated course relevance information is about 0.87, which indicates the degree of correlation between the preset course information and Xiaolin's real-time physiological parameter information. In this example, a higher value indicates that the difficulty setting of the course content has a strong correlation with Xiaolin's physiological response during the learning process, that is, the difficulty of the course content can cause Xiaolin's corresponding physiological state changes to a certain extent, which is a manifestation of the course matching influencing factor. If the value is close to 1, it means that the difficulty of the course content is highly matched with Xiaolin's physiological state. The course content can stimulate Xiaolin's interest and enthusiasm in learning (manifested as appropriate changes in physiological parameters, such as moderate increases in heart rate, reasonable fluctuations in skin conductivity, etc.), but will not make him feel too difficult or stressed; if the value is close to 0, it means that the difficulty of the course content is less matched with Xiaolin's physiological state, and the course content may need to be adjusted.

[0081] According to the course matching influencing factors, teachers can make personalized adjustments to Xiaolin's courses. For example, if in the subsequent oil painting course learning, it is found that Xiaolin's real-time physiological parameter information is less correlated with the difficulty of the course content (small value), it may be because the difficulty of the oil painting course is set too high, causing Xiaolin to be too stressed, manifested as a continuous increase in heart rate and unstable skin conductivity. Teachers can appropriately reduce the difficulty of the oil painting course, such as simplifying the teaching of some oil painting techniques, or increasing the time for some basic exercises, to improve the match between the course content and Xiaolin's physiological state, thereby improving the teaching effect. At the same time, teachers can also further optimize teaching methods based on Xiaolin's physiological parameter information such as changes in facial expressions when learning different course contents. For example, in creative courses, encourage Xiaolin to create according to his own interests and expertise to maintain his active learning state and high learning efficiency.

[0082] S106: Process the target user information based on the course adjustment information and the course matching influencing factor to generate course information of the target user.

[0083] In one implementation, the target user information is processed based on the course adjustment information to generate an environmental factor adjustment factor and an emotional value of the target user, and the target user information is processed based on the course matching influencing factor to generate a learning progress value of the target user. In the painting interest class, after a period of teaching observation and data analysis, it was found that some students had difficulty understanding the color theory course, which led to a setback in their learning enthusiasm. Therefore, the school decided to adjust the course, add more example demonstrations and practical operation links of color matching, and slow down the pace of theoretical explanation. At the same time, according to the situation that the classroom light has an impact on color discrimination, it is planned to adjust the lighting arrangement of the classroom to make it closer to natural light. Target user information: Xiaoyang, 12 years old, is passionate about painting, has a certain foundation in painting, is good at sketching, but is relatively weak in the use of colors. He has a cheerful personality, but tends to become impatient when encountering learning difficulties. In class, he usually actively participates in discussions, but sometimes he becomes depressed because he is dissatisfied with his work.

[0084] The school plans to replace the lighting in the painting classroom from the original ordinary white light to full-spectrum lighting that simulates natural light to improve the color reproduction. According to the spectral characteristics and brightness stability of the new light, combined with the requirements of the color theory course for light (such as the color rendering index should reach above 90, the illumination uniformity should be within a certain range, etc.), the environmental factor adjustment factor is calculated. Assuming that through professional measurement and evaluation, the new light has a color rendering index of 95 (full score 100) and good illumination uniformity, and combined with other environmental factors (such as temperature, humidity, etc., which remain within a suitable range), the environmental factor adjustment factor is calculated to be 0.85 (0 means the environment is completely unsuitable, 1 means the environment is very suitable). This means that the adjusted environment has a greater positive impact on the teaching of the color theory course.

[0085] Before the course adjustment, when Xiaoyang was studying the color theory course, he often looked anxious and frowned frequently due to difficulty in understanding. He was also impatient when communicating with his classmates. Through expression recognition technology and classroom behavior analysis, his emotional value was evaluated to be low, assuming it was 0.3 (0 means extremely negative, 1 means extremely positive). After the course adjustment, with the increase of example demonstrations and practical operation links, Xiaoyang was gradually able to keep up with the course progress and began to develop an interest in color theory. He smiled more in class, actively interacted with his classmates, and his emotions became more positive. His emotional value was evaluated again and increased to 0.7.

[0086] For Xiaoyang, his performance in the sketching course has always been good, indicating that the previous sketching course content and teaching methods are highly matched with his physiological and psychological characteristics. Before the color course was adjusted, the match was low, manifested as learning difficulties and negative emotions. After the course was adjusted, based on factors such as his concentration in class (monitored by eye tracking technology, such as average fixation time, distribution of fixation points, etc.), the accuracy of completing exercises (such as the correct proportion of color matching exercises), and the course matching influence factor calculation formula (assuming that the variables in the formula are calculated based on these actual monitoring data), it was concluded that the course matching influence factor has improved. Before the course was adjusted, Xiaoyang made slow progress in mastering color theory knowledge. For example, he often made mistakes in calculating color mixing ratios and understanding color contrast, and his learning progress value was only 0.4 (0 means no progress, 1 means complete mastery). After the course was adjusted, with the improvement of teaching methods and his own mood, his learning progress was significantly accelerated. He was able to correctly calculate the color mixing ratio and use color contrast well in his works, and his learning progress value increased to 0.75.

[0087] The environmental factor adjustment factor, the target user's emotional value and the target user's learning progress value are processed to generate the target user's course information. According to the environmental factor adjustment factor, emotional value and learning progress value, Xiaoyang's course information is further optimized. Since the environmental factor adjustment factor is high, it means that the new classroom lighting environment is conducive to color learning, so more color-related courses can be arranged in this classroom in subsequent courses. Taking into account the changes in Xiaoyang's emotional value and learning progress value, more personalized learning content is customized for him. For example, in the color practice course, some challenging but within his ability to create color creation tasks, such as creating a colorful and harmonious landscape painting based on a given theme, to further improve his ability to use color. Based on Xiaoyang's learning characteristics and current status, the teacher adjusts the teaching method. When explaining complex color theory, more vivid metaphors and examples are used to help Xiaoyang understand better. At the same time, Xiaoyang is encouraged to share his creative ideas and experiences in class to enhance his self-confidence and learning enthusiasm. In view of Xiaoyang's improvement in learning progress, the teacher can provide some extended learning resources, such as recommending related color art books, online tutorials, etc., so that he can continue to study in depth after class. And according to his learning progress, the difficulty and progress of the course can be adjusted in time to ensure that he is always in a positive learning state.

[0088] The school regularly evaluates Xiaoyang's learning situation, including the quality of his paintings (scored in terms of color application, composition, creativity, etc.), classroom performance (participation, cooperation with classmates, etc.), physiological parameters (such as heart rate and skin conductivity during the learning process, monitored by wearable devices to observe the stability of his learning state), etc. According to the evaluation results, the course information is continuously adjusted. If it is found that Xiaoyang encounters a bottleneck in learning at a certain stage, his mood value drops, and his learning progress slows down, the teacher will analyze the reasons in time and may further adjust the teaching content, teaching methods or environmental factors. For example, if it is found that Xiaoyang lacks creativity in color expression, the teacher can organize creative inspiration activities, such as visiting art exhibitions, creative brainstorming, etc., to stimulate his creativity, and at the same time adjust the layout of the classroom to create a more creative learning environment to adapt to Xiaoyang's changing learning needs, to ensure that he can continue to make progress in the painting interest class and have a good learning experience.

[0089] Through the above examples, we can see how to integrate course adjustment information, course matching influencing factors and various information of target users to generate personalized course information, and through continuous monitoring and adjustment, achieve precision and optimization of interest training class teaching, and improve students' learning effect and satisfaction. In practical applications, this method can be extended to other interest classes and students. According to different course characteristics and individual differences of students, various data and technical means can be flexibly used to provide students with the most suitable teaching services.

[0090] S107, processing the course information of the target user to generate teaching behavior information.

[0091] In one implementation, the course information of the target user is processed to generate teaching resource information and teaching activity type information. The painting interest class that Xiaomei signed up for mainly revolves around sketching, color, and creative painting. In the sketching stage, the course goal is to enable students to master the basic use of lines, light and dark contrast, and object structure expression; the color course aims to teach color theory knowledge, such as color matching, tonal harmony, etc., and through practice, students can skillfully use colors to express emotions; creative painting encourages students to use their imagination to create works with personal style. The course content includes theoretical explanations, demonstrations, practical exercises, and work reviews. For example, in the sketching course, the basic concepts of sketching and the use of tools are first explained, then the teacher demonstrates painting, and the students practice copying and sketching. Finally, the teacher comments on the students' works, points out the advantages and disadvantages, and gives suggestions for improvement.

[0092] The entire painting interest class course lasts for 12 weeks, with 2 classes per week and 2 hours per class. Sketching courses are held in the first 4 weeks, color courses in weeks 5-8, and creative painting in the last 4 weeks. The time allocation for each course is roughly: 30 minutes of theoretical explanation, 30 minutes of demonstration, 1 hour of practical exercises, and 20 minutes of work reviews. According to the course content, the required teaching resources include painting tools (sketch paper, pencils, erasers, drawing boards, etc. for sketching courses; watercolor paints, brushes, palettes, etc. for color courses; oil pastels, colored pencils, drawing paper, etc. for creative painting), teaching materials and reference books (such as "Basic Sketching Tutorial" and "Principles of Color Composition", etc.), multimedia materials (painting teaching videos, art work photo libraries, etc.) and gallery space for displaying works (used to display students' excellent works and stimulate learning motivation).

[0093] For example, in the sketching course, students are provided with pencils of different hardness (2H-8B) to meet the needs of different lines and light and shade expressions; the color course is equipped with professional-grade watercolor paints to ensure the vividness and purity of the colors; creative painting provides a variety of rich painting materials, and encourages students to try the combination effects of different materials. The types of teaching activities are diverse to adapt to different course contents and teaching objectives. In the theoretical explanation part, classroom lectures are mainly used. Teachers teach students painting knowledge through PPT presentations and oral explanations. In the demonstration and demonstration session, teachers conduct live painting demonstrations on the podium to allow students to learn painting skills intuitively. During practical exercises, students create independently or in groups according to the tasks assigned by the teacher (for example, in creative painting, sometimes groups are arranged to jointly create a large work to cultivate teamwork ability). In the work review session, teachers comment on students' works one-on-one, and also organize students to evaluate each other to promote communication and learning.

[0094] The target user is processed based on the teaching resource information and the teaching activity type information to generate the target user's emotional information, the target user's learning progress value and the target user's learning time. In the painting course, Xiaomei's emotional information is monitored in a variety of ways. In class, the teacher preliminarily judges Xiaomei's emotional state by observing her expression, eyes and body language. For example, when Xiaomei frowns and looks anxious during the painting process, it may mean that she has encountered difficulties and is in a low mood; when she smiles and focuses, it means that she is in an active learning state. At the same time, wearable devices (such as smart bracelets) are used to monitor Xiaomei's physiological parameters, such as heart rate and skin conductivity. When Xiaomei is emotional (such as when she is very satisfied with her work or encounters great difficulties), her heart rate may increase and her skin conductivity will also change. By analyzing these physiological parameters and combining expression observation, her emotions can be judged more accurately. For example, in a color course, when Xiaomei tried a new color combination, her heart rate increased slightly and her expression was excited, indicating that she was interested in new attempts and had a positive mood.

[0095] The learning progress value is determined based on Xiaomei's task completion and skill mastery at each stage of the course. In the sketching course, the teacher evaluates Xiaomei based on the smoothness of her lines, the contrast between light and dark, and the accuracy of the object structure. If Xiaomei can skillfully use pencils of different hardness to express the three-dimensional sense of the object within the specified time and control the lines properly, then her learning progress value in the sketching course is high, assuming it is 0.8 (0 means no mastery at all, 1 means complete mastery). In the color course, Xiaomei is scored based on the rationality of color matching, the accuracy of color tone use, and the degree of understanding of color emotional expression. If Xiaomei can accurately choose the appropriate color according to the theme and use color contrast to highlight the picture effect, her learning progress value in the color course may be 0.7. In the creative painting stage, Xiaomei's creative novelty, work integrity, and personal style are evaluated, and the learning progress value is determined based on these indicators. The learning time is counted by recording Xiaomei's actual learning time in class and the time for independent practice after class. In class, each class lasts 2 hours, and the classroom sign-in and sign-out system accurately records the time Xiaomei actually spends in the classroom studying, excluding non-study time such as breaks. After class, Xiaomei uses a painting learning app, which can record the length of time she opens and practices painting each time. For example, in a week, in addition to the 4 hours of study time in class, Xiaomei also practices independently for 3 hours after class, so her study time for the week is 7 hours.

[0096] The target user's emotional information, learning progress value and learning time are processed to generate the target user's learning behavior pattern information. According to Xiaomei's emotional information, learning progress value and learning time, her learning behavior pattern is analyzed. If Xiaomei is positive in class, has a fast learning progress and spends a long time on independent practice after class, then her learning behavior pattern can be classified as proactive. For example, in the sketching course, Xiaomei showed a strong interest in sketching, actively asked the teacher questions in class, and took the initiative to find more sketching materials for practice after class. Her learning progress has always been ahead of other students. However, if Xiaomei is emotionally unstable, has slow learning progress and a short learning time at a certain stage of the course, she may have learning difficulties or lack of motivation. For example, in the early stage of the color course, Xiaomei had difficulty understanding color theory, which led to some anxiety, poor concentration in class, and few active practices after class. Her learning progress was obviously behind. This behavior pattern may be regarded as difficult coping type.

[0097] For the proactive behavior pattern, its characteristics include higher enthusiasm for learning (manifested as positive emotions), stronger self-motivation (actively increasing the length of study time), good learning ability (fast learning progress) and positive interactivity (actively communicating with teachers and classmates). For the difficulty coping behavior pattern, its characteristics are lower self-confidence (emotional anxiety), weaker learning strategies (not knowing how to solve difficulties), shorter learning duration (insufficient learning time) and lower participation (lack of concentration in class).

[0098] The learning behavior pattern information of the target user is processed to generate teaching behavior information, which is used to characterize the learning progress and learning time information of the target user. For Xiaomei's proactive behavior pattern, the teacher can provide more challenging learning tasks, such as recommending advanced sketching skills exercises, guiding her to participate in painting competitions, etc., to further stimulate her potential. At the same time, encourage her to share her learning experience in class and drive other students to make progress together. For example, the teacher can arrange for Xiaomei to demonstrate sketching skills in class so that other students can learn her painting methods. For Xiaomei's difficult coping behavior pattern, the teacher needs to pay more attention and support. In teaching, simplify complex theoretical knowledge and use more vivid teaching methods to help her understand color theory. For example, through color games, color case analysis in real life, etc., Xiaomei can be more receptive to knowledge. Emotionally, give more encouragement and affirmation to enhance her self-confidence. For example, when Xiaomei makes a little progress in color matching, praise her in time so that she feels that her efforts are recognized.

[0099] The teaching behavior information is reflected in the personalized teaching plan and classroom teaching strategy formulated by the teacher for Xiaomei. In the personalized teaching plan, the teaching objectives and learning tasks for Xiaomei's different behavior pattern stages are clearly defined. For example, in the proactive stage, the teaching goal is to improve the depth and breadth of painting skills, and the learning tasks include completing difficult paintings and studying the painting styles of masters; in the difficulty coping stage, the teaching goal is to help Xiaomei overcome difficulties and improve her interest in learning. The learning tasks focus on consolidating basic knowledge and simple creative exercises. In terms of classroom teaching strategies, the teaching rhythm and interactive methods are adjusted according to Xiaomei's behavior patterns. For Xiaomei, who is proactive, the teaching rhythm is accelerated and the depth of interactive discussions is increased; for Xiaomei, who is difficult to cope with, the teaching rhythm is slowed down, one-on-one guidance time is increased, and she is guided to gradually overcome difficulties to ensure that her learning progress and learning experience are optimized. At the same time, by regularly evaluating Xiaomei's learning situation, the teaching behavior information is constantly adjusted to adapt to her dynamic changes in the painting learning process.

[0100] This application obtains target user information, target class information, teaching purpose information and corresponding teaching behavior warning models, where the target user may be in different classes at different time periods. Then, the target user information is processed to obtain physiological parameter information, including generating expression feature influencing factors from the initial facial image information, and then deriving emotional information, and obtaining real-time physical sign information from identity information, and combining the two to obtain physiological parameter information. At the same time, the target class information is processed to generate preset course information based on user interests and hobbies. Then, the teaching behavior warning model is used to process the preset course information and physiological parameter information to obtain course adjustment information, which is specifically achieved by generating a synergistic value of courses and environmental factors; the teaching purpose information and physiological parameter information can also be processed to obtain course matching influencing factors. The target user information is processed based on the course adjustment information and course matching influencing factors to generate course information containing environmental factor adjustment factors, emotional values ​​and learning progress values.

[0101] Finally, the target user's course information is processed to obtain teaching resource information and teaching activity type information, and then the user's emotional information, learning progress value and learning time are generated based on this information, and then the learning behavior pattern information is obtained, and finally the teaching behavior information representing the learning progress and time is generated. This method helps interest training classes adjust teaching according to students' situations, improve teaching quality and effect, and realize personalized teaching.

[0102] In one embodiment, if Figure 2 As shown, the present application also provides an intelligent collection device for teaching behavior, including:

[0103] The acquisition module 201 is used to acquire target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information, wherein the target user information includes user information to be monitored, and the users to be monitored are in different classes at different time periods;

[0104] The processing module 202 is used to process the target user information to generate physiological parameter information of the target user, wherein the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes; process the target class information to generate preset course information, wherein the preset course information is generated based on the interests and hobbies of the target user; process the preset course information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate course adjustment information; process the teaching purpose information matched with the target class information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate a course matching influence factor; process the target user information based on the course adjustment information and the course matching influence factor to generate the course information of the target user; process the course information of the target user to generate teaching behavior information, wherein the teaching behavior information is used to characterize the learning progress of the target user and the learning time information of the target user.

[0105] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the protection scope of the present application shall be subject to the scope defined by the claims.

Claims

1. An intelligent collection method for teaching behavior, characterized in that: include: Obtaining target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information, wherein the target user information includes user information to be monitored, and the users to be monitored are respectively in different classes at different time periods; Processing the target user information to generate physiological parameter information of the target user, wherein the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes; Processing the target class information to generate preset course information, wherein the preset course information is generated based on the interests and hobbies of the target user; Processing the preset course information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate course adjustment information; Based on the teaching behavior early warning model, the teaching purpose information matched with the target class information and the physiological parameter information of the target user are processed to generate a course matching influencing factor; Processing the target user information based on the course adjustment information and the course matching influencing factor to generate the course information of the target user; The course information of the target user is processed to generate teaching behavior information, wherein the teaching behavior information is used to characterize the learning progress of the target user and the learning time information of the target user.

2. The method according to claim 1, characterized in that Obtain a teaching behavior warning model that matches the teaching purpose information, including: Acquire a training sample set and a preset teaching behavior early warning model, wherein the training sample set is used to characterize training samples that match the teaching purpose information; Extract features from the training sample set and determine the original feature library; Dividing the original feature library into various feature data sets to generate a training set and a validation set; Based on the classifier, each verification set divided by the original feature library is predicted and processed to generate a prediction result; Based on the preset algorithm, each training set divided by the original feature library is trained and processed to generate the prediction results of the validation set class; Processing the prediction results and the validation set prediction results to generate target feature data, wherein the target feature data is used to characterize risk factors affecting teaching behavior; The preset teaching behavior early warning model is processed based on the training sample set with target feature data to generate a target teaching behavior early warning model.

3. The method according to claim 1, characterized in that Processing the target user information to generate physiological parameter information of the target user includes: Processing the target user information to generate initial facial image information and identity information of the target user, wherein the initial facial image information is facial information of the target user within a preset time period; Processing the initial facial image information to generate expression feature influencing factors of the target user; Processing the influencing factors of the target user's expression characteristics to generate the target user's emotion information, wherein the target user's emotion information is used to represent the target user's expression change information; Processing the identity information of the target user to generate real-time vital sign information of the target user; Based on the real-time vital sign information of the target user and the influencing factors of the expression characteristics of the target user, the physiological parameter information of the target user is generated.

4. The method according to claim 3, characterized in that The preset course information and the physiological parameter information of the target user are processed based on the teaching behavior early warning model to generate course adjustment information, including: Processing the preset course information and the target user's physiological parameter information based on the teaching behavior early warning model to generate the target user's emotional information for different courses and real-time environmental factors of the target class; Processing the target user's emotional information for different courses and the real-time environmental factors of the target class to generate a synergy value between the target course and the environmental factors; Process the synergy values ​​of the target course and environmental factors to generate course adjustment information; The method includes a calculation formula for obtaining the synergy value of the target course and the environmental factors, and the calculation formula is: Among them, S represents the synergy value between the target course and the environmental factors; n is the number of dimensions associated with the considered course and environmental factors; ω i is the weight coefficient of the i-th dimension; Is a course feature C i Environmental factors i The matching function.

5. The method according to claim 4, characterized in that Process the teaching purpose information that matches the target class information to generate course matching influencing factors, including: Processing the teaching purpose information that matches the target class information to generate preset course information, wherein the preset course information is used to represent the course information to be learned by the target user; Processing the preset course information based on the real-time physiological parameter information of the target user to generate course relevance information; Processing the course relevance information to generate a course matching influence factor, wherein the course matching influence factor is used to characterize physiological parameter information matched by different learning course information of the target user; The method includes a calculation formula for obtaining course relevance information, and the calculation formula is: Among them, X i and Y i are the observed values ​​of the two variables, and is their average value.

6. The method according to claim 5, characterized in that The target user information is processed based on the course adjustment information and the course matching influencing factor to generate the course information of the target user, including: Processing the target user information based on the course adjustment information to generate an environmental factor adjustment factor and an emotion value of the target user; Processing the target user information based on the course matching influence factor to generate a learning progress value for the target user; The environmental factor adjustment factor, the target user's emotion value and the target user's learning progress value are processed to generate the target user's course information.

7. The method according to claim 6, characterized in that Processing the course information of the target user to generate teaching behavior information includes: Processing the course information of the target user to generate teaching resource information and teaching activity type information; Processing the target user based on the teaching resource information and the teaching activity type information to generate the target user's emotion information, the target user's learning progress value, and the target user's learning time; Processing the target user's emotion information, the target user's learning progress value, and the target user's learning time to generate the target user's learning behavior pattern information; The learning behavior pattern information of the target user is processed to generate teaching behavior information.

8. An intelligent collection device for teaching behavior, characterized in that: The device comprises: an acquisition module, used to acquire target user information, target class information matching the target user information, teaching purpose information matching the target class information, and a teaching behavior early warning model matching the teaching purpose information, wherein the target user information includes user information to be monitored, and the users to be monitored are respectively in different classes at different time periods; A processing module is used to process the target user information to generate physiological parameter information of the target user, wherein the physiological parameter information of the target user is used to characterize the physiological state of the target user in different classes; process the target class information to generate preset course information, wherein the preset course information is generated based on the interests and hobbies of the target user; process the preset course information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate course adjustment information; process the teaching purpose information matched with the target class information and the physiological parameter information of the target user based on the teaching behavior early warning model to generate a course matching influence factor; process the target user information based on the course adjustment information and the course matching influence factor to generate the course information of the target user; process the course information of the target user to generate teaching behavior information, wherein the teaching behavior information is used to characterize the learning progress of the target user and the learning time information of the target user.

9. An electronic device, characterized in that: include: a first processor; and a memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the intelligent collection method of teaching behavior described in any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the second processor, the intelligent collection method of teaching behavior described in any one of claims 1 to 7 is implemented.

Images