Student offline classroom concentration identification research based on deep learning

Abstract

During the reform of the deep teaching model, students’ deep learning quality was affected and restricted by various factors. During the offline class learning process of students, the concentration of deep learning directly affects the quality of learning. This article analyzes the study focus of students in deep learning models, conducts research on the quality of class offline learning of different students, quantifies the factors that affect students’ deep learning, and builds an analysis model for quantitative comparison. Important influence factor affecting students’ offline classroom concentration, through targeted measures, improve teaching methods and quality, optimize classroom teaching models, use various methods and measures to effectively improve learning focus, and further promote the reform of teaching models. The level of concentration of students’ learning has been steadily improved, and the model of deep learning is proposed to help the teaching model reform.

Keywords

Deep learning offline learning classroom concentration identification measures

1. Introduction

With the rise of artificial intelligence technology, the application of deep learning analysis model is more and more extensive. Deep learning is a kind of machine learning algorithm, which can be analyzed and detected by intelligent means and methods that are difficult to observe by human eyes. In recent years, deep learning has been applied to identify students’ offline classroom focus. Deep learning, as an important means of analyzing students’ offline learning focus at this stage, is an efficient and high-quality analysis method and plays an important role in schools and education and training institutions. By analyzing students’ attention, deep learning can find out students’ problems and improve the quality of learning by widely applying various teaching resources. Deep learning, as an important means of reform of students’ learning model at this stage, is an efficient and high-quality learning method. It plays an important role in schools and education and training institutions. Deep learning can lead and drive students to improve the quality of learning by wide application of various teaching resources. Under the deepening development of artificial intelligence technology, by accelerating the application of deep learning models, expanding the reform of teaching models, creating a deep learning platform for intelligent teaching, and achieving a strong improvement of the quality of classrooms and offline teaching. As the main body of learning activities, students can better realize their mastery of teaching knowledge content by relying on the tutoring of teachers and the comprehensive application of teaching resources. The concentration of students’ learning is an important factor affecting the learning efficiency and learning quality. In the research process of the concentration of students’ offline classroom learning, the different behavioral performance and the efficiency of classroom interaction in class line learning and the efficiency of classroom interaction are used. Studies can better discover the influence factors of students’ concentration levels [1, 2]. This article conducts in – depth research on classroom teaching activities to break through the factors that affect students’ learning concentration. By adopting measures to improve the teaching model, the utilization rate of teaching resources, and in – depth discussions to effectively improve students’ classroom concentration.

At present, scholars have studied the learning focus of primary and secondary school students, and made suggestions for primary and secondary school students’ learning by using the conclusions of deep learning analysis. However, deep learning technology has not made great achievements due to the short time of entering the field of offline students’ learning focus. In the process of exploring the factors that affect students’ deep learning and focus levels, comprehensively analyzes the multi-faceted factors that affect students’ deep learning and concentration influencing factors by constructing convolutional neural networks. An isolation analysis of the impact factor and the degree of influence, so that different influence factors can be more accurate analysis, and formulate more effective avoidance measures [3].

In this paper, the deep learning method refers to a method system that allows students to enter the deep learning state, and the deep learning of students refers to the learning state of students. There are essential differences between the two.

2. Offline concentration of students in deep learning mode

2.1 Student offline concentration typical behavior

In this paper, the deep learning model refers to the use of deep learning methods of machine learning to analyze the factors that affect student learning. This chapter analyzes the typical behavior of students’ offline learning through survey data.

In the deep learning mode, during the process of offline learning, there are various forms on the level of focus. After expert interviews and field surveys, it is found that 7 typical behaviors are important indicators to reflect the level of classroom concentration. The behaviors are: listening to the lesson, looking around, raising your hands to answer questions, sleeping, standing, reading, and writing. These behaviors not only reflect the basic behaviors of students in offline classroom learning, but also constitute an important influencing factor in learning quality in the process of learning under the classroom [4, 5]. For students of these typical scenes, we can effectively identify and classify, which can effectively grasp the quality of students’ deep learning, and play an important role in improving learning concentration.

A total of 300 students were selected from a middle school in this survey. They were selected in different grades according to the proportion of grades, and randomly selected in the same grade according to the number of samples needed. Then, in the case of 300 students who do not know, the offline classroom learning focus data acquisition. Students have collected and decomposed classroom behaviors, effectively analyzed the influence factors and deep learning quality of classroom learning concentration, and analyzed gender, dress posture, and learning environment factors, so that students’ classroom behavior is more fitting teaching goals, and the quality of learning has gradually improved. In Table 1, people with different learning status analyzes the learning status from three aspects: emotional cognition, thinking cognition, and action cognition [6].

Table 1
The study status of people at different learning status

Learning status		Study	Basically meet the requirements	Deep learning
Emotional	Avoid punishment	Avoid punishment	Over-done	Over-done
cognition	Avoid learning	Avoid learning	High interest	High interest
Thinking	Hate learning	Hate learning	There are clear plans for learning	There are clear plans for learning
	Unsustainable	Unsustainable	Very autonomous	Very autonomous
Cognition	Without active learning	Without active learning	Learn a lot of time	Learn a lot of time
	Do whatever you want	Do whatever you want	Strong self-management ability	Strong self-management ability

2.2 Concentration based on deep learning

After conducting actual research on the learning behavior and status of the 300 students, the influencing factors of classroom teaching behaviors were obtained in the education and teaching stage. In the classroom, teachers, as an important influencing factors as leading classroom learning progress, in the process of guiding students’ offline classroom deep learning process, by adopting a variety of learning forms and educational models, further attracting students, improving their interest in learning, and enabling steady improvement. Through the questionnaire survey of the 300 students, we can find that the intelligentization of classrooms and the plot of the story can greatly improve students’ interest in classroom learning. In terms of improving learning concentration and improving the quality of classroom teaching, it plays an important role. By constructing a platform for typical scenes, communication, and interaction, students’ learning is more comprehensive, objective and scientific, and the intelligence level of education also affects the quality and efficiency of deep learning. The influence factor of students’ concentration is multifaceted. It is necessary to improve the teaching mode and the teaching level of the teacher itself, so that the teacher’s personal charm deeply attracts students, and further improves students’ learning quality and interest in classrooms.

2.3 Student offline classroom learning behavior test

By using information technology to conduct real-time monitoring and monitoring of the quality and behavior of students’ classroom learning, to effectively reflect the influencing factors that affect students’ learning behavior and concentration, and then analyze differently through graphic dataization to analyze different personalities, different genders, different ages, and students in different teaching environments in deep learning quality, and then find concentration influence factors. By analyzing the classroom data of 300 students and the quality of class concentration, combined with the results of the follow-up test after the classroom teaching, the quality of students’ deep learning and the practical application of knowledge. Establishing classroom behavioral image data processing libraries, providing effective measures for students who have low concentration of learning in the later period, provide important technical support. Building such a graphic database can make accurate matching in classroom teaching, and find that students are not concentrated. Teachers can get prompts as soon as possible to make timely reminders to effectively improve students’ learning quality.

2.4 Technical factors affecting students’ deep learning

In the process of analyzing the technical factors affecting students’ deep learning, we found that by focusing on different classroom behaviors of students and intelligent identification matching, we can accurately evaluate students’ learning quality and learning efficiency. Aiming at some frequent behaviors that affect classroom concentration, timely rectification and correction can effectively improve the quality of learning. In the process of information technology application and classroom behavior recognition, the database of absorbing and expanding students’ learning behavior can make students’ learning monitoring more comprehensive. At the same time, the use of intelligent technical means to build a more scientific evaluation and analysis method, so that students’ learning quality and classroom behavior can be effectively monitored. Belly established in a real classroom environment, evaluate the quality of students’ learning, and use the deep convolutional neural network evaluation model to scientifically evaluate the quality of students’ learning to effectively improve students’ learning concentration. In the usual learning, focusing on the use of training and classroom teaching models can be cultivated. In actual teaching, it can steadily improve students’ offline classroom learning concentration and improve the quality of deep learning [7, 8]. In addition, in the process of student learning behavior correction, we must pay attention to methods and correct guidance to prevent students from reverse psychology. In addition, in the comprehensive monitoring process of data processing and student learning behavior, we must make full use of technical means and the dual role of school families, so that students’ overall learning quality and learning efficiency have been steadily improved.

3. In deep learning mode, students offline classroom concentration influence analysis

3.1 Concentration impact factor

By analyzing the influencing factors that affect students’ deep learning, the use of information technology means, analysis of voice recognition image analysis and processing, comprehensive database construction, etc., in deep learning models, the influencing factors of students’ offline classroom concentration. In Fig. 1, we investigated the quality of student training in domestic extracurricular training institutions. In terms of support for extra-curricular training, general parents believe that extracurricular training and offline learning have an important role in promoting the quality of learning. In the survey of extra-curricular learning models, 45% of parents believe that deep learning through the combination of class and offline models can further improve the quality of learning. And in the deep learning mode, in order to effectively improve the quality and efficiency of learning, through the establishment of an information management platform, it provides an important management basis for students’ classroom learning behavior. Under the reform of the new teaching model, students’ learning is closer to the needs of realistic teaching, and students’ offline classroom learning concentration is enhanced [9, 10].

Figure 1.

Parents’ attitude towards online and offline classroom integration teaching model.

Through the application of network model-based mathematical process monitoring and management platform, students’ learning and concentration levels have been steadily improved. In the process of system construction, by building a comprehensive learning management resource library, the quality and efficiency of students’ learning have been steadily improved. While improving the method of learning efficiency of students, this article is different from students who learn offline learning. Analysis of behaviors, different languages, and concentration levels, as well as actual effects to investigate, use convolutional neural network evaluation and analysis methods to comprehensively evaluate the quality efficiency of learning and test results. For multiple influencing factors affecting students’ classroom concentration, comprehensive analysis and influencing factors are performed. Let deep learning become an important means for students to improve the quality of learning, improve learning efficiency, and effectively promote students’ learning concentration. After summing up a large number of classroom behaviors, the monitoring platform and database obtained are more targeted by matching and accurate control of database resources. Teachers in classrooms can be targeted. Problem students should treat timely treatment [11, 12]. In Fig. 2, the child’s participation in extra-curricular training was conducted, and parental satisfaction was investigated. The results showed that most of them were more agreed. The application of information technology, the analysis of speech recognition image processing and the construction of comprehensive database are all factors to promote the development of deep learning model. Through information technology, massive data can be obtained to enhance the analysis ability of deep learning platform. Using databases, etc., can help the deep learning platform to establish its own local database. The speech recognition image analysis technology department analyzes the learning status of children by identifying the behavior of students.

Figure 2.

The survey results of children participating in extra-curricular training.

3.2 The way to improve concentration

In the process of improving the depth of learning, by strengthening the concentration of students’ learning, it can efficiently improve the efficiency of learning. In the process of using technical means for student class behavior monitoring and identification, the students’ learning status is effective through comparative analysis of massive data. Monitoring, propose models and methods of concentration correction. On the basis of building a student learning concentration analysis model, use targeted methods. On the one hand, let the teachers have measures for real-time monitoring and adjustment of students’ learning status, and timely point out the problems of students in the classroom. Through the establishment of a learning platform, the monitoring system fully plays a role. By contacting each desk and the corresponding learning location, students’ learning can improve efficiency and focus under the monitoring of the information system. In the process of building a student concentration management platform, comprehensively analyzes the actual assessment effect of learning with the level of concentration of students, so that students’ learning methods and quality can be quantified, so that students’ learning focus is steadily improved. For example, the steps of the AI intelligent class offline learning auxiliary detection system running and the auxiliary methods of each learning stage are analyzed.

Table 2
AI intelligent class offline learning auxiliary detection system running

Number	Learning phase	Learn auxiliary means
1	Preschool test	Specified learning plan
2	Pre-class preview	Preview textbook knowledge
3	Classroom study	Online and offline classroom teaching
4	Post-class review	Teaching Resources Application
5	Consolidate practice	Exercise test
6	Expand learning	AI intelligent expansion
7	Growth report	Exercise, classroom, teacher feedback
8	Phase evaluation	Staged general review

3.3 Use deep learning to improve concentration

The quality and level of deep learning are an important manifestations of the concentration of learning in students. In the actual model construction process, through analysis of classic deep learning network models, we can find that the identification of students’ offline learning behavior can obtain more accurate predictions and prevention through the monitoring of network models. For the quality collection of students’ offline learning, study the quality of curriculum learning under different classroom behavior habits, process data processing of students’ learning behavior through the network model, and obtain the impact of different learning environments and different courses on students’ learning concentration. As a training model, students can effectively improve the quality of students’ learning. In addition, by quantifying the students’ deep learning effects and linked to their learning concentration, they can more accurately discover important impact factors affecting students’ offline deep learning concentration.

As shown in Fig. 3, the majority of the 300 students we surveyed were ‘listening carefully’, ‘looking forward to the right’ and ‘raising hands to interact’. ‘Going to sleep’, ‘reading extracurricular books’ and ‘writing other homework’ are the minority. The former belongs to students with good learning status, and the latter belongs to students with poor learning status. The results show that students’ learning state directly affects students’ learning quality [13, 14].

Figure 3.

The proportion of different students in the classroom.

4. Student concentration analysis model based on key points of bone key points

In this paper, the model is constructed by convolutional neural network method, and the problem is found by analyzing students’ learning behavior. Based on deep learning technology and other information technologies, a deep learning platform is constructed. At the same time, the students’ concentration is analyzed by the bone key point detection algorithm. Skeleton point detection algorithm is a top-down algorithm, that is, to detect inverted human body, and then get the key points and skeleton. Its advantage is that the key points of the occluded part will not be arbitrarily obtained (that is, only the visible part can be displayed). Its accuracy and AP value are higher than open-pose, which can be used to effectively detect student concentration analysis. But it also has shortcomings. As the number of people on the picture increases, his calculation increases and his speed slows down. However, the use of bone point detection algorithm in this analysis has certain advantages, and the results are satisfactory.

In this experiment, SPSS analysis shows that the questionnaire results have good reliability and validity. This shows that the questionnaire has high reliability. At the same time, the object of this study is students’ learning behavior. Research indicators are students’ classroom behavior.

4.1 Model construction basis

Taking the important influence of students’ deep learning as a quantitative analysis indicator. This article uses convolutional neural networks to build algorithm models on deep learning influence factors, and analyzes the influencing factors of students’ offline learning concentration from multiple aspects and multiple perspectives. Based on deep learning analysis models to achieve the level of concentration of different students, different personalities, different genders, and different learning efficiency in the classroom, and achieve a comprehensive evaluation of the overall learning concentration. By using the human morphology identification model, compare the learning concentration of different students, so that the impact factor of the students’ learning quality and concentration level is fully demonstrated. Through input images for comparative analysis, build an analysis model of convolutional neural network extraction image. The state and form of learning are displayed in the analysis process through quantitative analysis methods. At the same time, comprehensive research on the study concentration of different time nodes, different stages and different teaching content inside the classroom, and analyzing the influencing factors that affect the course resources that affect students’ learning concentration [15, 16]. Comprehensive teaching resources and students’ learning status and analysis of students, build a more reasonable teaching model, so that students’ learning quality is based on their excellent learning habits and state of learning, but also rely on schools to provide more reasonable and scientific teaching resources and teaching. model. In Table 3, the proportion of students’ offline learning status and concentration in Table 3 investigations.

Table 3
Student class line learning status and concentration proportion

Number	Classroom learning status	Class learning concentration	Proportion
1	Full-time authentication learning	More than 80%	34.15%
2	Occasionally	60%–80%	63.53%
3	Basic hang-up do things that have nothing to do with classroom teaching	30–59%	1.85%
4	Completely hanging up, not learning	Below 30%	0.46%

4.2 Parameter collection and calculation

By comprehensively judging the concentration of a student in the deep learning of the classroom by answering questions in classrooms and completing the follow-up test in the classroom. During the data collection process, we applied the analysis model of convolutional neural networks to comprehensively collect several factor that affects the learning status of students. During the data collection process of students’ learning forms, we use the comprehensive form of various parts of the human body as the standard for collection. By comparing different states in the database, the corresponding learning quality and learning efficiency standards in the database match level and timely rectification. After the construction of the above model, we will influence the multiple factors that affect the learning of students as an indicator of analysis to enter the comprehensive analysis model of students’ learning as follows:

$\displaystyle S(x)=\exp\left(-\frac{\left\|{x-\alpha}\right\|_{i}^{n}}{\alpha^% {2}}\right)$ (1) $\displaystyle S(x)=\max S_{i}^{n}(x)$ (2)

Equations (1) and (2) are the processing of experimental results. Where ${x}$ represents different factors, ${\alpha}$ represents different multiple factor analysis standard constants. ${S}({x})$ is the level of influence of a factor on students’ concentration. The influence level of a factor can be known from Eq. (1). The maximum impact factor can be obtained from Eq. (2). After the operation and processing of the above evaluation model, we have a comprehensive evaluation and conclusions for students’ learning status in the classroom. On the other hand, we must analyze the matching and teaching model of teaching resources in the classroom. The degree of interest in learning teaching resources in different states and comprehensive interest in teaching resources to judge the level of concentration of students’ learning. By constructing a convolutional neural network analysis model, the influence factor is entered into the model, and the influencing factors that comprehensively affect students’ deep learning concentration have been obtained. Factors, compare two or two, comprehensively analyze the following evaluation model for calculation:

$\displaystyle E=\int_{n=0}^{n=1}{L_{n}}(S(x))\cdot\frac{d_{i1}-d_{i2}}{\left\|% {d_{i1}-d_{i2}}\right\|}d_{i}$ (3) $\displaystyle S(x)=(1-x)d_{i1}+xd_{i2}$ (4)

In Eqs (3) and (4), $d_{i1}$ and $d_{i2}$ represent different important influence factors. $D_{i}$ is the weight of this factor, and $S(x)$ is the influence level of a factor on students’ concentration. $L_{n}$ represents the weight of different factors. $E$ represents expectation. $X$ represents different factors. Equation (3) calculates the expectation of the results, and Eq. (4) calculates the comprehensive evaluation results of the results. After a comprehensive analysis of the above-mentioned influence factors, we can draw. On the one hand, due to the different learning habits of students, different learning statuses, and their respective concentration levels, students are in the process of deep learning. The degree of knowledge in teaching is different, which will affect the quality of students’ teaching in class. On the other hand, due to the difference in teaching resources and the different teaching model, the attractiveness and interest of students are different, which has led to different concentration in the course of class line learning [17, 18]. Comprehensive influence factor in two aspects, through the analysis of convolutional neural networks, we have concluded that if a student is based on its excellent learning habits and learning status, it is based on the more scientific and efficient teaching resources set up by schools and education institutions. Application, fully attract students’ information learning interest, steadily improved the level of learning status and deep learning concentration, can more efficiently promote and promote the reform of teaching models, so that students’ deep learning quality has been steadily improved.

4.3 Detection of student concentration results

The convolutional neural network uses fewer weights and bias parameters in the convolutional layer, and has translation invariance. That is to say, by moving the research content to another location, the convolutional neural network can still identify this target well, and the output results are consistent with the results before the movement. This feature of convolutional neural network well meets the needs of this study for research methods, so this study selected convolutional neural network method.

Through the construction of quantitative analysis model of convolutional neural network, we find that the influence factors of students’ learning focus affect students’ comprehensive learning effect from different angles. Using the results of the analysis to take targeted measures can effectively improve the quality of students’ learning and attention levels. Based on the construction of the quantitative analysis model of the above convolutional neural network, we found that the impact factor affecting students’ learning concentration, from two aspects, affect the comprehensive learning effect of students from different angles. Through the two aspects of learning influence factors, targeted measures can effectively improve the quality and concentration level of students’ learning. On the one hand, in the process of the impact factor of student learning status and personal learning efficiency, it is necessary to improve the state of learning, and to implement the learning status of students’ learning status, real-time monitoring of learning, and learning concentration by building a quantitative analysis model and networked monitoring management platform. Remind that students can achieve steadily improvement of individual learning ability under the dual action of self-restraint and information network monitoring, so that in the process of personal learning, good behavior habits are formed under long-term supervision and constraints [19, 20].

On the other hand, schools and education departments should reform the teaching resources and teaching models, fully integrate and use a variety of teaching resources, and through various forms of teaching, such as teaching interaction, further enhance students’ learning interests, and allow students to focus on their learning. Improve important measures and means. And in the process of teaching model reform, the teaching content of different stages and classrooms of students’ learning efficiency is made reasonable planning, and different difficulties, different stories, and different attractive knowledge points will be made reasonable planning. Let students learn from scientific teaching models, and improve the depth of learning and concentration.

5. Conclusion

This article conducts a comprehensive investigation of the students’ learning status in the influencing factors of the students’ offline teaching concentration in deep learning models, and conducts comprehensive analysis and collection of two factors that affect the level of concentration of students, and build convolutional nerves. The network analysis model is quantified for the risk factors of the two aspects, and targeted measures are proposed to improve students’ learning status, improve the utilization of teaching resources, and promote the reform of teaching models. In the process of overall resource application and student learning in depth, we must pay attention to the quality of students’ learning, not blindly take measures for learning, and fully pay attention to the healthy growth of students’ physical and psychological, so that learning is based on a more positive environment, students’ depth and concentration level. Let students build a good learning habit from the school days, laying a good foundation for future growth.

In this paper, in the process of analyzing the influence factor model based on neural network, through the comprehensive analysis of the influence factors of different aspects and different angles, the size of the influence factors is quantified, so that the students’ concentration and deep learning influence factors can be obtained. It is of great significance to improve students’ learning efficiency and the development of education industry.

During the integration and improvement of learning efficiency throughout the learning process, teachers and education departments played a key role, and provided good guidance and traction for students to learn. To build a good learning environment and teaching resources, in the innovation of teaching models, make full use of students’ interest in teaching resources to improve the quality and efficiency of learning in all aspects.

In summary, deep learning has great advantages in the study of students’ offline learning state, which can capture the details that traditional research methods are difficult to find. In this study, through model construction, deep learning analysis technology is used to find the influencing factors of students’ attention, so as to take targeted measures, which is of great help to the development of students’ learning and the education industry.

Footnotes

Fund

This work supported by General Project of Hunan Provincial Social Science Achievement Evaluation Committee. project number: XSP22YBC431.

References

Wang

Feng

. The basic connotation, value implication and practical path of deep learning. Western Qual Edu. 2022; 8(14): 30-32.

Ding

. A study on the intelligent classroom teaching model of college English audio-visual speaking from the perspective of deep learning. China Med Edu Technol. 2022; 36(4): 438-441+448.

Yuan

Guo

. Research on the practice path of intelligent education in colleges and universities with deep learning as the core. J Hebei Norm Univ (Edu Sci Ed). 2022; 24(4): 68-74.

Shi

. Research on multi-factor short-term load forecasting based on AdaBelief optimized deep learning model. Shaanxi University of Technology; 2022.

Zhu

. Research on high school mathematics modeling teaching based on deep learning. Shanghai Normal University; 2022.

Chen

. Research on apple flower recognition detection method based on deep learning. Shandong Agricultural University; 2022.

Liu

. Intensive classroom exercises to promote deep learning. Huaxia Teacher. 2022; (8): 59-60.

Zhao

. Research and application of automatic detection of Thangka elements based on deep learning. Qinghai Normal University; 2022.

Sun

. Effect evaluation and attribution analysis of online classroom learning. Integr Circuit Appl. 2022; 39(7): 88-89.

10.

Wang

. Comparative analysis of simple online teaching and online and offline mixed teaching effects. China Mod Edu Equip. 2022; (5): 75-78.

11.

Yan

. Urban-rural differences and impact analysis of college students’ online learning. J Taiyuan City Polytech. 2022; (1): 46-51.

12.

Xue

Chen

Zhou

Song

. Design of teacher teaching quality evaluation index system under mixed teaching mode. Comput Knowl Technol. 2021; 17(30): 253-255.

13.

Guo

. Design and implementation of attention detection system based on students’ learning process. Beijing University of Posts and Telecommunications; 2021.

14.

Liu

. Concentration recognition based on multi-feature fusion. Guangdong University of Technology; 2021.

15.

. Research on students’ online learning engagement based on multi-dimensional information fusion. Huazhong Normal University; 2020.

16.

Xin

. Concentration analysis and application research of junior high school students based on mirror system. Tianjin University; 2019.

17.

Shu

Zhou

. Design of classroom focus based on face recognition. Mod Inf Technol. 2021; 5(12): 29-32.

18.

Zhai

. Research on emotional computing of online education students based on rPPG. Beijing University of Posts and Telecommunications; 2021.

19.

Sun

Liu

. Application research of attention recognition. Sci Technol Inf. 2021; 19(14): 30-32.

20.

Zhang

. A survey on intercultural sensitivity of English majors in Chinese-Foreign Cooperative Universities. J Shanxi Norm Univ (Soc Sci Ed). 2014; 41(3): 158-160.