A data management system based on a live broadcast interaction platform

Abstract

The application discloses a data management system based on a live broadcast interaction platform, which comprises a data acquisition module, an interaction management module and a feedback module. The data acquisition module is used for collecting real-time barrage data, comprehensive data of students watching live broadcast and recorded broadcast on a student end, and question data of students, collecting visual images and environmental comprehensive data of students watching live broadcast during live broadcast teaching, and inputting comprehensive data of historical live broadcast into a system. The interaction management module is used for training an identification model by using the historical live broadcast comprehensive data, identifying and counting live broadcast barrages according to the identification model, analyzing visual images of students watching live broadcast, analyzing iris data of students and environmental data of students watching live broadcast, evaluating learning states of students according to analysis results, analyzing student barrages and watching records of students on live broadcast recording screens, and obtaining knowledge points of students. The data management system has the characteristics of high accuracy and strong interaction.

Description

Technical Field

[0001] This invention relates to the field of live-streaming teaching interaction technology, specifically a data management system based on a live-streaming interactive platform. Background Technology

[0002] With the rapid development of internet technology, live-streaming teaching has become an emerging educational method. It breaks through the time and space limitations of traditional teaching, allowing students to receive education anytime, anywhere.

[0003] Currently, most online education is conducted via live streaming or pre-recorded videos. However, existing technologies only identify whether students are facing the camera, making it difficult to accurately determine whether students are paying attention or assessing their engagement. This results in poor teaching effectiveness. Furthermore, current live streaming teaching has relatively weak interactivity, especially with a large number of students and numerous comments, making it difficult for teachers to accurately read students' questions. This leads to insufficient communication between students and teachers and poor interactive teaching. Therefore, it is essential to design a data management system based on a live streaming interactive platform that offers high accuracy and strong interactivity. Summary of the Invention

[0004] The purpose of this invention is to provide a data management system based on a live streaming interactive platform to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a data management system based on a live interactive platform, comprising a data acquisition module, an interaction management module, and a feedback module;

[0006] The data acquisition module is used to collect barrage data, comprehensive data of students watching live and recorded broadcasts on the student's end, and data of students' questions in real time. It also collects comprehensive data of visual images and environment of students watching live broadcasts during live teaching and inputs comprehensive data of historical live broadcasts into the system.

[0007] The interactive management module is used to train a recognition model using historical live streaming data, to identify and statistically analyze live streaming comments based on the recognition model, to analyze the visual images of students watching the live stream, to analyze the students' iris data and environmental data when students watch the live stream, to evaluate the students' learning status based on the analysis results, to analyze the students' comments and the students' viewing records of the live stream recordings, to obtain the knowledge points that the students are confused about, to match materials for the students based on the above knowledge points, and to analyze and integrate the characteristics of the students' comments.

[0008] The data management module is used to integrate students' bullet screen data and live broadcast viewing data and transmit them to the teacher's end. Students can establish a connection with the teacher's end through the student end, provide feedback and interact with the teacher.

[0009] According to the above technical solution, the data acquisition module includes a data collection module, a sensor module, and a historical live broadcast comprehensive data entry module. The data collection module is used to collect real-time barrage data and comprehensive data of students watching live broadcasts and recorded broadcasts on the student's end. The sensor module is used to collect visual images and comprehensive environmental information of students watching live broadcasts during live teaching. The historical live broadcast comprehensive data entry module is used to enter the historical live broadcast comprehensive data of various subjects into the system.

[0010] According to the above technical solution, the interactive management module includes a live streaming supervision module and a live streaming interaction module;

[0011] The live streaming monitoring module further includes an iris analysis submodule, a status assessment submodule, and an environmental analysis submodule. The iris analysis submodule is used to identify the iris position of students when watching the live stream and further analyze the impact of students' iris fixation deviation on students' status. The environmental analysis submodule is used to analyze the impact of comprehensive environmental data on students' status. The status assessment submodule is used to assess students' status based on the above analysis results.

[0012] The after-class interaction module further includes an operation analysis submodule, a knowledge matching submodule, and a model training submodule. The operation analysis submodule is used to analyze the time characteristics of students watching live broadcasts and screen recordings. The knowledge matching submodule is used to mark the live broadcast and screen recording videos according to the above time characteristics, further analyze the marked videos and extract knowledge point features, and further match the corresponding teaching videos according to the knowledge point features. The model training submodule is used to train a recognition model based on historical live broadcast comprehensive data, use the above recognition model to identify bullet screen features and student question features, and match and integrate knowledge points according to the above bullet screen features and student question features.

[0013] According to the above technical solution, the data management module includes a summary and feedback module, which is used to feed back the knowledge points corresponding to the bullet screen features and the knowledge points corresponding to the students watching the marked videos to the teacher.

[0014] According to the above technical solution, the data management module also includes a question feedback module, which is used to feed back students' question data to the system.

[0015] According to the above technical solution, the operation method of the data management system of the live streaming interactive platform mainly includes the following steps:

[0016] Step S1: Collect real-time data on bullet comments, combined data on students watching live and recorded broadcasts on their devices, and data on students' questions. Collect visual images and environmental data of students watching live broadcasts during live teaching, and input the combined data of historical live broadcasts into the system.

[0017] Step S2: When conducting live teaching, the system activates the live monitoring module, starts analyzing the comprehensive data of the student's environment, analyzes the iris fixation deviation of the student while watching, and evaluates the student's status based on the above analysis results;

[0018] Step S3: During the live teaching process, the system starts the live interaction module, begins to train the recognition model using historical live data, identifies bullet screen features and student question features based on the recognition model, matches the corresponding knowledge points based on the above features, further analyzes the records of students watching the live broadcast and screen recording and extracts the knowledge points, and further matches the corresponding teaching videos based on the above knowledge points.

[0019] Step S4: During the live teaching process, the system enters the questions raised by students into the system, integrates the knowledge points involved in the questions raised by students and the questions raised in the bullet comments, and sends them to the teacher.

[0020] According to the above technical solution, step S2 further includes the following steps:

[0021] Step S21: Construct a visual image database, acquire visual images of students watching the live broadcast during the live teaching period, identify the orientation features of the visual images, search the visual image database based on the orientation features, match the corresponding shooting position and shooting distance L0 of the student's end camera device, identify the student's facial features, anchor the position of the student's iris, establish a coordinate system, identify the coordinates of the student's end camera device and the coordinates of the student's left and right irises, calculate the distances L1 and L2 between the camera device and the student's iris using the distance formula, and establish the equation. In the formula, θ1 represents the angle between the line connecting the student's camera device and the student's right iris and the vertical plane containing the student's irises, and θ2 represents the angle between the line connecting the student's camera device and the student's left iris and the vertical plane containing the student's irises.

[0022] Step S22: Retrieve the corresponding influence coefficients on student status ratings from the database based on θ1 and θ2 respectively. Weight and fuse these influence coefficients to obtain the fused influence coefficient α. Retrieve the corresponding line-of-sight intersections from the database based on θ1 and θ2, identify the coordinates of the line-of-sight intersections, and retrieve the corresponding influence weights φ of the aforementioned influence coefficients from the database based on the coordinates of the line-of-sight intersections. i ;

[0023] Step S23: Obtain the shooting distance of the student-end camera device, retrieve the corresponding viewing area S1 from the database based on the shooting distance, obtain the screen area S2 of the student-end device, and calculate the ratio of the screen area of ​​the student-end device to the viewing area corresponding to the shooting distance. According to the ratio Retrieve the corresponding influence coefficient β on student status rating from the database.

[0024] According to the above technical solution, step S23 further includes the following steps:

[0025] Step S231: Obtain comprehensive environmental data, identify sound features in the environment, compare with audio features in the live video. If the similarity is greater than the threshold, the system continues to detect; otherwise, identify the sound volume in the environment and retrieve the corresponding influence coefficient ω on the student status score from the database based on the sound volume.

[0026] Step S232: Obtain the student's status score P0 under the set conditions from the database, and calculate the student's actual status score P1 using the formula (α·φ). i )·β·ω·P0, where P1 represents the student's actual status score. If the student's status score P1 is less than the system's set threshold, the system obtains the student's IP address and sends a pop-up reminder to the teacher.

[0027] According to the above technical solution, step S3 further includes the following steps:

[0028] Step S31: Obtain historical live streaming data, identify the characteristics of historical live streaming courses and bullet screen features, and use the historical live streaming course features and bullet screen features as sample data to train the recognition model;

[0029] Step S32: Obtain real-time bullet screen data, use the trained recognition model to identify bullet screen features in the real-time bullet screen data, identify the bullet screen features of each bullet screen and count them, identify the number of bullet screens with the same bullet screen features within the set recognition period T, if the number of bullet screens is greater than the system set threshold, the system will display a pop-up reminder for the bullet screen feature on the teacher's end, otherwise the system will continue to detect.

[0030] Step S33: When students are watching the screen recording, the timestamp of the progress bar is identified, the timestamp is marked and entered into the first table of the database, the difference R between adjacent first and second timestamps in the database is calculated, and when the difference R is greater than the system-set threshold, the second timestamp is entered into the second table. The timestamp of the first progress bar and the timestamp of the last progress bar in each table are identified, the video is segmented according to the above timestamps, the knowledge point features of each video segment are identified, the corresponding teaching video in the database is retrieved according to the knowledge point features, the knowledge point features are marked as the first questionable knowledge point features and saved to the database.

[0031] According to the above technical solution, step S4 further includes the following steps:

[0032] Step S41: At the end of the live teaching, acquire student question data, identify student question characteristics, and search the knowledge graph in the database to match the corresponding second question knowledge point characteristics based on the student question characteristics;

[0033] Step S42: Obtain student bullet screen features, first question knowledge point features, and second question knowledge point features, integrate the features and send them to the teacher's end.

[0034] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: By analyzing the angular deviation between the student's iris and the student's camera device while watching a live stream, this invention can accurately determine whether the student is focused on the live teaching, helping teachers monitor students' attention. Furthermore, based on the angular deviation between the iris and the camera device, it can accurately determine the student's line of sight, thus significantly improving the system's accuracy. By analyzing the proportion of the screen area in the viewing angle area while the student is watching the live stream, it can accurately determine whether the student is focused on the live video, further improving the system's accuracy. By judging whether the clarity of ambient sound content is the same as the clarity of sound in the live video, it can further analyze the impact of ambient sound volume on students' listening comprehension, accurately... By assessing students' engagement during class, the system significantly improves accuracy. The steps outlined above enable monitoring of online learning, preventing teachers from wasting class time monitoring individual student chat windows and thus greatly improving classroom efficiency. Real-time detection and counting of chat messages during live teaching allows teachers to address student questions while focusing on lecturing, enhancing teacher-student interaction. Identifying the characteristics of repeatedly viewed video clips and their corresponding knowledge points automatically and accurately matches users with relevant instructional videos, further improving system accuracy and interactivity. These steps allow teachers to accurately grasp student questions, strengthening teacher-student interaction and enabling them to develop more precise teaching plans tailored to students' specific needs. Attached Figure Description

[0035] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0036] Figure 1 This is a schematic diagram of the system module composition of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Please see Figure 1 The present invention provides a technical solution: a data management system based on a live interactive platform, comprising a data acquisition module, an interactive management module, and a feedback module;

[0039] The data acquisition module is used to collect real-time data on bullet comments, combined data on students watching live and recorded broadcasts on the student's end, and data on students' questions. It also collects visual images and environmental data of students watching live broadcasts during live teaching and inputs the combined data of historical live broadcasts into the system.

[0040] The interaction management module is used to train a recognition model using historical live streaming data. Based on the recognition model, it identifies and statistically analyzes the live stream comments, analyzes the visual images of students watching the live stream, analyzes the students' iris data and environmental data when students watch the live stream, evaluates the students' learning status based on the analysis results, analyzes the students' comments and the students' viewing records of the live stream recordings, obtains the knowledge points that students have questions about, matches materials for students based on the above knowledge points, and analyzes and integrates the characteristics of students' comments.

[0041] The data management module is used to integrate students' bullet screen data and live broadcast viewing data and transmit them to the teacher's end. Students can establish a connection with the teacher's end through the student end, provide feedback and interact with the teacher.

[0042] The data acquisition module includes a data collection module, a sensor module, and a historical live broadcast comprehensive data entry module. The data collection module is used to collect real-time data on bullet comments and comprehensive data on students watching live broadcasts and recorded broadcasts on their devices. The sensor module is used to collect visual images and comprehensive environmental information of students watching live broadcasts during live teaching. The historical live broadcast comprehensive data entry module is used to enter the comprehensive historical live broadcast data of various subjects into the system.

[0043] The interactive management module includes a live streaming monitoring module and a live streaming interaction module;

[0044] The live streaming monitoring module further includes an iris analysis submodule, a status assessment submodule, and an environmental analysis submodule. The iris analysis submodule is used to identify the iris position of students when watching the live stream and further analyze the impact of students' iris gaze deviation on students' status. The environmental analysis submodule is used to analyze the impact of comprehensive environmental data on students' status. The status assessment submodule is used to assess students' status based on the above analysis results.

[0045] The after-class interaction module further includes an operation analysis submodule, a knowledge matching submodule, and a model training submodule. The operation analysis submodule is used to analyze the time characteristics of students watching live broadcasts and screen recordings. The knowledge matching submodule is used to mark the live broadcast and screen recording videos according to the above time characteristics, further analyze the marked videos and extract knowledge point features, and further match the corresponding teaching videos according to the knowledge point features. The model training submodule is used to train a recognition model based on historical live broadcast comprehensive data, use the above recognition model to identify bullet screen features and student question features, and match and integrate knowledge points according to the above bullet screen features and student question features.

[0046] The data management module includes a summary and feedback module, which is used to feed back the knowledge points corresponding to the characteristics of bullet comments and the knowledge points corresponding to the videos marked by students to the teacher.

[0047] The data management module also includes a question feedback module, which is used to feed back students' question data to the system.

[0048] The operation of the data management system for a live streaming interactive platform mainly includes the following steps:

[0049] Step S1: Collect real-time data on bullet comments, combined data on students watching live and recorded broadcasts on their devices, and data on students' questions. Collect visual images and environmental data of students watching live broadcasts during live teaching, and input the combined data of historical live broadcasts into the system.

[0050] Step S2: When conducting live teaching, the system activates the live monitoring module, starts analyzing the comprehensive data of the student's environment, analyzes the iris fixation deviation of the student while watching, and evaluates the student's status based on the above analysis results;

[0051] Step S3: During the live teaching process, the system starts the live interaction module, begins to train the recognition model using historical live data, identifies bullet screen features and student question features based on the recognition model, matches the corresponding knowledge points based on the above features, further analyzes the records of students watching the live broadcast and screen recording and extracts the knowledge points, and further matches the corresponding teaching videos based on the above knowledge points.

[0052] Step S4: During the live teaching process, the system enters the questions raised by students into the system, integrates the knowledge points involved in the questions raised by students and the questions raised in the bullet comments, and sends them to the teacher.

[0053] Step S2 further includes the following steps:

[0054] Step S21: Construct a visual image database, acquire visual images of students watching the live broadcast during the live teaching period, identify the orientation features of the visual images, search the visual image database based on the orientation features, match the corresponding shooting position and shooting distance L0 of the student's end camera device, identify the student's facial features, anchor the position of the student's iris, establish a coordinate system, identify the coordinates of the student's end camera device and the coordinates of the student's left and right irises, calculate the distances L1 and L2 between the camera device and the student's iris using the distance formula, and establish the equation. In the formula, θ1 represents the angle between the line connecting the student's camera device and the student's right iris and the vertical plane containing the student's irises, and θ2 represents the angle between the line connecting the student's camera device and the student's left iris and the vertical plane containing the student's irises.

[0055] Step S22: Retrieve the corresponding influence coefficients on student status ratings from the database based on θ1 and θ2 respectively. Weight and fuse these influence coefficients to obtain the fused influence coefficient α. Retrieve the corresponding line-of-sight intersections from the database based on θ1 and θ2, identify the coordinates of the line-of-sight intersections, and retrieve the corresponding influence weights φ of the aforementioned influence coefficients from the database based on the coordinates of the line-of-sight intersections. i By analyzing the angular deviation between the pupil and iris of students and the camera device at the student's end while watching the live broadcast, it is possible to accurately determine whether students are focused on listening to the live broadcast teaching, which can help teachers monitor students' listening. Furthermore, based on the angular deviation between the iris and the camera device, it is possible to accurately determine the point of intersection of students' eyes, which can accurately determine whether students are focused on watching the live video, thereby greatly improving the accuracy of the system.

[0056] Step S23: Obtain the shooting distance of the student-end camera device, retrieve the corresponding viewing area S1 from the database based on the shooting distance, obtain the screen area S2 of the student-end device, and calculate the ratio of the screen area of ​​the student-end device to the viewing area corresponding to the shooting distance. According to the ratio By retrieving the corresponding influence coefficient β on student status rating from the database and analyzing the proportion of screen area in the viewing area when students watch live streams, it is possible to accurately determine whether students are focused on watching live stream videos, thereby further improving the accuracy of the system.

[0057] Step S23 further includes the following steps:

[0058] Step S231: Obtain comprehensive environmental data, identify sound features in the environment, and compare them with audio features in the live video. If the similarity is greater than the threshold, the system continues to detect; otherwise, it identifies the sound volume in the environment. Based on the sound volume, it retrieves the corresponding influence coefficient ω on student status rating from the database. By judging whether the clarity of the sound content in the environment is the same as the clarity of the sound in the live video, it further analyzes the impact of sound volume in the environment on students' listening, which can accurately judge students' listening status and thus greatly improve the accuracy of the system.

[0059] Step S232: Obtain the student's status score P0 under the set conditions from the database, and calculate the student's actual status score P1 using the formula (α·φ). i )·β·ω·P0, where P1 represents the student's actual status score. If the student's status score P1 is less than the system's set threshold, the system obtains the student's IP address and sends a pop-up reminder to the teacher. Through the above steps, the system can monitor students' online learning, avoid teachers wasting class time monitoring students' online learning, and thus greatly improve classroom efficiency.

[0060] Step S3 further includes the following steps:

[0061] Step S31: Obtain historical live streaming data, identify the characteristics of historical live streaming courses and bullet screen features, and use the historical live streaming course features and bullet screen features as sample data to train the recognition model;

[0062] Step S32: Obtain real-time bullet screen data, use the trained recognition model to identify bullet screen features in the real-time bullet screen data, identify the bullet screen features of each bullet screen and count them, identify the number of bullet screens with the same bullet screen features within the set recognition period T, if the number of bullet screens is greater than the system set threshold, the system will display a pop-up reminder for the bullet screen feature on the teacher's end, otherwise the system will continue to detect. By detecting and counting the bullet screen features in real time during the teacher's live teaching, the teacher can concentrate on teaching while also addressing students' questions, thereby greatly improving the interactivity between teachers and students in online teaching.

[0063] Step S33: When a student is watching a screen recording, the timestamp of the progress bar is identified, marked, and entered into the first table of the database. The difference R between adjacent first and second timestamps in the database is calculated. If the difference R is greater than a system-set threshold, the second timestamp is entered into the second table. The timestamps of the first and last progress bars in each table are identified. The video is segmented according to the timestamps, and the knowledge point features of each video segment are identified. The corresponding teaching video is retrieved from the database based on the knowledge point features. The knowledge point features are marked as the first questionable knowledge point features and saved to the database. By identifying the knowledge point features of video segments that students watch repeatedly, the system can automatically and accurately match the corresponding teaching video for the user, thereby improving the accuracy and interactivity of the system.

[0064] Step S4 further includes the following steps:

[0065] Step S41: At the end of the live teaching, acquire student question data, identify student question characteristics, and search the knowledge graph in the database to match the corresponding second question knowledge point characteristics based on the student question characteristics;

[0066] Step S42: Obtain student bullet screen features, first question knowledge point features, and second question knowledge point features, integrate the features and send them to the teacher's end. Through the above steps, teachers can accurately grasp students' questions, thereby greatly enhancing the interaction between teachers and students in online teaching, and at the same time, making it easier for teachers to formulate more accurate teaching plans for students' question knowledge points.

[0067] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0068] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data management system based on a live streaming interactive platform, comprising a data acquisition module, an interaction management module, and a data management module, characterized in that: The data acquisition module is used to collect barrage data, comprehensive data of students watching live and recorded broadcasts on the student's end, and data of students' questions in real time. It also collects comprehensive data of visual images and environment of students watching live broadcasts during live teaching and inputs comprehensive data of historical live broadcasts into the system. The interactive management module is used to train a recognition model using historical live streaming data, to identify and statistically analyze live streaming comments based on the recognition model, to analyze the visual images of students watching the live stream, to analyze the students' iris data and environmental data when students watch the live stream, to evaluate the students' learning status based on the analysis results, to analyze the students' comments and the students' viewing records of the live stream recordings, to obtain the knowledge points that the students are confused about, to match materials for the students based on the above knowledge points, and to analyze and integrate the characteristics of the students' comments. The data management module is used to integrate students' bullet screen data and live broadcast viewing data and transmit them to the teacher's end. Students can establish a connection with the teacher's end through the student end, provide feedback and interact with the teacher. The data acquisition module includes a data collection module, a sensor module, and a historical live broadcast comprehensive data entry module. The data collection module is used to collect real-time barrage data and comprehensive data of students watching live broadcasts and recorded broadcasts on the student's end. The sensor module is used to collect visual images and comprehensive environmental information of students watching live broadcasts during live teaching. The historical live broadcast comprehensive data entry module is used to enter the historical live broadcast comprehensive data of various subjects into the system. The interactive management module includes a live streaming supervision module and a live streaming interaction module; The live streaming monitoring module further includes an iris analysis submodule, a status assessment submodule, and an environmental analysis submodule. The iris analysis submodule is used to identify the iris position of students when watching the live stream and further analyze the impact of students' iris fixation deviation on students' status. The environmental analysis submodule is used to analyze the impact of comprehensive environmental data on students' status. The status assessment submodule is used to assess students' status based on the above analysis results. The live streaming interaction module further includes an operation analysis submodule, a knowledge matching submodule, and a model training submodule. The operation analysis submodule is used to analyze the time characteristics of students watching live streams and screen recordings. The knowledge matching submodule is used to mark the live stream and screen recording videos according to the time characteristics, further analyze the marked videos and extract knowledge point features, and further match the corresponding teaching videos according to the knowledge point features. The model training submodule is used to train a recognition model based on historical live stream comprehensive data, use the recognition model to identify bullet screen features and student question features, and match and integrate knowledge points according to the bullet screen features and student question features. The data management module includes a summary and feedback module, which is used to feed back the knowledge points corresponding to the bullet screen features and the knowledge points corresponding to the marked videos watched by students to the teacher's end. The data management module also includes a question feedback module, which is used to feed back students' question data to the system. The operation method of the data management system of the live streaming interactive platform mainly includes the following steps: Step S1: Collect real-time data on bullet comments, combined data on students watching live and recorded broadcasts on their devices, and data on students' questions. Collect visual images and environmental data of students watching live broadcasts during live teaching, and input the combined data of historical live broadcasts into the system. Step S2: When conducting live teaching, the system activates the live monitoring module, starts analyzing the comprehensive data of the student's environment, analyzes the iris fixation deviation of the student while watching, and evaluates the student's status based on the above analysis results; Step S3: During the live teaching process, the system starts the live interaction module, begins to train the recognition model using historical live data, identifies bullet screen features and student question features based on the recognition model, matches the corresponding knowledge points based on the above features, further analyzes the records of students watching the live broadcast and screen recording and extracts the knowledge points, and further matches the corresponding teaching videos based on the above knowledge points. Step S4: During the live teaching process, the system enters the questions raised by students into the system, integrates the knowledge points involved in the questions raised by students and the questions raised in the bullet comments, and sends them to the teacher's end; Step S2 further includes the following steps: Step S21: Construct a visual image database, acquire visual images of students watching the live broadcast during the live teaching period, identify the orientation features of the visual images, search the visual image database based on the orientation features, match the corresponding shooting position and shooting distance L0 of the student's end camera device, identify the student's facial features, anchor the position of the student's iris, establish a coordinate system, identify the coordinates of the student's end camera device and the coordinates of the student's left and right irises, calculate the distances L1 and L2 between the camera device and the student's iris using the distance formula, and establish the equation. In the formula, θ1 represents the angle between the line connecting the student's camera device and the student's right iris and the vertical plane containing the student's irises, and θ2 represents the angle between the line connecting the student's camera device and the student's left iris and the vertical plane containing the student's irises. Step S22: Retrieve the corresponding influence coefficients on student status ratings from the database based on θ1 and θ2 respectively. Weight and fuse these influence coefficients to obtain the fused influence coefficient α. Retrieve the corresponding line-of-sight intersections from the database based on θ1 and θ2, identify the coordinates of the line-of-sight intersections, and retrieve the corresponding influence weights φ of the aforementioned influence coefficients from the database based on the coordinates of the line-of-sight intersections. i ; Step S23: Obtain the shooting distance of the student-end camera device, retrieve the corresponding viewing area S1 from the database based on the shooting distance, obtain the screen area S2 of the student-end device, and calculate the ratio of the screen area of ​​the student-end device to the viewing area corresponding to the shooting distance. According to the ratio Retrieve the corresponding influence coefficient β on student status rating from the database; Step S23 further includes the following steps: Step S231: Obtain comprehensive environmental data, identify sound features in the environment, compare with audio features in the live video. If the similarity is greater than the threshold, the system continues to detect; otherwise, identify the sound volume in the environment and retrieve the corresponding influence coefficient ω on the student status score from the database based on the sound volume. Step S232: Obtain the student's status score P0 under the set conditions from the database, and calculate the student's actual status score P1 using the formula (α·φ). i )·β·ω·P0, where P1 represents the student's actual status score. If the student's status score P1 is less than the system's set threshold, the system obtains the student's IP address and sends a pop-up reminder to the teacher. Step S3 further includes the following steps: Step S31: Obtain historical live streaming data, identify the characteristics of historical live streaming courses and bullet screen features, and use the historical live streaming course features and bullet screen features as sample data to train the recognition model; Step S32: Obtain real-time bullet screen data, use the trained recognition model to identify bullet screen features in the real-time bullet screen data, identify the bullet screen features of each bullet screen and count them, identify the number of bullet screens with the same bullet screen features within the set recognition period T, if the number of bullet screens is greater than the system set threshold, the system will display a pop-up reminder for the bullet screen feature on the teacher's end, otherwise the system will continue to detect. Step S33: When students are watching the screen recording, the timestamp of the progress bar is identified, the timestamp is marked and entered into the first table of the database, the difference R between adjacent first and second timestamps in the database is calculated, and when the difference R is greater than the system set threshold, the second timestamp is entered into the second table. The timestamp of the first progress bar and the timestamp of the last progress bar in each table are identified, the video is segmented according to the above timestamps, the knowledge point features of each video segment are identified, the corresponding teaching video in the database is retrieved according to the knowledge point features, the knowledge point features are marked as the first questionable knowledge point features and saved to the database. Step S4 further includes the following steps: Step S41: At the end of the live teaching, acquire student question data, identify student question characteristics, and search the knowledge graph in the database to match the corresponding second question knowledge point characteristics based on the student question characteristics; Step S42: Obtain student bullet screen features, first question knowledge point features, and second question knowledge point features, integrate the features and send them to the teacher's end.

Images