A Method for Emotional Attention in AI Online Education

Abstract

The present invention discloses an emotion attention method for AI online education, which is applied among the student side, the teacher side and the server. The server sends the live video stream of the teacher side to each student side for playing. This emotion attention method for AI online education calculates the average comprehension degree value according to the learning expressions of students to evaluate the understanding and attention degree of students to the current course lecture, and dynamically and intelligently adjusts the playing speed of each student side to provide a personalized learning experience for each student, thereby improving learning efficiency; and sends a reminder message about the overall understanding situation of students to the teacher side, enabling the teacher to better interact with students and timely adjust teaching methods and content; the student side with the camera turned on sends the speed adjustment status information to all other student sides and the server, without the need for the server to interact with each student side, which can reduce the server processing pressure, and the server adjusts the bit rate of the live video stream by calculating each playing ratio.

Description

Technical Field

[0001] The present invention belongs to the technical field of online education, and specifically relates to an emotion attention method for AI online education. Background Art

[0002] Online learning refers to learning conducted in an electronic environment facilitated by communications, microcomputer, computer, artificial intelligence, network, and multimedia technologies. It is technology-based learning. It involves teaching and learning online through the internet, or via mobile wireless networks, in a virtual classroom. Interactive learning is an emerging learning approach that leverages multimedia, computer, and network technologies to foster interactive communication between teachers and learners, as well as between groups of learners. This learning approach facilitates one-on-one learning and teaching, fully respects individual learners, and stimulates learning motivation. It is not restricted by time, location, or space, and allows for the same level of interaction as in real life.

[0003] In existing online learning technologies, students passively watch course videos by default and are generally unable to adjust the playback speed of live courses. Even if a small number of online education systems support real-time playback functions, students are required to manually replay or adjust the playback speed. It is impossible to achieve adaptive course playback speed adjustment based on each student's situation. Even if manual adjustment is used, the operation is not convenient. Moreover, if students do not actively provide feedback to the teacher on their level of understanding, the teacher cannot obtain the students' understanding information, resulting in low educational efficiency. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems raised in the background technology and propose an emotion attention method for AI online education.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] The present invention proposes an emotion-focused method for AI online education, which is applied between a student terminal, a teacher terminal, and a server. The server sends the live video stream from the teacher terminal to each student terminal for playback, wherein:

[0007] For students with cameras turned on, a preset number of facial images are collected, and the collected facial images are used for expression recognition through the trained deep learning model to obtain the learning expressions corresponding to each facial image;

[0008] Based on the pre-set mapping relationship between learning expressions and understanding levels, each student terminal with a camera turned on calculates the average understanding level of the corresponding student;

[0009] Compare each average understanding degree value with the fast - forward playback threshold and / or the slow - down playback threshold, and adjust the playback speed of each student terminal according to the comparison result;

[0010] Each student terminal with the camera turned on uses a broadcast mechanism to send speed - adjustment status information to all other student terminals and the server;

[0011] Each student terminal determines whether the first preset condition is met based on the current playback time point of the video stream, the speed - adjustment status information, and the total number of student terminals, and the server determines whether the second preset condition is met based on the current playback time point of the live video stream, the speed - adjustment status information, and the total number of student terminals;

[0012] For each student terminal, when the first preset condition is met, the current student terminal calculates the first normal - playback proportion, the first slow - down playback proportion, and the first fast - forward playback proportion respectively according to the speed - adjustment status information, and the current student terminal adjusts its own playback speed according to the magnitudes of these three proportions;

[0013] When the second preset condition is met, the server calculates the second normal - playback proportion, the second slow - down playback proportion, and the second fast - forward playback proportion respectively according to the speed - adjustment status information, and the server sends a reminder message about the overall understanding situation of the students to the teacher terminal according to the magnitudes of these three proportions.

[0014] Preferably, data caching is performed while the live video stream is being played on the student terminal, and the data - caching speed is greater than the playback speed.

[0015] Preferably, the categories of learning expressions include understanding and concentrating, understanding but distracted, confused but not giving up, confused and giving up, and absent - minded.

[0016] Preferably, each student terminal with the camera turned on calculates the average understanding degree value of the corresponding student The formula is as follows:

[0017] ;

[0018] Wherein, represents the number of face images collected by each student terminal, represents the understanding degree value corresponding to the learning expression of the th face image collected, represents the weight corresponding to the understanding degree value of the th face image collected.

[0019] Preferably, the step of comparing each average understanding degree value with the fast - forward playback threshold and / or the slow - down playback threshold, and adjusting the playback speed of each student terminal according to the comparison result includes:

[0020] Each student terminal with the camera turned on compares the current playback time point of its own video stream with the current playback time point of the live video stream. When the difference between the two is within a preset range, the student terminal with the camera currently turned on is watching the live broadcast; otherwise, the student terminal with the camera currently turned on is watching a non-live broadcast;

[0021] For the student terminals that are watching the live broadcast and have the camera turned on, compare the average comprehension level value of each corresponding student with the deceleration playback threshold:

[0022] If the average comprehension level value is less than or equal to the deceleration playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to decelerate playback at the first preset speed;

[0023] If the average comprehension level value is greater than the deceleration playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to play normally at the second preset speed;

[0024] For the student terminals that are watching a non-live broadcast and have the camera turned on, compare the average comprehension level value of each corresponding student with the deceleration playback threshold and the acceleration playback threshold respectively:

[0025] If the average comprehension level value is less than or equal to the deceleration playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to decelerate playback at the first preset speed;

[0026] If the average comprehension level value is greater than or equal to the acceleration playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to accelerate playback at the third preset speed;

[0027] If the average comprehension level value is greater than the deceleration playback threshold and less than the acceleration playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to play normally at the second preset speed.

[0028] Preferably, the speed adjustment state information includes the student ID, the current playback time point T1 of the video stream when each student terminal with the camera turned on sends the speed adjustment state information, and the adjusted playback speed state, where the adjusted playback speed state is normal playback, deceleration playback, or acceleration playback.

[0029] Preferably, each student terminal determines whether the first preset condition is met based on the current playback time point of the video stream, the speed adjustment state information, and the total number of student terminals, and the server determines whether the second preset condition is met based on the current playback time point of the live video stream, the speed adjustment state information, and the total number of student terminals, including:

[0030] After each student terminal receives the speed adjustment state information, it performs a first difference calculation on the current playback time point T2 of the student terminal itself and the T1 in each speed adjustment state information, and counts the number A of the first differences less than the first threshold;

[0031] Determine whether the ratio of quantity A to the total number of student terminals is greater than the second threshold. If it is greater than the second threshold, the first preset condition is satisfied;

[0032] After the server receives each speed adjustment state information, it calculates the second difference between the current playback time point T3 of the live video stream and T1 in each speed adjustment state information respectively, and counts the number B of the second differences less than the first threshold;

[0033] Determine whether the ratio of quantity B to the total number of student terminals is greater than the second threshold. If it is greater than the second threshold, the second preset condition is satisfied.

[0034] Preferably, for each student terminal, when the first preset condition is satisfied, the current student terminal calculates the first normal playback ratio, the first deceleration playback ratio, and the first acceleration playback ratio respectively according to each speed adjustment state information, and the current student terminal adjusts its own playback speed according to the magnitudes of these three ratios, including:

[0035] The first normal playback ratio is the ratio of the number of normal playback in all received speed adjustment state information to quantity A, the first deceleration playback ratio is the ratio of the number of deceleration playback in all received speed adjustment state information to quantity A, and the first acceleration playback ratio is the ratio of the number of acceleration playback in all received speed adjustment state information to quantity A;

[0036] For each student terminal, among the first normal playback ratio, the first deceleration playback ratio, and the first acceleration playback ratio, if one of the ratios reaches the third threshold, the current student terminal plays at the playback speed in the speed adjustment state information corresponding to this ratio. If none of the three ratios reaches the third threshold, the current student terminal continues to play at its current speed.

[0037] Preferably, when the second preset condition is satisfied, the server calculates the second normal playback ratio, the second deceleration playback ratio, and the second acceleration playback ratio respectively according to each speed adjustment state information, and the server sends a reminder message about the overall understanding situation of the students to the teacher terminal according to the magnitudes of these three ratios, including:

[0038] The second normal playback ratio is the ratio of the number of normal playback in all received speed adjustment state information to quantity B, the second deceleration playback ratio is the ratio of the number of deceleration playback in all received speed adjustment state information to quantity B, and the second acceleration playback ratio is the ratio of the number of acceleration playback in all received speed adjustment state information to quantity B;

[0039] Among the second normal playback ratio, the second slow - down playback ratio, and the second speed - up playback ratio, if the second slow - down playback ratio reaches the fourth threshold, the server sends a reminder message indicating that the students generally have difficulty in understanding to the teacher terminal. If the second normal playback ratio or the second speed - up playback ratio reaches the fourth threshold, the server sends a reminder message indicating that the students generally have normal understanding to the teacher terminal.

[0040] Preferably, when the server sends a reminder message indicating that the students generally have difficulty in understanding to the teacher terminal, the server also reduces the bitrate of the live video stream of each student terminal.

[0041] When the server sends a reminder message indicating that the students generally have normal understanding to the teacher terminal, the server also restores the bitrate of the live video stream of each student terminal to the original bitrate.

[0042] Among them, the original bitrate is the bitrate at which the server sends the live video stream of the teacher terminal to each student terminal for playback.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0044] The emotion - focused method for this AI online education calculates the average understanding degree value based on the learning expressions of students to evaluate the understanding and attention of students to the current course lecture, and dynamically and intelligently adjusts the playback speed of each student terminal to provide a personalized learning experience for each student, thereby improving learning efficiency; and sends a reminder message about the overall understanding situation of students to the teacher terminal, enabling the teacher to better interact with students and timely adjust teaching methods and content.

[0045] In this method, the student terminals with the camera enabled send speed - adjustment status information to all other student terminals and the server. There is no need for the server to interact with each student terminal, which can reduce the server processing pressure. Moreover, the server adjusts the bitrate of the live video stream by calculating each playback ratio to prevent the video cache of the student terminal from overloading, so as to optimize the user experience and performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of the emotion - focused method for the AI online education of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] In order to make the objectives, technical solutions, and advantages of the present application clearer, the following further describes the present application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] In one embodiment, as Figure 1As shown, an emotion attention method for AI online education is provided. This method is applied among the student side, the teacher side, and the server. The server sends the live video stream from the teacher side to each student side for playback (the teacher starts live teaching on an intelligent terminal (such as a computer, PAD, mobile phone, etc.) through educational software (such as an APP, a small program, etc.). The educational software is called the teacher side. The live teaching content of the teacher side is sent in the form of a video stream and in a multicast manner through the server. Multiple students join the multicast group through the educational software client on the intelligent terminal (called the student side) to watch the live teaching content. While the live video stream is being played on the student side, data caching is performed, and the data caching speed is greater than the playback speed, where:

[0049] Step 1: For the student side with the camera turned on, collect a preset number of face images, and perform facial expression recognition on the collected face images through a trained deep learning model to obtain the learning expressions corresponding to their respective face images.

[0050] It should be noted that for the student side with the camera turned on, the camera samples face images periodically (such as once every five minutes), and to reduce random errors, multiple face images are collected each time, such as three. If the deep learning model is a convolutional neural network model, the process of how to train the deep learning model and how to perform facial expression recognition belongs to the prior art, and this solution will not be elaborated in detail.

[0051] Among them, in this embodiment, the categories of learning expressions include understanding and concentrating, understanding but being distracted, being confused but not giving up, being confused and giving up, and being out of state.

[0052] Step 2: According to the pre-set mapping relationship between the learning expression and the understanding degree value, each student side with the camera turned on calculates the average understanding degree value of the corresponding student.

[0053] In this embodiment, the pre-set mapping relationship between the learning expression and the understanding degree value is: the understanding degree value corresponding to understanding and concentrating is 1, the understanding degree value corresponding to understanding but being distracted is 0.7, the understanding degree value corresponding to being confused but not giving up is 0.3, the understanding degree value corresponding to being confused and giving up is 0, and the understanding degree value corresponding to being out of state is 0.5;

[0054] Among them, each student side with the camera turned on calculates the average understanding degree value of the corresponding student

[0055] ;

[0056] Among them, represents the number of face images collected by each student side (in this embodiment is 3), Indicates the degree of understanding corresponding to the learned expression of the th face image collected, Indicates the weight corresponding to the degree of understanding of the th face image collected.

[0057] In this embodiment, the weight of being out of state is 1, and the weights of understanding and concentrating, understanding but being distracted, being confused but not giving up, and being confused and giving up are all 3, so as to reduce the influence degree of being out of state (possibly understanding, possibly not understanding) on the average degree of understanding value. Through the average degree of understanding value to evaluate the understanding and attention degree of students to the current course lecture.

[0058] Step 3: Compare each average degree of understanding value with the fast-forward playback threshold and / or the slow-down playback threshold, and adjust the playback speed of each student terminal according to the comparison result, including:

[0059] Each student terminal with the camera turned on compares the current playback time point of its own video stream with the current playback time point of the live video stream. When the difference between the two is within a preset range (such as within 1 minute), the current student terminal with the camera turned on is watching the live broadcast, otherwise the current student terminal with the camera turned on is watching a non-live broadcast;

[0060] For the student terminal that is watching the live broadcast and has the camera turned on, compare the average degree of understanding value of each corresponding student with the slow-down playback threshold (in this embodiment, the slow-down playback threshold is 0.3):

[0061] If the average degree of understanding value is less than or equal to the slow-down playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to slow down playback at the first preset speed (such as slow down playback at 0.8 times speed);

[0062] If the average degree of understanding value is greater than the slow-down playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to play normally at the second preset speed (that is, play normally at 1.0 times speed);

[0063] For the student terminal that is watching a non-live broadcast and has the camera turned on, compare the average degree of understanding value of each corresponding student with the slow-down playback threshold and the fast-forward playback threshold respectively (in this embodiment, the fast-forward playback threshold is 0.7):

[0064] If the average degree of understanding value is less than or equal to the slow-down playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to slow down playback at the first preset speed (such as slow down playback at 0.8 times speed);

[0065] If the average degree of understanding value is greater than or equal to the fast-forward playback threshold, the corresponding student terminal with the camera turned on adjusts the playback speed state to accelerate playback at the third preset speed (such as accelerate playback at 1.2 times speed);

[0066] If the average comprehension degree value is greater than the deceleration playback threshold and less than the acceleration playback threshold, the corresponding student side of the camera is turned on to adjust the playback speed state to play normally at the second preset speed (that is, play normally at 1.0 times speed).

[0067] Step 4: Each student side with the camera turned on uses the broadcast mechanism to send the speed adjustment state information to all other student sides (including the server) (where the broadcast mechanism means that the student side with the camera turned on sends the speed adjustment state information to all other students (including the student side with the camera turned on and the student side without the camera turned on) and the server);

[0068] Among them, the speed adjustment state information includes the student ID, the current playback video stream time point T1 when each student side with the camera turned on sends the speed adjustment state information (that is, T1 is the current playback video timestamp when the student side with the camera turned on sends the speed adjustment state information), and the adjusted playback speed state, where the adjusted playback speed state is normal playback, deceleration playback or acceleration playback.

[0069] Step 5: Each student side judges whether the first preset condition is met according to the current playback video stream time point, the speed adjustment state information and the total number of student sides, and the server judges whether the second preset condition is met according to the current playback time point of the live video stream, the speed adjustment state information and the total number of student sides, including:

[0070] For judging whether the first preset condition is met: After each student side receives the speed adjustment state information, it performs the first difference calculation on the current playback video stream time point T2 of the student side itself (that is, T2 is the current playback video stream time point of the student side that receives the speed adjustment state information) and T1 in each speed adjustment state information respectively, and counts the number A of the first differences less than the first threshold (such as the first threshold is 30 seconds);

[0071] Judge whether the ratio of the number A to the total number of student sides is greater than the second threshold (such as the second threshold is 30%, and by comparing with the second threshold, if it is greater than the second threshold, it means that there is a certain reference value). If it is greater than the second threshold, the first preset condition is met (otherwise it is not met);

[0072] For judging whether the second preset condition is met: After the server receives each speed adjustment state information, it performs the second difference calculation on the current playback time point T3 of the live video stream and T1 in each speed adjustment state information respectively, and counts the number B of the second differences less than the first threshold;

[0073] Judge whether the ratio of the number B to the total number of student sides is greater than the second threshold. If it is greater than the second threshold, the second preset condition is met (otherwise it is not met).

[0074] Step 6. For each student terminal, when the first preset condition is met (if the first preset condition is not met, the current student terminal does not need to adjust its own playback speed and continues to play at its current speed), the current student terminal calculates the first normal playback ratio, the first decelerated playback ratio, and the first accelerated playback ratio based on each speed adjustment status information respectively. And the current student terminal adjusts its own playback speed according to the magnitudes of these three ratios, including:

[0075] The first normal playback ratio is the ratio of the number of normal playback in all received speed adjustment status information to quantity A. The first decelerated playback ratio is the ratio of the number of decelerated playback in all received speed adjustment status information to quantity A. The first accelerated playback ratio is the ratio of the number of accelerated playback in all received speed adjustment status information to quantity A.

[0076] For each student terminal, among the first normal playback ratio, the first decelerated playback ratio, and the first accelerated playback ratio, if one of the ratios reaches the third threshold (i.e., is greater than or equal to the third threshold, such as the third threshold is 60%), the current student terminal plays at the playback speed in the speed adjustment status information corresponding to this ratio (i.e., the playback speed corresponding to the first normal playback ratio is normal playback, the playback speed corresponding to the first decelerated playback ratio is decelerated playback, and the playback speed corresponding to the first accelerated playback ratio is accelerated playback). If none of the three ratios reaches the third threshold, the current student terminal continues to play at its current speed. That is, Steps 4 - 6 achieve global speed adjustment judgment through the broadcast mechanism of the camera - enabled student terminals.

[0077] Step 7. When the second preset condition is met (if the second preset condition is not met, the server does not need to send a reminder message to the teacher terminal, nor does it need to adjust the bitrate of the live video stream of each student terminal, and continues with the current bitrate), the server calculates the second normal playback ratio, the second decelerated playback ratio, and the second accelerated playback ratio based on each speed adjustment status information respectively. And the server sends a reminder message about the overall understanding situation of the students to the teacher terminal according to the magnitudes of these three ratios, including:

[0078] The second normal playback ratio is the ratio of the number of normal playback in all received speed adjustment status information to quantity B. The second decelerated playback ratio is the ratio of the number of decelerated playback in all received speed adjustment status information to quantity B. The second accelerated playback ratio is the ratio of the number of accelerated playback in all received speed adjustment status information to quantity B.

[0079] Among the second normal playback ratio, the second slow - down playback ratio, and the second speed - up playback ratio, if the second slow - down playback ratio reaches the fourth threshold (i.e., greater than or equal to the fourth threshold, for example, the fourth threshold is 60%), the server sends a reminder message of overall understanding difficulty of students to the teacher terminal. If the second normal playback ratio or the second speed - up playback ratio reaches the fourth threshold, the server sends a reminder message of overall normal understanding of students to the teacher terminal; the teacher terminal reminds the teacher of the reminder message, and the teacher adjusts the teaching content, speed, or method according to the reminder message.

[0080] Among them, when the server sends a reminder message of overall understanding difficulty of students to the teacher terminal, the server also reduces the bitrate of the live video stream of each student terminal (specifically, how much to reduce can be set according to actual needs. By adjusting the bitrate, it is possible to prevent the video cache of the student terminal from being overloaded);

[0081] When the server sends a reminder message of overall normal understanding of students to the teacher terminal, the server also restores the bitrate of the live video stream of each student terminal to the original bitrate;

[0082] Among them, the original bitrate is the bitrate at which the server sends the live video stream of the teacher terminal to each student terminal for playback.

[0083] Then return to step 1 to continue continuous detection.

[0084] In another embodiment, the present application further includes an emotion - attention device for AI online education, including a processor and a memory storing a number of computer instructions. When the computer instructions are executed by the processor, the steps of the method described in steps 1 - 7 are implemented.

[0085] This emotion - attention method for AI online education calculates the average understanding degree value based on the learning expressions of students to evaluate the understanding and attention of students to the current course lecture, and dynamically and intelligently adjusts the playback speed of each student terminal to provide a personalized learning experience for each student, thereby improving learning efficiency; and sends a reminder message of the overall understanding situation of students to the teacher terminal, enabling the teacher to better interact with students and timely adjust teaching methods and content; in this method, the student terminals with the camera turned on send speed - adjustment status information to all other student terminals and the server, without the need for message interaction between the server and each student terminal, which can reduce the server processing pressure, and the server adjusts the bitrate of the live video stream by calculating each playback ratio to prevent the video cache of the student terminal from being overloaded, so as to optimize the user experience and performance.

[0086] It should be understood that although Figure 1The steps in the flowchart are shown in sequence according to the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least a part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0087] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for emotional attention in AI online education, characterized in that: The method is applied among the student side, the teacher side and the server. The server sends the live video stream of the teacher side to each student side for playing, where: For the student side with the camera turned on, a preset number of face images are collected, and the collected face images are subjected to expression recognition through a trained deep learning model to obtain the learning expressions corresponding to their respective face images; According to the mapping relationship between the preset learning expressions and the understanding degree values, each student side with the camera turned on calculates the average understanding degree value of the corresponding student; Compare each average understanding degree value with the acceleration playback threshold and / or the deceleration playback threshold, and adjust the playback speed of each student side according to the comparison result; Each student side with the camera turned on uses the broadcast mechanism to send the speed adjustment status information to all other student sides and the server; Each student side judges whether the first preset condition is met according to the current playback time point of the video stream, the speed adjustment status information and the total number of student sides, and the server judges whether the second preset condition is met according to the current playback time point of the live video stream, the speed adjustment status information and the total number of student sides; For each student side, when the first preset condition is met, the current student side calculates the first normal playback ratio, the first deceleration playback ratio and the first acceleration playback ratio respectively according to the speed adjustment status information, and the current student side adjusts its own playback speed according to the magnitudes of the three ratios; When the second preset condition is met, the server calculates the second normal playback ratio, the second deceleration playback ratio and the second acceleration playback ratio respectively according to the speed adjustment status information, and the server sends a reminder message about the overall understanding situation of the students to the teacher side according to the magnitudes of the three ratios.

2. The emotional attention method for AI online education according to claim 1, characterized in that: Data caching is performed while the live video stream is being played on the student side, and the data caching speed is greater than the playback speed.

3. The emotion attention method for AI online education as described in claim 1, wherein: The categories of the learning expressions include understanding and concentrating, understanding but distracted, confused but not giving up, confused and giving up, and absent-minded.

4. The emotional attention method for AI online education according to claim 1, wherein: Each student terminal that turns on the camera calculates the average understanding degree value of the corresponding student The formula is as follows: ； Among them, represents the number of face images collected by each student terminal, represents the comprehension degree value corresponding to the learning expression of the th face image collected, represents the weight corresponding to the comprehension degree value of the th face image collected.

5. The emotional attention method for AI online education according to claim 1, characterized in that: The comparison of each average understanding degree value with the acceleration playback threshold and / or the deceleration playback threshold, and the adjustment of the playback speed of each student side according to the comparison result include: Each student side with the camera turned on compares its own current playback time point of the video stream with the current playback time point of the live video stream. When the difference between the two is within the preset range, the current student side with the camera turned on is watching the live broadcast, otherwise the current student side with the camera turned on is watching the non-live broadcast; For the student side that is watching the live broadcast and has the camera turned on, compare the average understanding degree value of each corresponding student with the deceleration playback threshold: If the average understanding degree value is less than or equal to the deceleration playback threshold, the corresponding student side with the camera turned on adjusts the playback speed state to decelerate playback at the first preset speed; If the average understanding degree value is greater than the deceleration playback threshold, the corresponding student side with the camera turned on adjusts the playback speed state to normal playback at the second preset speed; For the student side that is watching the non-live broadcast and has the camera turned on, compare the average understanding degree value of each corresponding student with the deceleration playback threshold and the acceleration playback threshold respectively: If the average comprehension degree value is less than or equal to the slow-down playback threshold, the playback speed state of the student side with the camera enabled is adjusted to slow down playback at the first preset speed; If the average comprehension degree value is greater than or equal to the speed-up playback threshold, the playback speed state of the student side with the camera enabled is adjusted to speed up playback at the third preset speed; If the average comprehension degree value is greater than the slow-down playback threshold and less than the speed-up playback threshold, the playback speed state of the student side with the camera enabled is adjusted to play normally at the second preset speed.

6. The emotional attention method for AI online education according to claim 5, characterized in that: The speed adjustment state information includes the student ID, the current playback video stream time point T1 when each student side with the camera enabled sends the speed adjustment state information, and the playback speed adjustment state, where the playback speed adjustment state is normal playback, slow-down playback, or speed-up playback.

7. The emotional attention method for AI online education according to claim 6, characterized in that: Each student side determines whether the first preset condition is met based on the current playback video stream time point, the speed adjustment state information, and the total number of student sides, and the server determines whether the second preset condition is met based on the current playback time point of the live video stream, the speed adjustment state information, and the total number of student sides, including: After each student side receives the speed adjustment state information, it calculates the first difference between the current playback video stream time point T2 of the student side itself and T1 in each speed adjustment state information, and counts the number A of the first differences less than the first threshold; Determine whether the ratio of the number A to the total number of student sides is greater than the second threshold. If it is greater than the second threshold, the first preset condition is met; After the server receives each speed adjustment state information, it calculates the second difference between the current playback time point T3 of the live video stream and T1 in each speed adjustment state information, and counts the number B of the second differences less than the first threshold; Determine whether the ratio of the number B to the total number of student sides is greater than the second threshold. If it is greater than the second threshold, the second preset condition is met.

8. The emotional attention method for AI online education according to claim 7, characterized in that: For each student side, when the first preset condition is met, the current student side calculates the first normal playback ratio, the first slow-down playback ratio, and the first speed-up playback ratio based on each speed adjustment state information, and the current student side adjusts its own playback speed according to the magnitudes of these three ratios, including: The first normal playback ratio is the ratio of the number of normal playback in all received speed adjustment state information to the number A, the first slow-down playback ratio is the ratio of the number of slow-down playback in all received speed adjustment state information to the number A, and the first speed-up playback ratio is the ratio of the number of speed-up playback in all received speed adjustment state information to the number A; For each student side, among the first normal playback ratio, the first slow-down playback ratio, and the first speed-up playback ratio, if one of the ratios reaches the third threshold, the current student side plays at the playback speed in the speed adjustment state information corresponding to this ratio. If none of the three ratios reaches the third threshold, the current student side continues to play at its own current speed.

9. The emotional attention method for AI online education according to claim 7, wherein: When the second preset condition is satisfied, the server calculates the second normal playback ratio, the second slow-down playback ratio, and the second speed-up playback ratio based on each speed adjustment state information respectively, and the server sends a reminder message about the overall understanding situation of the students to the teacher side according to the magnitudes of these three ratios, including: The second normal playback ratio is the ratio of the number of normal playback in all received speed adjustment state information to the quantity B, the second slow-down playback ratio is the ratio of the number of slow-down playback in all received speed adjustment state information to the quantity B, and the second speed-up playback ratio is the ratio of the number of speed-up playback in all received speed adjustment state information to the quantity B; Among the second normal playback ratio, the second slow-down playback ratio, and the second speed-up playback ratio, if the second slow-down playback ratio reaches the fourth threshold, the server sends a reminder message that the students have overall difficulty in understanding to the teacher side, and if the second normal playback ratio or the second speed-up playback ratio reaches the fourth threshold, the server sends a reminder message that the students have normal overall understanding to the teacher side.

10. The emotion attention method for AI online education according to claim 9, wherein: When the server sends a reminder message that the students have overall difficulty in understanding to the teacher side, the server also reduces the bitrate of the live video stream of each student side; When the server sends a reminder message that the students have normal overall understanding to the teacher side, the server also restores the bitrate of the live video stream of each student side to the original bitrate; Wherein, the original bitrate is the bitrate at which the server sends the live video stream of the teacher side to each student side for playback.

Images