An interactive intelligent teaching device and method

Abstract

The application relates to the technical field of teaching devices, in particular to an interactive intelligent teaching device and method. The device comprises a processing server internally provided with an AA-ViT expression recognition model of dynamic occlusion perception, a dual-mode teaching blackboard, a teacher operation interactive terminal, a student interactive terminal arranged on a desktop and an image acquisition unit, and a method based on the device is disclosed. The above-mentioned interactive intelligent teaching device and method are used, the AA-ViT expression recognition model of dynamic occlusion perception is arranged aiming at the occlusion problem of a teaching scene, and the expression recognition accuracy in the occlusion scene is improved. The whole-process intelligent teaching assistance of "state perception-strategy generation-interactive execution-effect evaluation" is realized, and the pertinence and effectiveness of teaching interaction are improved.

Description

Technical Field

[0001] This invention relates to the field of teaching device technology, and in particular to an interactive intelligent teaching device and method. Background Technology

[0002] In traditional teaching, teachers face two major pain points: First, it is difficult to perceive students' state of mind. With a large number of students in the classroom (typically 30-50 students), it is difficult for teachers to pay attention to each student's listening status at the same time. Second, the teaching interaction is not targeted enough. Although some teaching aids use image acquisition and facial expression recognition, when key facial areas are obscured due to hair, hands supporting the face, or wearing glasses, it is impossible to accurately judge whether students are focused or understand the knowledge points. As a result, teachers cannot develop interaction strategies based on students' true state and can only adopt a "one-size-fits-all" teaching model, which cannot meet the needs of personalized teaching.

[0003] With the acceleration of digital transformation in education, personalized and interactive teaching has become a key direction for improving teaching quality. There is an urgent need for an interactive intelligent teaching device and method that can accurately perceive student status (including occluded scenes), dynamically generate interactive strategies, and have flexible adaptability, so as to solve the core problems of "inaccurate perception, untargeted interaction, and inflexible equipment" in traditional teaching. Summary of the Invention

[0004] The purpose of this invention is to provide an interactive intelligent teaching device and method that accurately extracts students' facial expression features through a dynamic occlusion perception facial expression recognition model, transforms them into quantifiable learning psychological states, and combines flexible hardware devices and closed-loop interaction strategies to achieve intelligent teaching assistance throughout the entire process of "state perception - strategy generation - interaction execution - effect evaluation", thereby improving the pertinence and effectiveness of teaching interaction.

[0005] To achieve the above objectives, the present invention provides an interactive intelligent teaching device, including a processing server with a built-in AA-ViT facial expression recognition model that has dynamic occlusion perception, a dual-mode teaching blackboard, a teacher operation interactive terminal, a student interactive terminal set on the desktop, and an image acquisition unit. The image acquisition unit is installed in the classroom to capture student images and transmit them to the processing server. The server uses the AA-ViT facial expression recognition model with built-in dynamic occlusion perception to recognize facial expressions and determine the student group status based on the recognition results. Based on the student group status, the server determines the interaction strategy and sends the student group status and interaction strategy to the teacher's interactive terminal. The teacher's interactive terminal asks questions based on the facial expression recognition results. The teacher's interactive terminal communicates with the dual-mode teaching blackboard and recognizes the teacher's written content, then transmits the recognition results to the dual-mode teaching blackboard for display.

[0006] Preferably, the dual-mode teaching blackboard includes a fixed frame, in which an electronic blackboard and a mechanical blackboard are slidably installed. The back of the fixed frame is installed on the classroom wall via two angle-adjustable telescopic cylinders and several mechanical telescopic rods.

[0007] Preferably, the teacher's interactive terminal is installed on a following mechanism, which includes a following vehicle body, a mounting frame on the following vehicle body, a lifting and telescopic cylinder on the mounting frame, an angle adjustment motor on the top of the lifting and telescopic cylinder, and the output shaft of the angle adjustment motor is connected to the teacher's interactive terminal through a connector.

[0008] The specific steps of the method based on the above-mentioned interactive intelligent teaching device are as follows: Step S1: Acquire student images through an image acquisition agency and transmit them to the processing server; Step S2: The processing server preprocesses the collected student images, including face cropping, size normalization, and numbering, to obtain a standardized face image dataset. Step S3: Construct a dynamic occlusion-aware AA-ViT expression recognition model. Input the standardized face image dataset into the dynamic occlusion-aware AA-ViT expression recognition model to output the basic expression probability distribution matrix. Step S4: Calculate the students' psychological states based on the basic facial expression probability distribution matrix to obtain the students' psychological state dataset; Step S5: Develop an interaction strategy based on the student psychological state dataset, and push the interaction strategy to the teacher's interactive terminal for display. After setting a time, repeat steps S1-S4 to obtain the student psychological state dataset after interaction, calculate the improvement score, and send it to the teacher's interactive terminal for display.

[0009] Preferably, in step S2, the face in the image is located and the facial region is cropped using a multi-task cascaded convolutional neural network, and the resulting image size after size normalization is 224×224 pixels.

[0010] Preferably, in step S3, the AA-ViT expression recognition model with dynamic occlusion perception includes the ViT basic skeleton, the MSA module in the ViT basic skeleton is replaced with the ADA module, the MLP Head module is replaced with the DB-MLP module, and an Int Re module is added between the input layer and the normalization layer.

[0011] Preferably, the ADA module adaptively weights the standard dot product attention score using an adaptive weight matrix to initially suppress high-frequency responses in occluded regions, as shown in the following expression: ; in, Attention score matrix after adaptive weight calibration; For querying the matrix, The key matrix; ; For value matrices, For the input image features, This is a shared linear layer function used to generate the query matrix, key matrix, and value matrix; The key vector dimension for each attention head, C represents the feature dimension, and h represents the number of attention heads. For element-wise multiplication, This is an adaptive weight matrix; ; For activation function, It is an adaptive weighted linear layer function; The effective region is enhanced by using a mask, as shown in the following expression: ; in, For obscured mask Suppressed attention score matrix To mask the area, Set the attention score of the occluded region to negative infinity, and then apply the mask. An occlusion mask is generated using a random erasure strategy to simulate real-world learning scenarios where hair, hands, and glasses cause occlusion. The expression is as follows: ; in, For batch size, The length of the image pixel sequence. Indicates the obscured area. Indicates the valid region.

[0012] Preferably, the DB-MLP module is used to improve the accuracy of facial expression classification, and the expression is as follows: ; in, It is a non-linear activation function. For the first compressed linear layer function, This is the second compressed linear layer function. For normalization function, The ADA module outputs the facial feature vector after the normalization layer. Features of facial feature vectors The number of emoji types, Output categorized emojis. Let the probability distribution vector for facial expression classification be... , For the first The probability of expression types, .

[0013] Preferably, in step S4, the student's psychological state includes boredom, cognitive load, anxiety, and fatigue; The formula for calculating a single psychological state is as follows: ; in, For the first Psychological state For the first Facial expressions for the first Weighting coefficients for psychological states; Based on state classification threshold The expression for determining a student's psychological state is as follows: ; in, For the state classification results, Time indicates the first The psychological state is significant. Indicates the first The psychological state was not significant.

[0014] Preferably, in step S5, the data in the student psychological state dataset is sorted to determine the main state of the student group, and an interaction strategy is specified based on the main state of the student group. When the main state of the student group is bored, the interaction strategy is to ask questions and increase the speed of explanation. The question targets are randomly selected from students in the bored state, and the questions are extracted from a question bank of medium difficulty. When the dominant state of the student group is cognitive load, the interaction strategy is to reduce the speed of explanation; When the student group's dominant state is anxiety, the interaction strategy is to ask questions and slow down the explanation speed. The questions are randomly selected from students who are bored, and the questions are extracted from a question bank with corresponding low difficulty. When the student group is in a state of fatigue, the interaction strategy is to rest or switch the knowledge points to be explained. The knowledge points to be switched are extracted from the lesson plan with low difficulty. Students input their answers through the student interactive terminal, which then sends the answers to the teacher's interactive terminal. The teacher selects the correct answer or chooses the best answer from the students' answers and sends it to the student interactive terminal corresponding to the student who was asked the question.

[0015] Therefore, the present invention, employing the above-mentioned interactive intelligent teaching device and method, has the following beneficial effects: (1) To address the frequent issues of eyeglasses, hair, hands, and masks obscuring the face in teaching scenarios, an AA-ViT expression recognition model with dynamic occlusion perception was set up. Through the ADA module and Int Re module in the model, the accuracy of expression recognition in occlusion scenarios was improved.

[0016] (2) Based on the probability and weight coefficient of facial expressions, the quantitative calculation of four types of learning psychological states such as boredom and cognitive load is realized. Combined with the group's main state, targeted interactive strategies are generated, which significantly improves student participation and realizes the whole process of intelligent teaching assistance from "state perception - strategy generation - interactive execution - effect evaluation", thereby improving the pertinence and effectiveness of teaching interaction.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of an interactive intelligent teaching device according to the present invention; Figure 2 This is a schematic diagram of the dual-mode teaching blackboard structure of the present invention; Figure 3 This is a schematic diagram of the AA-ViT facial expression recognition model with dynamic occlusion perception according to the present invention.

[0019] Figure Labels 1. Processing server; 2. Dual-mode teaching blackboard; 21. Fixed frame; 22. Electronic blackboard; 23. Mechanical blackboard; 24. Angle adjustment telescopic cylinder; 3. Teacher operation interactive terminal; 4. Student interactive terminal; 5. Following mechanism; 51. Following vehicle body; 52. Mounting frame; 53. Lifting telescopic cylinder; 54. Angle adjustment motor; 6. Image acquisition unit. Detailed Implementation

[0020] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0022] like Figure 1 As shown, an interactive intelligent teaching device is characterized by including: a processing server with a built-in AA-ViT facial expression recognition model that has dynamic occlusion perception, a dual-mode teaching blackboard, a teacher operation interactive terminal, a student interactive terminal set on the desktop (using a 7-inch touch terminal, one per student, placed at the front of the desk, supporting answer input, status feedback, and lesson plan viewing functions), and an image acquisition unit.

[0023] The processing server has functions such as image preprocessing, facial expression recognition, psychological state calculation, interaction strategy generation, and effect evaluation. It adopts an Intel Core i7-13700K CPU, 32GB DDR5 memory, and NVIDIA GeForce RTX 4090 GPU, and supports parallel processing of student images at 30 frames per second, meeting the real-time recognition needs of a class of 50 students. Higher performance processors can be selected according to actual needs.

[0024] The image acquisition unit is installed in the classroom to capture student images and transmit them to the processing server. The server uses the AA-ViT facial expression recognition model with built-in dynamic occlusion perception to recognize facial expressions and determine the student group status based on the recognition results. Based on the student group status, the server determines the interaction strategy and sends the student group status and interaction strategy to the teacher's interactive terminal. The teacher's interactive terminal asks questions based on the facial expression recognition results. The teacher's interactive terminal communicates with the dual-mode teaching blackboard and recognizes the teacher's written content, then transmits the recognition results to the dual-mode teaching blackboard for display.

[0025] like Figure 2As shown, the dual-mode teaching blackboard includes a fixed frame, within which an electronic blackboard and a mechanical blackboard are slidably mounted. The electronic blackboard and the mechanical blackboard are positioned vertically. The back of the fixed frame is mounted to the classroom wall via two angle-adjustable telescopic cylinders and several mechanical telescopic rods, allowing for easy adjustment of the angle of the dual-mode teaching blackboard according to reflective conditions to adapt to different classroom lighting conditions. Simultaneously, the electronic blackboard is used to display classroom writing content and lesson plans, while the mechanical blackboard facilitates manual writing by the teacher.

[0026] The teacher-operated interactive terminal is installed on the following mechanism. It is a 15.6-inch touchscreen tablet running Android 14, supporting handwriting recognition, interactive strategy display, student answer reception and feedback, and includes a built-in 5-megapixel camera and dual speakers. Communication latency is ≤100ms. The following mechanism includes a following vehicle with a mounting frame. The mounting frame has a lifting and telescopic cylinder, and the top of the cylinder has an angle adjustment motor. The output shaft of the angle adjustment motor is connected to the teacher-operated interactive terminal via a connector, allowing for easy adjustment of the following position according to the classroom location.

[0027] The specific steps of the method based on the above-mentioned interactive intelligent teaching device are as follows: Step S1: Collect student images through an image acquisition agency and transmit them to the processing server.

[0028] Step S2: The processing server preprocesses the collected student images, including face cropping, size normalization, and numbering, to obtain a standardized face image dataset. A multi-task cascaded convolutional neural network is used to locate and crop the facial regions in the images. After size normalization, the resulting image size is 224×224 pixels.

[0029] Step S3: Construct a dynamic occlusion-aware AA-ViT expression recognition model. Input a standardized facial image dataset into the dynamic occlusion-aware AA-ViT expression recognition model to output the basic expression probability distribution matrix.

[0030] like Figure 3 As shown, the AA-ViT expression recognition model with dynamic occlusion awareness includes the ViT basic skeleton. The MSA module in the ViT basic skeleton is replaced with the ADA module, the MLP Head module is replaced with the DB-MLP module, and an Int Re module is added between the input layer and the normalization layer.

[0031] The ADA module adaptively weights the standard dot product attention score using an adaptive weight matrix to initially suppress high-frequency responses in occluded regions, as shown in the following expression: ; in, Attention score matrix after adaptive weight calibration; For querying the matrix, The key matrix; ; For value matrices, For the input image features, This is a shared linear layer function used to generate the query matrix, key matrix, and value matrix; The key vector dimension for each attention head, C represents the feature dimension, and h represents the number of attention heads. For element-wise multiplication, This is an adaptive weight matrix; ; For activation function, It is an adaptive weighted linear layer function; The effective region is enhanced by using a mask, as shown in the following expression: ; in, For obscured mask Suppressed attention score matrix To mask the area, Set the attention score of the occluded region to negative infinity, and then apply the mask. An occlusion mask is generated using a random erasure strategy to simulate real-world learning scenarios where hair, hands, and glasses cause occlusion. The expression is as follows: ; in, For batch size, The length of the image pixel sequence. Indicates the obscured area. Indicates the valid region.

[0032] The DB-MLP module is used to improve the accuracy of facial expression classification. The expression is as follows: ; in, It is a non-linear activation function. For the first compressed linear layer function, This is the second compressed linear layer function. For normalization function, The ADA module outputs the facial feature vector after the normalization layer. Features of facial feature vectors The number of expression types is 8 basic expressions, including anger, contempt, disgust, fear, happiness, neutrality, sadness, and surprise.

[0033] Output categorized emojis. Let the probability distribution vector for facial expression classification be... , For the first The probability of expression types, .

[0034] Step S4: Calculate students' psychological states based on the basic facial expression probability distribution matrix to obtain a student psychological state dataset. Student psychological states include boredom, cognitive load, anxiety, and fatigue.

[0035] The formula for calculating a single psychological state is as follows: ; in, For the first Psychological state For the first Facial expressions for the first Weighting coefficients for the four psychological states. The weighting coefficients for the four psychological states are shown in Table 1.

[0036] Table 1. Weighting coefficients of the four psychological states

[0037] Based on state classification threshold The expression for determining a student's psychological state is as follows: ; in, For the state classification results, Time indicates the first The psychological state is significant. Indicates the first The mental states are not significant. The classification threshold obtained through five-fold cross-validation is the critical value for "significant / insignificant state" in the teaching scenario. The state classification threshold values ​​are shown in Table 2.

[0038] Table 2. Threshold values ​​for state classification

[0039] Step S5: Develop an interaction strategy based on the student psychological state dataset and push the strategy to the teacher's interactive terminal for display. After a set time, repeat steps S1-S4 to obtain the student psychological state dataset after interaction. Calculate the improvement score and send it to the teacher's interactive terminal for display. In step S5, the data in the student psychological state dataset is sorted to determine the dominant state of the student group, and the interaction strategy is developed based on the dominant state of the student group.

[0040] When the main state of the student group is bored, the interaction strategy is to ask questions and increase the speed of explanation. The question targets are randomly selected from students in the bored state, and the questions are extracted from a question bank of medium difficulty.

[0041] When the student group's dominant state is cognitive load, the interaction strategy is to reduce the speed of explanation.

[0042] When the student group's dominant state is anxiety, the interaction strategy is to ask questions and slow down the explanation speed. The questions are randomly selected from students who are bored, and the questions are extracted from a question bank with corresponding low difficulty.

[0043] When the student group is in a state of fatigue, the interaction strategy is to rest or switch the knowledge points to be explained. The knowledge points to be switched are extracted from the lesson plan with lower difficulty.

[0044] Students input their answers through the student interactive terminal, which then sends the answers to the teacher's interactive terminal. The teacher selects the correct answer or chooses the best answer from the students' answers and sends it to the student interactive terminal corresponding to the student who was asked the question.

[0045] The recognition test was conducted in a classroom that can accommodate 50 students, and the results are shown in Table 3. Table 3 Test Results

[0046] Eight students showed no significant state, indicating that the dominant state of the student group was boredom. The interaction strategies involved asking questions and increasing the speed of explanation. Questions were randomly selected from students in the bored state and were drawn from a question bank of moderate difficulty. After the interaction, images were collected again for recognition. The test results after the interaction are shown in Table 4.

[0047] Table 4. Post-interaction test results

[0048] Among them, 22 people had no significant state, which reduced the number of people in the bored state.

[0049] We also tested it against existing mature models, where the scene had 40% occlusion from glasses, 50% from hair, and 10% from hands. The test data is as follows:

[0050] For the most common obstacles in teaching, such as hair, glasses, and hands obstructing the view, this embodiment achieves an accuracy rate of over 90%, which meets the needs of teaching.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An interactive intelligent teaching device, characterized in that: It includes a processing server with a built-in AA-ViT facial expression recognition model that has dynamic occlusion perception, a dual-mode teaching blackboard, a teacher operation interactive terminal, a student interactive terminal set on the desktop, and an image acquisition unit. The image acquisition unit is installed in the classroom to capture student images and transmit them to the processing server. The server uses the AA-ViT facial expression recognition model with built-in dynamic occlusion perception to recognize facial expressions and determine the student group status based on the recognition results. Based on the student group status, the server determines the interaction strategy and sends the student group status and interaction strategy to the teacher's interactive terminal. The teacher's interactive terminal asks questions based on the facial expression recognition results. The teacher's interactive terminal communicates with the dual-mode teaching blackboard and recognizes the teacher's written content, then transmits the recognition results to the dual-mode teaching blackboard for display.

2. The interactive intelligent teaching device according to claim 1, characterized in that: The dual-mode teaching blackboard includes a fixed frame, within which an electronic blackboard and a mechanical blackboard are slidably installed. The back of the fixed frame is mounted to the classroom wall via two angle-adjustable telescopic cylinders and several mechanical telescopic rods.

3. The interactive intelligent teaching device according to claim 2, characterized in that: The teacher's interactive terminal is installed on the following mechanism, which includes a following vehicle body. The following vehicle body is equipped with a mounting frame, and the mounting frame is equipped with a lifting and telescopic cylinder. The top of the lifting and telescopic cylinder is equipped with an angle adjustment motor, and the output shaft of the angle adjustment motor is connected to the teacher's interactive terminal through a connector.

4. A method based on the interactive intelligent teaching device according to claim 3, characterized in that, The specific steps are as follows: Step S1: Acquire student images through an image acquisition agency and transmit them to the processing server; Step S2: The processing server preprocesses the collected student images, including face cropping, size normalization, and numbering, to obtain a standardized face image dataset. Step S3: Construct a dynamic occlusion-aware AA-ViT expression recognition model. Input the standardized face image dataset into the dynamic occlusion-aware AA-ViT expression recognition model to output the basic expression probability distribution matrix. Step S4: Calculate the students' psychological states based on the basic facial expression probability distribution matrix to obtain the students' psychological state dataset; Step S5: Develop an interaction strategy based on the student psychological state dataset, and push the interaction strategy to the teacher's interactive terminal for display. After setting a time, repeat steps S1-S4 to obtain the student psychological state dataset after interaction, calculate the improvement score, and send it to the teacher's interactive terminal for display.

5. The method based on an interactive intelligent teaching device according to claim 4, characterized in that: In step S2, the face in the image is located and the facial region is cropped using a multi-task cascaded convolutional neural network. After size normalization, the resulting image size is 224×224 pixels.

6. The method based on an interactive intelligent teaching device according to claim 3, characterized in that: In step S3, the AA-ViT expression recognition model with dynamic occlusion awareness includes the ViT basic skeleton. The MSA module in the ViT basic skeleton is replaced with the ADA module, the MLP Head module is replaced with the DB-MLP module, and an Int Re module is added between the input layer and the normalization layer.

7. The method based on an interactive intelligent teaching device according to claim 3, characterized in that: The ADA module adaptively weights the standard dot product attention score using an adaptive weight matrix to initially suppress high-frequency responses in occluded regions, as shown in the following expression: ； in, Attention score matrix after adaptive weight calibration; For querying the matrix, The key matrix; ； For value matrices, For the input image features, This is a shared linear layer function used to generate the query matrix, key matrix, and value matrix; The key vector dimension for each attention head, C represents the feature dimension, and h represents the number of attention heads. For element-wise multiplication, This is an adaptive weight matrix; ； For activation function, It is an adaptive weighted linear layer function; The effective region is enhanced by using a mask, as shown in the following expression: ； in, For obscured mask Suppressed attention score matrix To mask the area, Set the attention score of the occluded region to negative infinity, and then apply the mask. An occlusion mask is generated using a random erasure strategy to simulate real-world learning scenarios where hair, hands, and glasses cause occlusion. The expression is as follows: ； in, For batch size, The length of the image pixel sequence. Indicates the obscured area. Indicates the valid region.

8. The method based on an interactive intelligent teaching device according to claim 3, characterized in that: The DB-MLP module is used to improve the accuracy of facial expression classification. The expression is as follows: ； in, It is a non-linear activation function. For the first compressed linear layer function, This is the second compressed linear layer function. For normalization function, The ADA module outputs the facial feature vector after the normalization layer. Features of facial feature vectors The number of emoji types, Output categorized emojis. Let the probability distribution vector for facial expression classification be... , For the first The probability of expression types, .

9. The method based on an interactive intelligent teaching device according to claim 3, characterized in that: In step S4, the students' psychological states include boredom, cognitive load, anxiety, and fatigue; The formula for calculating a single psychological state is as follows: ； in, For the first Psychological state For the first Facial expressions for the first Weighting coefficients for psychological states; Based on state classification threshold The expression for determining a student's psychological state is as follows: ； in, For the state classification results, Time indicates the first The psychological state is significant. Indicates the first The psychological state was not significant.

10. The method based on an interactive intelligent teaching device according to claim 3, characterized in that: In step S5, the data in the student psychological state dataset is sorted to determine the main state of the student group, and an interaction strategy is specified based on the main state of the student group. When the main state of the student group is bored, the interaction strategy is to ask questions and increase the speed of explanation. The question targets are randomly selected from students in the bored state, and the questions are extracted from a question bank of medium difficulty. When the dominant state of the student group is cognitive load, the interaction strategy is to reduce the speed of explanation; When the student group's dominant state is anxiety, the interaction strategy is to ask questions and slow down the explanation speed. The questions are randomly selected from students who are bored, and the questions are extracted from a question bank with corresponding low difficulty. When the student group is in a state of fatigue, the interaction strategy is to rest or switch the knowledge points to be explained. The knowledge points to be switched are extracted from the lesson plan with low difficulty. Students input their answers through the student interactive terminal, which then sends the answers to the teacher's interactive terminal. The teacher selects the correct answer or chooses the best answer from the students' answers and sends it to the student interactive terminal corresponding to the student who was asked the question.

Images