CN-115659221-B - Teaching quality assessment method and device and computer readable storage medium

CN115659221BCN 115659221 BCN115659221 BCN 115659221BCN-115659221-B

Abstract

The invention provides a method and a device for evaluating teaching quality and a computer readable storage medium, wherein the method comprises the steps of collecting multi-modal data of students and teachers in a preset time period, wherein the multi-modal data comprise student brain wave data, student behavior data, teacher behavior data, student expression data and teacher expression data, inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality evaluation score, and sending the teaching quality evaluation score to a classroom terminal to enable the classroom terminal to receive and display the predicted teaching quality evaluation score. The method and the device can solve the problem that the accuracy of the evaluation result is low because only one mode data is used or only the mode data on one side of a student or a teacher is considered in the evaluation of the traditional teaching quality evaluation method.

Inventors

WU HAORAN
LI YAMENG
WANG WEI
LIAO JUN

Assignees

中国联合网络通信集团有限公司

Dates

Publication Date: 20260512
Application Date: 20221031

Claims (11)

1. A method for evaluating teaching quality, comprising: collecting multi-mode data of students and teachers in a preset time period, wherein the multi-mode data comprises student brain wave data, student behavior data, teacher behavior data, student expression data and teacher expression data; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score; Transmitting the teaching quality assessment score to a classroom terminal, so that the classroom terminal receives and displays the predicted teaching quality assessment score; the trained multi-modal feature fusion classification model comprises a first layer, a second layer, a third layer, a fourth layer, a multi-modal feature fusion layer and a full connection layer; The outputs of the first layer, the second layer, the third layer and the fourth layer are respectively connected with the input of the multi-mode feature fusion layer, and the output of the multi-mode feature fusion layer is connected with the input of the full-connection layer; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score, which specifically comprises the following steps: inputting the student brain wave data into the first layer, and extracting the student brain wave characteristics through the first layer; inputting the student behavior data and the teacher behavior data into the second layer, and extracting fusion characteristics of the student behavior and the teacher behavior through the second layer; Inputting the student expression data into the third layer, and extracting teacher behaviors and student expression fusion characteristics through the third layer; inputting the teacher expression data into the fourth layer, and extracting the teacher expression characteristics through the fourth layer; inputting the features extracted by the first layer, the second layer, the third layer and the fourth layer into the multi-mode feature fusion layer for normalization processing; inputting the normalized characteristics into the full-connection layer for full connection to obtain the predicted teaching quality evaluation score; Inputting the student behavior data and the teacher behavior data into the second layer, and extracting fusion characteristics of the student behavior and the teacher behavior through the second layer, wherein the method specifically comprises the following steps: Inputting student behavior data in 10 x m x n dimensions and teacher behavior data in 10 x m x n dimensions into the hourglass Hourglass model of the second layer to detect key points of a human body, so as to obtain a detection result of the key points of the human body; inputting the detection result of the human body key points into a cyclic neural network RNN for behavior recognition to respectively obtain 1*n-dimensional student behavior characteristics and 1*n-dimensional teacher behavior characteristics; and fusing the 1*n-dimensional student behavior characteristics and the 1*n-dimensional teacher behavior characteristics by using an L2 paradigm to obtain 1*n-dimensional fused behavior characteristics.
2. The method of claim 1, wherein after collecting the multi-modal data of the student and teacher over the predetermined period of time, the method further comprises: aligning the multi-mode data according to the acquisition time stamp; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score, which specifically comprises the following steps: And inputting the aligned multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score.
3. The method of claim 1, wherein before inputting the multimodal data into a pre-trained multimodal feature fusion classification model to obtain a predicted quality of teaching assessment score, the method further comprises: Dividing multi-mode data of students and teachers in the hall pupil time according to a certain time step; training the multi-modal feature fusion classification model based on the segmented multi-modal data and the corresponding teaching quality scoring result to obtain the trained multi-modal feature fusion classification model.
4. The method according to claim 1, wherein the first layer includes an activation function layer and a normalization layer, and the inputting the student brain wave data into the first layer, extracting the student brain wave features through the first layer specifically includes: inputting 1*n-dimensional student brain wave data into the activation function layer for activation treatment to obtain an output result of the activation function layer; And inputting an output result of the activation function layer into the normalization layer for normalization processing to obtain 1*n-dimensional features after normalization processing, and taking the 1*n-dimensional features after normalization processing as the brain wave features of the students.
5. The method according to claim 1, wherein the inputting the student expression data into the third layer, extracting teacher behavior and student expression fusion features through the third layer, specifically comprises: Inputting student expression data in 10 x m x n dimensions into the third layer, and extracting expression features by using a convolutional neural network CNN model shared by the third layer and the fourth layer to obtain 1*n-dimension student expression features; and fusing the 1*n-dimensional student expression characteristics and the 1*n-dimensional teacher behavior characteristics according to priori knowledge to obtain 1*n-dimensional teacher behavior and student expression fusion characteristics.
6. The method according to claim 5, wherein the inputting the teacher expression data into the fourth layer and extracting the teacher expression feature through the fourth layer specifically includes: Inputting 10 x m x n-dimensional teacher expression data into the fourth layer, and extracting expression features by using the CNN model to obtain 1*n-dimensional teacher expression features.
7. The method of claim 1, wherein the normalization process is formulated as: Wherein, the For the features extracted by the i-th layer, Training weights corresponding to the ith layer.
8. The method according to claim 7, wherein the features after normalization processing of the multi-modal feature fusion layer are multi-modal fusion features of 1 x 4n dimensions, and the inputting the features after normalization processing of the multi-modal feature fusion layer into the full connection layer for full connection, to obtain the predicted teaching quality assessment score, specifically includes: and inputting the 1 x 4 n-dimensional multi-mode fusion characteristics into a 4n x 11-dimensional full-connection layer for full connection to obtain a predicted 1 x 11-dimensional teaching quality evaluation score.
9. An apparatus for evaluating teaching quality, comprising: The system comprises a modal data acquisition module, a data processing module and a data processing module, wherein the modal data acquisition module is used for acquiring multi-modal data of students and teachers in a preset time period, and the multi-modal data comprise student brain wave data, student behavior data, teacher behavior data, student expression data and teacher expression data; the evaluation score prediction module is connected with the modal data acquisition module and is used for inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality evaluation score; The receiving and displaying module is connected with the assessment score predicting module and used for sending the teaching quality assessment score to a classroom terminal so that the classroom terminal receives and displays the predicted teaching quality assessment score; the trained multi-modal feature fusion classification model comprises a first layer, a second layer, a third layer, a fourth layer, a multi-modal feature fusion layer and a full connection layer; The outputs of the first layer, the second layer, the third layer and the fourth layer are respectively connected with the input of the multi-mode feature fusion layer, and the output of the multi-mode feature fusion layer is connected with the input of the full-connection layer; the evaluation score prediction module specifically comprises: the first processing unit is used for inputting the student brain wave data into the first layer and extracting the student brain wave characteristics through the first layer; the second processing unit is used for inputting the student behavior data and the teacher behavior data into the second layer, and extracting fusion characteristics of the student behavior and the teacher behavior through the second layer; The third processing unit is used for inputting the student expression data into the third layer, and extracting teacher behaviors and student expression fusion characteristics through the third layer; a fourth processing unit, configured to input the teacher expression data into the fourth layer, and extract the teacher expression feature through the fourth layer; A fifth processing unit, configured to input the features extracted by the first layer, the second layer, the third layer, and the fourth layer into the multi-modal feature fusion layer for normalization processing; The sixth processing unit is used for inputting the normalized characteristics into the full-connection layer for full connection to obtain the predicted teaching quality evaluation score; the second processing unit specifically includes: The key point detection unit is used for inputting student behavior data in 10 x m x n dimensions and teacher behavior data in 10 x m x n dimensions into the hourglass Hourglass model of the second layer to detect key points of a human body, so as to obtain a detection result of the key points of the human body; the behavior recognition unit is used for inputting the detection result of the human body key points into the cyclic neural network RNN to perform behavior recognition to respectively obtain 1*n-dimensional student behavior characteristics and 1*n-dimensional teacher behavior characteristics; And the behavior feature fusion unit is used for fusing the 1*n-dimensional student behavior features and the 1*n-dimensional teacher behavior features by using the L2 paradigm to obtain 1*n-dimensional fused behavior features.
10. An assessment device for teaching quality, characterized by comprising a memory and a processor, the memory having stored therein a computer program, the processor being arranged to run the computer program to implement the method for assessment of teaching quality according to any of claims 1-8.
11. A computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements a method for evaluating teaching quality according to any of claims 1-8.

Description

Teaching quality assessment method and device and computer readable storage medium Technical Field The present invention relates to the field of teaching technologies, and in particular, to a method and apparatus for evaluating teaching quality, and a computer readable storage medium. Background The assessment of teacher teaching quality is a major concern in the current education community, and the analysis of various data related to the teacher teaching quality by using methods such as big data and AI algorithm is a development direction actively explored in the current education community. However, the existing teaching quality evaluation method is easy to cause the problem of low accuracy of the evaluation result due to the fact that only one mode data is used or only one mode data of a student or a teacher is considered in the evaluation. Disclosure of Invention The technical problem to be solved by the invention is to provide a teaching quality assessment method, a teaching quality assessment device and a computer readable storage medium aiming at the defects of the prior art, so as to solve the problem that the accuracy of an assessment result is low easily caused by using only one mode data or considering only the mode data of a single side of a student or a teacher in the existing teaching quality assessment method. In a first aspect, the present invention provides a method for evaluating teaching quality, including: collecting multi-mode data of students and teachers in a preset time period, wherein the multi-mode data comprises student brain wave data, student behavior data, teacher behavior data, student expression data and teacher expression data; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score; and sending the teaching quality assessment score to a classroom terminal so that the classroom terminal receives and displays the predicted teaching quality assessment score. Further, after the multi-mode data of the students and the teachers in the preset time period is collected, the method further includes: aligning the multi-mode data according to the acquisition time stamp; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score, which specifically comprises the following steps: And inputting the aligned multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score. Further, before the multi-modal data is input into the pre-trained multi-modal feature fusion classification model to obtain the predicted teaching quality evaluation score, the method further includes: Dividing multi-mode data of students and teachers in the hall pupil time according to a certain time step; training the multi-modal feature fusion classification model based on the segmented multi-modal data and the corresponding teaching quality scoring result to obtain the trained multi-modal feature fusion classification model. Further, the trained multi-modal feature fusion classification model comprises a first layer, a second layer, a third layer, a fourth layer, a multi-modal feature fusion layer and a full connection layer; The outputs of the first layer, the second layer, the third layer and the fourth layer are respectively connected with the input of the multi-mode feature fusion layer, and the output of the multi-mode feature fusion layer is connected with the input of the full-connection layer; Inputting the multi-modal data into a pre-trained multi-modal feature fusion classification model to obtain a predicted teaching quality assessment score, which specifically comprises the following steps: inputting the student brain wave data into the first layer, and extracting the student brain wave characteristics through the first layer; inputting the student behavior data and the teacher behavior data into the second layer, and extracting fusion characteristics of the student behavior and the teacher behavior through the second layer; Inputting the student expression data into the third layer, and extracting teacher behaviors and student expression fusion characteristics through the third layer; inputting the teacher expression data into the fourth layer, and extracting the teacher expression characteristics through the fourth layer; inputting the features extracted by the first layer, the second layer, the third layer and the fourth layer into the multi-mode feature fusion layer for normalization processing; and inputting the normalized characteristics into the full-connection layer for full connection to obtain the predicted teaching quality evaluation score. Further, the first layer includes an activation function layer and a normalization layer, the student brain wave data is input into the first layer, and the student brain wave features are extra