Classroom learning situation analysis method based on computer vision

Abstract

The present invention relates to a computer vision-based classroom learning situation analysis method, comprising the following steps: step S1: obtaining human body image data, and performing preprocessing to obtain a human body image data set; step S2: training according to the obtained human body image data set The YOLOv4 target detection algorithm model, and based on the trained YOLOv4 target detection algorithm model, performs frame-by-frame human detection on the classroom real-time video, and obtains the human body position information of each target; Step S3: According to the obtained human body position information of each target, use CrowdPose The crowded scene pose estimation algorithm carries out pose estimation, and obtains the distribution information of skeleton key points of each target; Step S4: according to the obtained distribution information of skeleton key points of each target, conduct behavior detection and emotion analysis respectively; Step S5: according to behavior detection , face recognition, sentiment analysis results, further achieve target classroom attendance and gain recognition target classroom participation. The present invention effectively acquires the classroom status of each student for teachers to improve teaching strategies and improve the quality of classroom teaching.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a computer vision-based classroom learning situation analysis method. Background technique [0002] With the continuous maturity of artificial intelligence scenario-based applications, the importance and possibility of education + AI attracts more and more scholars to devote themselves to research. At the same time, the drawbacks presented by traditional education are slowly emerging, such as the difficulty of attendance due to large-class teaching, and the lack of time for teachers to take care of the status of many students in class. In order to solve the above pain points, the present invention proposes a classroom learning situation analysis method based on computer vision technologies such as object detection and attitude estimation. Learning situation analysis, using the data analysis of students' classroom status as a reference for teachers to improve teaching strategies. ...

AI Technical Summary

Problems solved by technology

At the same time, the disadvantages of traditional education are slowly emerging, such as the difficulty of attendance due to large class teaching, and the lack of time for teachers to take care of the class status of many students, etc.

Images