Analysis of Students' Role Perceptions and their Tendencies in Classroom Education Based on Visual Inspection

Area of Science:

• Computer Science
• Educational Technology
• Artificial Intelligence

Background:

• Student engagement is crucial for effective learning, yet traditional assessment methods in classrooms are limited.
• While computer vision is used for online learning analysis, its application in physical classrooms for engagement monitoring is less explored.
• Understanding student engagement patterns, such as declining motivation over time, is vital for pedagogical interventions.

Purpose of the Study:

• To propose and evaluate a multi-example learning student engagement assessment method for classroom education.
• To analyze students' role perceptions and engagement tendencies using visual inspection and a one-dimensional convolutional neural network (1D CNN).
• To investigate the effectiveness of a classroom attention evaluation detection system in real-world teaching scenarios.

Main Methods:

• Utilized visual features like head posture, eye gaze, and facial movements for engagement assessment.
• Employed a 1D CNN with multi-example learning and a multi-example pooling layer for analyzing video features.
• Extracted relative change features from video using the Open Face toolset to capture dynamic visual cues.
• Applied the developed system to actual classroom teaching activities for validation.

Main Results:

• The proposed method accurately assesses student engagement by analyzing visual cues from classroom videos.
• Identified a common student engagement trend: high motivation at the start, followed by increased absenteeism later.
• The system demonstrated effectiveness and accuracy when applied in real classroom settings, validated through interviews.

Conclusions:

• The developed 1D CNN-based system offers a viable approach for monitoring student engagement in physical classrooms.
• Visual feature analysis provides valuable insights into student attention and motivation dynamics.
• The study highlights the potential of computer vision technologies to enhance educational assessment and intervention strategies.