Classroom Behavior Recognition Using Computer Vision: A Systematic Review

Area of Science:

• Educational Technology
• Computer Vision
• Behavioral Science

Background:

• Behavioral computing using visual cues is vital for real-time classroom state analysis.
• A lack of consensus exists regarding the status and future of computer vision-based classroom behavior recognition.

Purpose of the Study:

• To systematically review the research status and future trends of computer vision-based classroom behavior recognition.
• To address research questions on goal orientation, recognition techniques, and challenges.

Main Methods:

• Systematic literature review of 80 peer-reviewed journal articles.
• Adherence to Preferred Reporting Items for Systematic Assessment and Meta-Analysis (PRISMA) guidelines.

Main Results:

• Recognition targets include physical action, learning engagement, attention, and emotion, with focus on the former two.
• Behavioral categorizations lack standardization and connection to instructional content.
• Studies predominantly focus on college students in traditional classrooms.
• Deep learning, especially the YOLO series, is the primary recognition method.
• Identified challenges in experimental design, recognition methods, practical applications, and pedagogical research.

Conclusions:

• Computer vision offers potential for classroom behavior analysis but requires methodological and pedagogical advancements.
• Future research should address current limitations for broader and more effective application.