Glancee: An Adaptable System for Instructors to Grasp Student Learning Status in Synchronous Online Classes

Document Title

Glancee: An Adaptable System for Instructors to Grasp Student Learning Status in Synchronous Online Classes

Document Information

• Subject Area: Online Learning, User Experience (HCI)
• Keywords: Synchronous online classes, learning status detection, instructor-adaptable interface design, affective computing, video conferencing, mixed-method research

Research Background and Issues

• Problems and Challenges:

  • Synchronous online learning is increasingly prevalent, especially during the COVID-19 pandemic, but instructors struggle to observe students' learning status through video in online classes.
  • Students often refuse to turn on their cameras due to privacy concerns, and existing video conferencing tools (e.g., Zoom) fail to effectively support instructors in real-time adjustment of teaching content.
  • Current teaching support systems are primarily applied in offline or asynchronous scenarios, lacking analysis of real-world needs in online real-time teaching contexts.

• Research Significance:

  • The absence of face-to-face interaction weakens teacher-student connections, affecting teaching effectiveness.
  • Exploring system designs to address teaching needs in online classrooms has academic value and application potential.

• Motivation and Related Work:

  • Existing work, based on traditional educational theories, has proposed various methods for detecting students' learning status.
  • Two major research gaps exist:
    1. Lack of empirical studies on the types of student learning statuses expected by instructors.
    2. Lack of flexible systems adaptable to different instructor preferences.

Solution

• Methods and Solutions:

  • The proposed Glancee system combines computer vision algorithms and customizable interfaces to detect and display students' learning status in real time.
  • Glancee features a sidebar-based interface that allows instructors to customize the display of learning status, including information types, data visualization formats, and notification mechanisms.

• Innovations:

  • Integrates multiple learning status detection algorithms (e.g., engagement, emotion, gaze behavior).
  • Provides adjustable instructor-adaptable interfaces, supporting "post-class review" functionality.
  • Designs teacher-centered visualization methods and lightweight notification mechanisms to avoid excessive distraction.

• Implementation Steps and Technology:

  • Student Side: Information is collected via cameras, with computer vision algorithms running in real time to detect learning status while protecting privacy.
  • Instructor Side and Server:
    1. The server receives anonymized student status data and generates overall classroom status.
    2. The instructor side displays real-time data and allows instructors to review detailed classroom statistics.

• User Configuration Features:

  • Customizable display formats (e.g., pie charts, bar charts).
  • Customizable notification mechanisms: e.g., whether to notify and notification frequency.
  • Provides timestamp alignment of course status with teaching slides after class.

Research Outcomes

• Specific Results:

  • Glancee supports efficient feedback on students' multi-level learning statuses (emotion, focus, engagement, etc.).
  • The system was evaluated as significantly superior to two baseline methods (EngageClass and ZoomOnly), demonstrating great potential in helping instructors improve teaching quality.

• System Advantages:

  • Innovative adaptable interface design meets personalized needs, allowing instructors to avoid reliance on uniform, fixed learning status displays.
  • Simultaneously satisfies real-time feedback and post-class information review needs, with good adaptability to different teaching scenarios.

• Experimental Results:

  • Survey: A survey of 67 instructors and 62 students clarified core functional design requirements.
    • Instructors were most concerned about student statuses such as engagement, confusion, emotion, and gaze behavior.
    • Students had significant privacy concerns, requiring data to be anonymized.
  • User Experiment:
    • Validated system usability (e.g., average ratings higher than baselines) and teaching improvement effects (easier to focus on student reactions) among 18 instructors and 53 students.
    • Observed instructors adjusting teaching pace in dynamic classrooms, proving the system's positive impact on instructor behavior and perception.

• Limitations and Future Directions:

  • Limitations:
    • Computer vision algorithms may be affected by lighting and camera angles, potentially leading to unstable performance in different environments.
    • Experiments were limited to formal courses, excluding discussion-based or interactive courses.
  • Future Improvements:
    • Optimization for hybrid learning models (online + offline).
    • Long-term deployment experiments for continuous improvement of teaching effects.
    • Exploration of more AR- or immersive-based teaching support interface designs.

Conclusion

• This study showcases the Glancee system and its practical application potential, emphasizing the importance of flexible adaptability and real-time feedback.
• The research results provide insights for the development of future intelligent teaching support tools, effectively promoting teacher-student communication and improving teaching quality in online classrooms.