Scholarlink

With the increasing popularity and preference of online classes, teachers are at a disadvantage of losing face-time with students and are faced with the challenge of monitoring their students’ attention levels remotely. This research, VisioFocus, presents a non-invasive student monitoring system using deep learning and computer vision to measure and analyze the attention levels of students. It has two main components namely Pose Estimation and Emotion Recognition which processes the captured facial images of the student through the webcam at 15 frames per second to measure the attention level. The Pose Estimation model uses pure computer vision and the Emotion Recognition model uses a modified ResNet deep convolutional network that could classify emotions portrayed into 6 different labels at about 70% accuracy. This research proposes a novel and reliable approach to measuring attention levels, which will have a significant impact on the education field. The analytical survey of attention levels after each lecture will enable educators to tailor their teaching approach to specific students, resulting in better academic development.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Online Learning: A Pose and Emotion-Based Approach for Student Attention Monitoring

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sarath Krishna Chingapurathu,
  • L. Jeganathan,
  • M. Janaki Meena

摘要

With the increasing popularity and preference of online classes, teachers are at a disadvantage of losing face-time with students and are faced with the challenge of monitoring their students’ attention levels remotely. This research, VisioFocus, presents a non-invasive student monitoring system using deep learning and computer vision to measure and analyze the attention levels of students. It has two main components namely Pose Estimation and Emotion Recognition which processes the captured facial images of the student through the webcam at 15 frames per second to measure the attention level. The Pose Estimation model uses pure computer vision and the Emotion Recognition model uses a modified ResNet deep convolutional network that could classify emotions portrayed into 6 different labels at about 70% accuracy. This research proposes a novel and reliable approach to measuring attention levels, which will have a significant impact on the education field. The analytical survey of attention levels after each lecture will enable educators to tailor their teaching approach to specific students, resulting in better academic development.