Student Distraction Detection and Concentration Analysis via Transfer Learned Convolutional Neural Networks

In the midst of the significant shift from offline to online education, accurately measuring student concentration and identifying distractions have become crucial. Many institutions face difficulties in monitoring their students remotely. With limited existing research addressing this issue comprehensively, this work offers a groundbreaking study on the application of deep learning techniques to tackle these challenges. Leveraging a newly curated dataset specifically designed for the online learning environment, this study employs advanced convolutional neural networks (CNNs) such as DenseNet201, InceptionV3, Xception, and EfficientNetB6 through transfer learning methodologies. Extensive experimentation and analysis were conducted across 15 different classes, resulting in an impressive accuracy of 98.27%. This work not only makes a vital contribution to educational technology but also provides valuable insights into enhancing student engagement and concentration in the digital age. By demonstrating the efficacy of these advanced CNN models, this study paves the way for more effective monitoring and support systems in online education, ensuring that students remain focused and engaged in their virtual learning environments.