Advanced Student Monitoring System for Physical Classes Using Pattern Recognition and Behavioral Analysis

This research aims to develop an intelligent classroom system that utilizes Computer Vision and Deep Learning to measure students' concentration levels and performance, enabling teachers to gain insight into overall performance and adjust their teaching accordingly. The main goal of this research is to develop a precise way of tracking student concentration level and automate attendance marking in physical classroom. The proposed system consists of four components. To take attendance, it utilizes OpenCV and dlib for precise face recognition and eye gaze tracking to measure student focus levels. In the emotion analysis component, it employs DeepFace, dlib, and MediaPipe to identify students' emotional states and behaviors. For improving low-resolution video footage, the low-resolution enhancement component combines dlib and ESRGANG. In the concentration level component, Convolutional Neural Networks (CNN) are used to build a precise model for assessing students' concentration levels. The implemented system offers real-time attendance tracking, emotional and behavioral insights, and updates on students' concentration levels. It improves the detection of students at the back of the classroom using low-resolution footage. These features support educators in identifying students who need assistance, adjusting teaching methods, and creating a more engaging learning environment, ultimately enhancing classroom performance.