TeacherEye: An AI-Powered System forMonitoring Student Engagement in OnlineEducation

Abstract In online education, maintaining student engagement and academic integrity poses significant challenges. While multiple solutions have been introduced, none have provided a comprehensive system. We introduce TeacherEye, an Artificial Intelligence (AI) powered system for monitoring student behavior and engagement in virtual classrooms and exams. TeacherEye integrates State-Of-The-Art (SOTA) AI models to analyze video data from the webcam stream and audio data from the microphone in real time. The system first verifies students through face detection and identification. It then monitors behavior using advanced vision features like sleep detection, multi-person detection, phone usage detection, focus tracking, and emotion recognition. Audio is simultaneously converted to text and analyzed by a Large Language Model (LLM) to detect suspicious behavior, enhancing online exam proctoring. If irregularities are found, a Vision-Language Model (VLM) suggests actions to the teacher, while all detected behavior is logged for review. TeacherEye presents a comprehensive solution for improving engagement and ensuring integrity in online education.