<![CDATA[This study originates from the need for a real-time student concentration monitoring system in response to challenges in maintaining learner engagement during the instructional process. The aim of the research is to develop Smart Focus Detection, an early warning system based on deep learning and computer vision designed to automatically detect students’ focus levels. The system was developed using a modern client–server architecture integrating the YOLOv8 behavior detection model, a FastAPI-based backend service, and a Vue.js-based user interface dashboard. Development followed the Agile methodology through the stages of planning, design, implementation, and testing. The YOLOv8 model was trained using public datasets from Kaggle and Roboflow, with data preprocessing and augmentation techniques applied, and evaluated using metrics including Precision, Recall, F1 Score, and Mean Average Precision (mAP). Results show that the model achieved an mAP@0.5 score of 85%, indicating high accuracy in detecting both focused and unfocused student behaviors. Limited trials in classroom settings demonstrated that the system’s interactive dashboard effectively displays data through live monitoring features, statistical visualizations, and automated notifications. The study concludes that Smart Focus Detection holds strong potential as a proactive pedagogical assistant, enabling teachers to conduct timely interventions to enhance student engagement and learning effectiveness. The findings imply promising opportunities for leveraging artificial intelligence in the development of data-driven adaptive learning systems.]]>
Copyrights © 2025
