The rapid growth of online education has emphasized the importance of maintaining active student engagement to ensure effective learning outcomes. Traditional engagement detection methods use manual observation and they are time-consuming for large-scale datasets. This research presents a Quantum-Enhanced Facial Engagement Recognition (Q-FER) system to overcome these constraints. This framework is inspired by the principles of Quantum Machine Learning (QML). The suggested system merges the advantages of traditional Deep Learning (DL) and Quantum Computing (QC) by utilizing DL methods for feature extraction and employing a Quantum Convolutional Neural Network (QCNN) for classification. This approach utilizes Quantum Principles (QP) such as Entanglement and Superposition which enable the model to analyze numerous data patterns simultaneously. The dataset comprises 2,120 facial images of students captured during virtual classes. These images are classified into six emotional states, which are again categorized into two engagement levels: Engaged and Not Engaged. The model employs Z-score normalization along with data augmentation methods to enhance generalization and minimize sensitivity to changes in lighting, orientation and facial expressions. Experimental results show that the Q-FER achieves a classification accuracy of 99.76%, significantly outperforming traditional Convolutional Neural Networks (CNN). These results underscore the promise of Quantum-assisted Deep Learning systems for analyzing real-time engagement in E-learning.