In many domains, including online education, healthcare, security, and human-computer interaction, facial emotion recognition (FER) is essential. Real-world FER is still difficult despite its significance because of some factors such as variable head positions, occlusions, illumination shifts, and demographic diversity. Engagement detection, which is essential for applications like virtual learning and customer services, is frequently challenging due to FER limitations by many current models. In...