<p>Emotions are central to education, influencing both instructors’ well-being and students’ learning outcomes. Emotion regulation skills are critical for creating effective teaching environments and teacher-student relationships. Despite increasing interest in the area, a complete synthesis of the effects of emotion regulation on teaching effectiveness is yet to be described. This review discusses the emerging field of facial expression recognition for instructors in video lectures and utilizes machine learning and deep learning technologies in emotional assessment. Literature since 2019 is systematically reviewed, highlighting advances, challenges, and opportunities presented by facial expression recognition within the educational context. Pre-processing of images, feature extraction, classification techniques, and performance evaluation of deep learning models, including both convolutional and recurrent neural networks, are key topics discussed. Additionally, this review article provides a critical analysis of the datasets used for emotion recognition and details regarding the nature and generation of such datasets. The review covers the current research gaps and moves to the future directions for integrating artificial intelligence-driven emotion recognition into professional enhancement programs with a view toward improvement in emotional intelligence and effective teaching. </p>