The exploration of multimodal learning enhances our understanding of the educational process. However, it faces challenges, particularly regarding the significance of facial recognition, which is closely linked to its extensive potential in various multimedia applications. To address these challenges, a novel strategy known as the Horned Lizard Segformer Prediction System (HLSPS) has been developed. Initially, a dataset containing student online feedback was collected and utilized for training purposes. Facial recognition is first employed to analyze students' emotions. Subsequently, the feedback is analyzed and presented to educators to assess the effectiveness of learning activities. Finally, performance indicators such as precision, accuracy, recall, F-measure, and error rate are used to evaluate the robustness of HLSPS. The method has been validated against a range of existing models, demonstrating a significant improvement in performance.

Online Learning Emotion Recognition Intelligent Training System Feedback Generation