Recognizing Cognitive Emotions in E-Learning Environment

In the present work, we describe the development of a Facial Expressions Recognition (FER) system able to recognize cognitive emotions in a distance education context. In this case, many research works show that the recognition of basic emotions is not enough and that recognizing emotions more related to the presence/lack of engagement and flow would be more appropriate. Therefore, we developed a FER system able to classify the following cognitive emotions: enthusiasm, interest, surprise, boredom, perplexity, frustration, and the neutral one. After several experiments, we tested which was the best combination of features and the best algorithm for our classification problem. Results show that the combination of Action Units and gaze and a Multiclass Support Vector Machine achieves the best accuracy on the dataset. Results are encouraging and we plan to integrate the system into an e-learning platform to create a more personalized educational environment.