Reconhecimento dinâmico de emoções através de expressões faciais utilizando árvore de decisão

Resumo

The classification emotions through facial expressions in computing environments has been widely investigated because it allows you to enlarge and refine the interaction between user and system. In academia, this classification can be used in virtual learning environments in order to motivate the student to from his emotion. This paper presents a real-time system for the recognition of seven basic emotions (happiness, surprise, anger, fear, disgust, sadness, and neutral). For this, the detection of the facial expressions was made by inserting points on the face based on the motion units. After extraction of features found in the face, it used the technique Decision Tree for the precise classification of emotions. The developed tool has been integrated into Moodle and the recognition of emotions has been tested and validated with static and videos generated from the Cohn-Kanade database images. The results were satisfactory for all emotions, achieved an overall percentage of 86.4% accuracy for recognizing them.