Beihang University Facial Expression Database and Multiple Facial Expression Recognition

Over the last decade, significant effort has been made in developing methods of facial expression analysis. Because of the limited data sets of facial expressions, comprehensive emotional facial expression recognition that is essential in the areas such as affective computing and human-machine interaction has not been well studied. This paper provides the Beihang University facial expression database, which includes 25 facial expressions (18 pure facial expressions, 3 mixed facial expressions and 4 complex facial expressions) and provides data of many unique emotional facial expressions not appeared in other face databases. The questionnaire investigation of videos of 18 pure facial expressions shows the effectiveness of the data. Facial expression recognition experiment on 9 pure facial expressions demonstrates the feasibility of multiple facial expression recognition