期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2018
卷号：96
期号：4
页码：854
出版社：Journal of Theoretical and Applied
摘要：In the real world, occlusion is considered as an important problem in the domain of emotion recognition. This study investigates the occlusion in real time using a fully automatic system. The features are extracted effectively from special areas of the face using both texture and geometrical features on the face. The system uses a trained backpropagation neural network to find the seven basic emotions. The experiments were conducted on normal, and with occlusions on forehead, eyes, and mouth. The proposed system was able to detect emotions with high recognition rates. The results are presented for near frontal and multi view faces using UPM3DFE and BU3DFE 3D facial expression databases.
