Document Type : Research Paper

Authors

• Vida Esmaeili ¹
• Mahmood Mohassel Feghhi ¹
• Seyed Omid Shahdi ²

¹ Faculty of Electrical and Computer Engineering, University of Tabriz, Tabriz, Iran

² Faculty of Electrical, Biomedical and Mechatronics Engineering, Qazvin Branch, Islamic Azad University, nokhbegan Blvd., Qazvin, Iran.

Abstract

Micro-expression as the main way of non-verbal communication occurs quickly and subtle in highrisk situations. Since it cannot be misleading, it discloses the real human aim. Nonetheless, feature extraction is an arduous task due to its two particular features. To resolve this problem in this paper, we propose Local Binary Pattern on Four Diagonal Planes. In fact, we utilize it after motion magnification for feature extraction in apex detection task. Also, we combine it and optical flow for micro-expression recognition. Simulation results show that automatic micro-expression apex frame detection and micro-expression recognition is promising using our method on CASME2 and CASME in comparison with other methods.

Keywords

• Apex frame detection
• local binary pattern on four diagonal planes
• micro-expression recognition
• motion magnification
• optical flow