Emotive Facial Expression Detection

Facial expressions are important cues to observe human emotions. The thesis studies the involuntary type of facial expressions called micro-expressions. Micro-expressions are quick facial motions, appearing in high stake and stressful situations, typically when a subject tries to hide his or her emotions. Two attributes are present fast duration and low intensity. A simple detection method is proposed, which determines instants of micro-expressions in a video. The method is based on analyzing image intensity differences over a registered face sequence. The specific pattern is detected by an SVM classifier. The results are evaluated on standard micro-expression datasets SMIC and CASMEII. The proposed method outperformed the competing method in detection accuracy. Further, we collected a new real-world micro-expression dataset consisting mostly of poker game videos downloaded from YouTube. We achieved average cross-validation AUC 0.88 for the SMIC, and 0.81 on the new challenging “in the Wild” database called MEVIEW.