H-012
Facial Emotion Recognition System by Using Depth Sensor

◎Nattawat Chanthaphan・Keiichi Uchimura（Kumamoto Univ.）・Takami Satonaka・Tsuyoshi Makioka（Kumamoto Prefectural College of Tech.）

In this paper, we propose the novel approach to extract the facial feature from moving 3D facial wire-frame. We introduce the facial movement streams feature and Structured Streaming Skeleton (SSS) feature type which is derived from the movement streams. K-Nearest Neighbors (K-NN) and Support Vector Machine (SVM) are used in the experiment with our dataset. SSS feature vector shows a promising accuracy (82.75%), whereas stream feature obtains 72.45%. In spite of lower accuracy in stream feature, it shows an outstanding execution time which is 4.8 times faster. Lastly, our approach can slightly beat the state-of-the-art approach around 1% higher.