Facial expression recognition based on t-SNE and AdaBoostM2

In recent years, because of the great potential applications in various areas such as human-computer interaction (HCI), safety, surveillance, and animation, facial expression recognition (FER) has become a hot topic in computer vision and pattern recognition. Although much progress has been made on FER and many algorithms have been studied such as PCA for FER, FER remains a challenge work due to the difficulties of dimension reduction and classification. In this paper, we give a novel approach for FER, which uses t-Stochastic Neighbor Embedding (t-SNE) for reducing the high-dimensional data into a relatively low-dimensional subspace and then uses AdaBoostM2 as the multi-classifier for the expression classification. The performance evaluation is based on the Japanese Female Facial Expression (JAFFE) database. Experimental results show that the proposed new algorithm applied to FER gains the better performance compared with those traditional algorithms, such as PCA, LDA, LLE and SNE