Abstract. Traditional facial expression recognition methods assume that facial expression in the training and testing sets are collected under the same condition such that they are independent and identically distributed. However, the assumption is not satisfied in many real applications. This problem is referred to as cross-dataset facial expression recognition. On the other hand, the traditional facial expression recognition methods are based on basic emotion theory proposed by Ekman. Unfortunately, the theory is limited to express diverse and subtle emotion. To solve the problem of the cross-dataset facial expression recognition and enrich the emotion expression, a transfer learning algorithm TPCA and arousal-valence emotional model are adopted in this paper. A new facial emotion recognition method based on TPCA and two-level fusion is proposed, which combine weight fusion and correlation fusion between arousal and valence to improve the recognition performance under cross-dataset scenarios. The contrast experimental results show that the proposed method can get better recognition result than the traditional methods.

There are some problems when the discriminative features are used in the traditional kinship verification methods, such as focusing on the local region information, containing a lot of noisy in non-face regions and redundant information in overlapping regions, manual parameters setting and high dimension. To solve the above problems, a novel kinship verification method based on deep transfer learning and feature nonlinear mapping is proposed in this paper. Firstly, a new deep learning model trained on the face recognition dataset is transferred to the kinship datasets to extract high-level feature. Secondly, siamese multi-layer perceptrons and triangular similarity metric learning are combined to reduce the dimensionality of feature vector by nonlinear mapping. Meanwhile it would guarantee a smaller distance between kin pairs while a larger distance between non-kin pairs. Lastly, the cosine similarity of feature vector pairs is computed, and traditional classifier, such as SVM, is used. Experiments on the TSKinFace, KinFace W-I and KinFace W-II datasets indicate the proposed method could achieve better performance than the traditional methods.