Abstract:To investigate an effective representation model for 3D facial expression recognition, this paper proposes a multi-feature tensor neighborhood preserving embedding (MFTNPE) method, which seeks various attribute features from raw textured shape scan models of 3D facial expression to construct a multi-feature tensor model. With MFTNPE method, we can project the model into a tensor subspace, explore the interactions and correlations among multimodal data and then extract discriminative low-dimensional tensor features for 3D facial expression recognition. In the study, we first introduce a tensor model based on multi-feature integration for 3D facial expression recognition. Via exploring the textural and geometric features of 3D facial expression data, an original multi-feature tensor model is constructed by stacking various attribute maps orderly. Second, we utilize the tensor completion approach as the scheme of data preprocessing to correct the original multi-feature tensor model. Due to the presence of imperfect data in the projected attribute maps, the tensor completion can predict the missing data in tensor by the trace norm of multi-feature tensor model along each mode. Third, we redefine a weighted matrix for use in the orthogonal-based tensor neighborhood preserving embedding algorithm. In the weighted matrix, the class information is used to guarantee the correct neighborhood relationships for multi-feature tensor samples. This effectively avoids the disturbances and confusions between similar samples while preserving spatial structure and inter-class information. Finally, we conduct extensive experiments on two 3D facial expression recognition datasets. The experimental results show that the proposed method has much better efficiency and performance compared with the state-of-the-art methods.

Fast Nonnegative Tensor Factorization Based on Graph-Preserving for 3d Facial Expression Recognition

Fast Non-Negative Tensor Factorization based on CP decomposition for 3D Facial Expression Recognition

Nonnegative Tensor Factorization Based on Low-Rank Subspace for Facial Expression Recognition

A Sparse Neighborhood Preserving Non-Negative Tensor Factorization Algorithm for Facial Expression Recognition

FERLrTc: 2D+3D Facial Expression Recognition Via Low-Rank Tensor Completion

Graph-Preserving Sparse Nonnegative Matrix Factorization with Application to Facial Expression Recognition

A Discriminant Nonnegative Tensor Factorization Method Based on Sparse Representation Classifier

Facial Expression Recognition Based on Graph-Preserving Sparse Non-Negative Matrix Factorization

Sparse and Low-rank Tucker Decomposition with Its Application to 2D+3D Facial Expression Recognition

Sparse Orthogonal Tucker Decomposition for 2D+3D Facial Expression Recognition

2D+3D facial expression recognition via embedded tensor manifold regularization

Orthogonal Tucker Decomposition Using Factor Priors for 2D+3D Facial Expression Recognition

Facial Expression Recognition Based on Discriminant Neighborhood Preserving Nonnegative Tensor Factorization and ELM

Graph Regularized Discriminative Non-Negative Matrix Factorization for Face Recognition

Multi-Feature Tensor Neighborhood Preserving Embedding for 3D Facial Expression Recognition

Towards Reading Beyond Faces for Sparsity-aware 3D/4D Affect Recognition

FER Based on the Improved Convex Nonnegative Matrix Factorization Feature

Topology Preserving Non-negative Matrix Factorization for Face Recognition

Sparse Multi-Feature Tensor Representation for 3D Facial Expression Recognition

Neighborhood Preserving Non-Negative Tensor Factorization For Image Representation

Non-Negative Matrix Factorization Framework For Face Recognition