Hyper-Laplacian Regularized Multi-View Subspace Clustering With a New Weighted Tensor Nuclear Norm

In this paper, we present a hyper-Laplacian regularized method WHLR-MSC with new weighted tensor nuclear norm for multi-view subspace clustering. Specifically, firstly stack the representation matrices of different views into tensor, which neatly captures higher-order correlations between views. Secondly, in order to make all singular values have contributions based on tensor-Singular Value Decomposition (t-SVD), use constrain constructed can obtain class discrimination information sample distribution more accurately. Third, from geometric point view, data are usually sampled low-dimensional manifold embedded high-dimensional ambient space, model uses graph regularization capture local structure data. An effective algorithm solving optimization problem is proposed. Extensive experiments five benchmark image datasets show effectiveness our proposed method.