Dictionary Training for Sparse Representation as Generalization of K-Means Clustering