Non-negative matrix factorization for visual coding

This paper combines linear sparse coding and non-negative matrix factorization into sparse non-negative matrix factorization. In contrast to non-negative matrix factorization, the new model can learn much sparser representation via imposing sparseness constraints explicitly; in contrast to a close model - non-negative sparse coding, the new model can learn parts-based representation via fully multiplicative updates because of adapting a generalized Kullback-Leibler divergence instead of the conventional mean square error for approximation error. Experiments on MIT-CBCL training faces data demonstrate the effectiveness of the proposed method.