Bayesian Sparse Estimation for Background/Foreground Separation
S. Nakajima (Technische Universität Berlin), Masashi Sugiyama (University of Tokyo), S. Derin Babacan (Google Inc.)
[Project Page][Publications]
Chapter Description
In this chapter, we introduce a Bayesian sparse estimation method, called a sparse additive matrix factorization (SAMF). SAMF generalizes probabilistic matrix factorization which allows various types of sparsity design in a unified framework, and contains a Bayesian variant of robust PCA.