Background Subtraction Based on Nonparametric Bayesian Estimation

Background subtraction, the task of separating foreground pixels from background pixels in a video, is an important step in video processing. Comparing with the parametric background modeling methods, nonparametric methods use a model selection criterion to choose the right number of components for each pixel online. We model the background subtraction problem with the Dirichlet process mixture, which constantly adapts both the parameters and the number of components of the mixture to the scene.