A Bayesian framework for Gaussian mixture background modeling

Background subtraction is an essential processing component for many video applications. However, its development has largely been application driven and done in ad hoc manners. In this paper, we provide a Bayesian formulation of background segmentation based on Gaussian mixture models. We show that the problem consists of two density estimation problems, one application independent one dependent, and a set of intuitive and theoretically optimal solutions can be derived. The proposed framework was tested on meeting and traffic videos and compared favorably over well-known algorithms.

• Type: Conference
• Year: 2003
• Where: ICIP
• Authors: Dar-Shyang Lee, Jonathan J. Hull, Berna Erol
• Keywords: Ad Hoc Manners, Background Subtraction, Density Estimation Problems, Essential Processing Component, Gaussian Mixture Models, ICIP 2003, Image Processing