Effective Gaussian mixture learning for video background subtraction

TLDR

Adaptive Gaussian mixtures have been used for modeling nonstationary temporal distributions of pixels in video surveillance applications, but balancing convergence speed and stability remains a challenge. This paper proposes an effective scheme to improve the convergence rate without compromising model stability. The scheme replaces the global, static retention factor with an adaptive learning rate computed for each Gaussian at every frame. The adaptive learning rate yields significant improvements on synthetic and real video data, leading to better background subtraction segmentation than a standard method.

Abstract

Adaptive Gaussian mixtures have been used for modeling nonstationary temporal distributions of pixels in video surveillance applications. However, a common problem for this approach is balancing between model convergence speed and stability. This paper proposes an effective scheme to improve the convergence rate without compromising model stability. This is achieved by replacing the global, static retention factor with an adaptive learning rate calculated for each Gaussian at every frame. Significant improvements are shown on both synthetic and real video data. Incorporating this algorithm into a statistical framework for background subtraction leads to an improved segmentation performance compared to a standard method.