Background Modeling by Exploring Multi-Scale Fusion of Texture and Intensity in Complex Scenes

Zhong Zhang,Baihua Xiao,Chunheng Wang,Wen Zhou,Shuang Liu

DOI: https://doi.org/10.1109/acpr.2011.6166554

2011-01-01

Abstract:Background modeling is a fundamental yet challenging issue in video surveillance. Traditional methods usually adopt single feature type to solve the problem, while the performance is usually unsatisfactory when handling complex scenes. In this paper, we propose a multi-scale framework, which combines both texture and intensity feature, to achieve a robust and accurate solution. Our contributions are three folds: first, we provide a multi-scale analysis for the issue; second, for texture feature we propose a novel texture operator named Scale-invariant Center-symmetric Local Ternary Pattern, and a corresponding Pattern Adaptive Kernel Density Estimation technique for its probability estimation; third, we design a Simplified Gaussian Mixture Models for intensity feature. Our method is tested on several complex real world videos with illumination variation, soft shadows and dynamic backgrounds. The experimental results clearly demonstrate that our method is superior to the previous methods.