Adaptive fuzzy apporach to background modeling using PSO and KLMS

This paper presents a new adaptive fuzzy approach for background estimation in video sequences of complex scene from the function estimation point of view. A Takagi-Sugeno-Kang (TSK) type fuzzy system is used as the function estimator in the study. The proposed approach uses a hybrid learning method combining both the particle swarm optimization (PSO) and the Kernel Least Mean Square (KLMS) to train the fuzzy estimator. In order to estimate background, we first interpret foreground samples as outliers relative to the background ones and so propose an Outlier Separator (OS). Then, the obtained results of OS algorithm are employed in the fuzzy estimator in order to train and estimate background in each pixel. Experimental results show the high accuracy and effectiveness of the proposed method in background estimation and foreground detection for various scenes.