Background modeling is a critical component for various vision-basedapplications. Most traditional methods tend to be inefficient when solvinglarge-scale problems. In this paper, we introduce sparse representation intothe task of large scale stable background modeling, and reduce the video sizeby exploring its 'discriminative' frames. A cyclic iteration process is thenproposed to extract the background from the discriminative frame set. The twoparts combine to form our Sparse Outlier Iterative Removal (SOIR) algorithm.The algorithm operates in tensor space to obey the natural data structure ofvideos. Experimental results show that a few discriminative frames determinethe performance of the background extraction. Further, SOIR can achieve highaccuracy and high speed simultaneously when dealing with real video sequences.Thus, SOIR has an advantage in solving large-scale tasks.
展开▼
