PROJECT TITLE :

Background Subtraction Based on Low-Rank and Structured Sparse Decomposition - 2015

ABSTRACT:

Low rank and sparse representation based ways, that make few specific assumptions concerning the background, have recently attracted wide attention in background modeling. With these strategies, moving objects within the scene are modeled as pixel-wised sparse outliers. But, in many practical situations, the distributions of these moving parts aren't actually pixel-wised sparse but structurally sparse. Meanwhile a sturdy analysis mechanism is needed to handle background regions or foreground movements with varying scales. Based mostly on these two observations, we have a tendency to 1st introduce a class of structured sparsity-inducing norms to model moving objects in videos. In our approach, we regard the observed sequence as being constituted of 2 terms, a coffee-rank matrix (background) and a structured sparse outlier matrix (foreground). Next, in virtue of adaptive parameters for dynamic videos, we propose a saliency measurement to dynamically estimate the support of the foreground. Experiments on challenging well known knowledge sets demonstrate that the proposed approach outperforms the state-of-the-art methods and works effectively on a wide selection of advanced videos.