Background Subtraction Based on Low-Rank and Structured Sparse Decomposition

Abstract

Low rank and sparse representation based methods, which make few specific assumptions about the background, have recently attracted wide attention in background modeling. With these methods, moving objects in the scene are modeled as pixel-wised sparse outliers. However, in many practical scenarios, the distributions of these moving parts are not truly pixel-wised sparse but structurally sparse. Meanwhile a robust analysis mechanism is required to handle background regions or foreground movements with varying scales. Based on these two observations, we first 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 two terms, a low-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 data sets demonstrate that the proposed approach outperforms the state-of-the-art methods and works effectively on a wide range of complex videos.

Dynamic tree-structured sparse RPCA via column subset selection for background modeling and foreground detection

“…[22] and WallFlower [21] results: top row is the original image, second row is the ground truth, the third row is our unrefined results with no post-processing. We used the same frames as [24], [13], [25], [26], [27], and [28], for qualitative comparison. [29] .0903 (10) .2574 (9) .4473 (10) .4344 (10) .3602 (10) .6554 (8) .5713 (7) .3561 (10) .2751 (9) .3830 (10) Stauffer [30] .7570 (5) .6854 (6) .7948 (7) .7580 (8) .6519 (6) .5363 (10) .3335 (10) .3838 (9) .1388 (10) .4842 (9) Culibrk [27] .5256 (7) .4636 (8) .7540 (8) .7368 (9) .6276 (9) .5696 (9) .3923 (9) .4779 (8) .4928 (8) .5600 (8) DECOLOR [7] .3416 (9) .2075 (10) .9022 (5) .8700 (4) .646 (8) .6822 (5) .8169 (3) .6589 (4) .7480 …”

“…Structural information about nonzero patterns of variables have been developed and used in sparse signal recovery, and many approaches have been applied to these problems successfully, such as Lattice Matching Pursuit (LaMP) [8], Dynamic Group Sparsity (DGS) recovery [9], Bayesian Robust Matrix Factorization (BRMF) [10],and the Proximal Operator using Network Flow (ProxFlow) [11]. However, the majority of related methods [12], [13] typically assume that the block structure and its location is known or will suffer in regularization or bootstrapping. In contrast, our method does not assume a prior size or location or structure for sparsity, and dynamically updates these to best fit the natural object shape in the scene, without a separate training phase.…”

Foreground Segmentation via Dynamic Tree-Structured Sparse RPCA

“…i2R results: top row is the original image, second row is the ground truth, and the last row is our unrefined results without post-processing. We used the same frames as [15,16,[22][23][24][25], for qualitative comparison.…”

“…The authors of [14] used a two-pass RPCA framework, in which the first pass generates a saliency map that corresponds to locations of the outliers, and then the second pass uses pre-defined salient blocks in the image, to favor spatially contiguous outliers. In another effort [15] used a group sparse structure, in which overlapping pre-defined groups of pixels in a region of an image are used in conjunction with a maximum norm regularization to take into account the spatial connection of foreground regions. In a recent work [16] a superpixel-based max-norm matrix decomposition approach has been proposed, in which homogeneous static or dynamic regions of image are classified as a graph partitioning problem, via Generalized Fused Lasso.…”

Automatic identification and location technology of glass insulator self-shattering

“…Image segmentation is the premise of image analysis, 15,16 which is key to extract the image features and identify the defects. The maximum interclass variance method (OTSU) is used to segment the insulators; we find that it is difficult for the segmentation by OTSU in RGB space to identify the insulators under the complex background, as shown in Fig.…”