Compressive Online Video Background-Foreground Separation Using Multiple Prior Information and Optical Flow

Abstract

In the context of video background-foreground separation, we propose a compressive online Robust Principal Component Analysis (RPCA) with optical flow that separates recursively a sequence of video frames into foreground (sparse) and background (low-rank) components. This separation method can process per video frame from a small set of measurements, in contrast to conventional batch-based RPCA, which processes the full data. The proposed method also leverages multiple prior information by incorporating previously separated background and foreground frames in an n-1 minimization problem. Moreover, optical flow is utilized to estimate motions between the previous foreground frames and then compensate the motions to achieve higher quality prior foregrounds for improving the separation. Our method is tested on several video sequences in different scenarios for online background-foreground separation given compressive measurements. The visual and quantitative results show that the proposed method outperforms other existing methods.

Deep neural network concepts for background subtraction:A systematic review and comparative evaluation

“…Furthermore, a study on the influence of the input feature's type would be interesting. -Rather than working in the pixel domain, DNNs may also be applied in the measurement domain for use in conjunction with compressive sensing data like in RPCA models [44,149].…”

Denoising-based Turbo Message Passing for Compressed Video Background Subtraction

“…In this paper, we consider the compressed video background subtraction problem. Several compressive RPCA methods have been proposed in [21], [22], [36], [37], [23], [38], [39], [25], [41], [24]. In [21], a variant of the PCP method is proposed for noiseless compressed RPCA.…”

LPNN‐based approach for LASSO problem via a sequence of regularized minimizations

“…Compressive sampling (CS) considers models for recovering signals from incomplete or even highly incomplete observation measurements 1–3 . CS arises in various engineering fields including, but not limited to, computer vision, acoustic networking, geotechnical engineering, magnetic resonance imaging 4–9 . CS‐based approaches have provided remarkable and efficient improvements, for example, a myriad of challenging applications have been developed in computer vision field such as autonomous driving systems, video surveillance, object detection, and tracking 7,10–12 .…”

“…[1][2][3] CS arises in various engineering fields including, but not limited to, computer vision, acoustic networking, geotechnical engineering, magnetic resonance imaging. [4][5][6][7][8][9] CS-based approaches have provided remarkable and efficient improvements, for example, a myriad of challenging applications have been developed in computer vision field such as autonomous driving systems, video surveillance, object detection, and tracking. 7,[10][11][12] For this kind of problems and unlike conventional coding, CS is more valuable in reducing the required computational complexity and time consumption.…”

“…[4][5][6][7][8][9] CS-based approaches have provided remarkable and efficient improvements, for example, a myriad of challenging applications have been developed in computer vision field such as autonomous driving systems, video surveillance, object detection, and tracking. 7,[10][11][12] For this kind of problems and unlike conventional coding, CS is more valuable in reducing the required computational complexity and time consumption. 13,14 Basically, CS enables signal reconstruction from far fewer measurements than standard methods.…”