Abstract
Video analysis often begins with background subtraction, which consists of creation of a background model that allows distinguishing foreground pixels. Recent evaluation of background subtraction techniques demonstrated that there are still considerable challenges facing these methods. Processing per-pixel basis from the background is not only time-consuming but also can dramatically affect foreground region detection, if region cohesion and contiguity is not considered in the model. We present a new method in which we regard the image sequence to be made up of the sum of a low-rank background matrix and a dynamic tree-structured sparse matrix, and solve the decomposition using our approximated Robust Principal Component Analysis method extended to handle camera motion. Furthermore, to reduce the curse of dimensionality and scale, we introduce a low-rank background modeling via Column Subset Selection that reduces the order of complexity, decreases computation time, and eliminates the huge storage need for large videos.
Highlights
Background subtraction can be defined as segmentation of a video stream into foreground, which appears at unique moments in time, and the background which is always present
The research in this paper addresses this fundamental task using an approximated Robust Principal Component Analysis (RPCA) based method for background modeling
Given a data matrix containing the frames of a video sequence stacked as its columns, A ∈ Rm×n, RPCA [1] solves the matrix decomposition problem min L ∗ + S 1 s.t
Summary
Background subtraction can be defined as segmentation of a video stream into foreground, which appears at unique moments in time, and the background which is always present. Assuming that we have a long video of a scene at our disposal with hundreds or even thousands of frames, only a handful of these frames determine a model of the background; the rest will either contaminate the background or will be redundant to process To this end, we propose to model the background of the sequence using a low-rank approximation by the output of the CSSP algorithm. In a nutshell contributions of this paper are: inducing structuredsparsity in a novel group structure, namely a dynamic block structure; insensitivity to foreground object size, as a result of using within-patch normalized regularization; assumption of a Gaussian i.i.d. noise for discarding false positive pixels (false alarms); variable rank to accommodate illumination and small scene changes; a dimensionality reduction for RPCA problem via the column subset selection that reduces computational complexity and cost; and an exhaustive evaluation using four datasets that demonstrates top performance in comparison with the state-of-the-art alternatives
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
