R-SpaRCS: An algorithm for foreground-background separation of compressively-sensed surveillance videos

Detection, recognition and tracking are three of the primary tasks involved in surveillance video processing. Given the huge amount of data generated by surveillance systems, it is desirable to use compressed sensing based techniques for acquisition and subsequent processing of videos. For compressively-sensed videos, the task of object detection can be formulated as a matrix decomposition problem, namely, that of decomposing the video volume matrix into a low-rank background and a sparse foreground matrix from a small set of linear measurements corresponding to the video volume matrix. In this paper, we introduce a regularized version of the SpaRCS algorithm, which is a greedy algorithm for solving problems of the above kind. The proposed algorithm, Regularized-SpaRCS (R-SpaRCS), exploits the fact that the foreground component in natural videos exhibits connectedness. R-SpaRCS is a model-based greedy algorithm that takes into account the connectedness of the support of the sparse foreground component in videos. Experiments performed on surveillance video datasets show that R-SpaRCS achieves a given recovery RSNR faster than the SpaRCS algorithm.