Compressive online robust principle component analysis with multiple prior information
Host Publication: IEEE Global Conference on Signal and Information Processing
Authors: H. Van Luong, N. Deligiannis, J. Seiler, S. Forchhammer and A. Kaup
Publication Date: Nov. 2017
Number of Pages: 5
Online Robust Principle Component Analysis (RPCA) arises naturallyin time-varying signal decomposition problems such as videoforeground-background separation. We propose a compressive onlineRPCA algorithm that decomposes recursively a sequence of datavectors (e.g., frames) into sparse and low-rank components. Unlikeconventional batch RPCA, which processes all the data directly, ourmethod considers a small set of measurements taken per data vector(frame). Moreover, our method incorporates multiple prior informationsignals, namely previous reconstructed frames, to improve theseparation and thereafter, update the prior information for the nextframe. Using experiments on synthetic data, we evaluate the separationperformance of the proposed algorithm. In addition, we applythe proposed algorithm to online video foreground and backgroundseparation from compressive measurements. The results show thatthe proposed method outperforms the existing methods.