Efficient and Robust Background Modeling with Dynamic Mode Decomposition

Abstract A large number of modern video background modeling algorithms deal with computational costly minimization problems that often need parameter adjustments. While in most cases spatial and temporal constraints are added artificially to the process, our approach is exploit Dynamic Mode Decomposition (DMD), a spectral decomposition technique naturally extracts spatio-temporal patterns from data. Applied data, DMD can compute models. However, original algorithm for neither efficient nor robust. In this paper, we present an equivalent reformulation leading more suitable into fore- background. Due reformulation, which uses sparse low-dimensional structures, robust derived computes accurate Moreover, show how be extended RGB data periodic parts, streaming enabling versatile use.