Quaternion-based dynamic mode decomposition for background modeling in color videos

Scene Background Initialization (SBI) is one of the challenging problems in computer vision. Dynamic mode decomposition (DMD) a recently proposed method to robustly decompose video sequence into background model and corresponding foreground part. However, this needs convert color image grayscale for processing, which leads neglect coupling information between three channels image. In study, we propose quaternion-based DMD (Q-DMD), extends by quaternion matrix analysis, so as ultimately preserve inherent structure video. We exploit standard eigenvalues compute its spectral calculate Q-DMD modes eigenvalues. The results on publicly available benchmark datasets prove that our outperforms exact method, experiment also demonstrate performance approach comparable state-of-the-art ones.