Fast motion estimation and motion segmentation using multi-scale approach (MA-P4)

Author(s) : Cedric Demonceaux (CREA/LAMFA, France) Djemaa Kachi-Akkouche (CREA, France)

Abstract : The goal of this work is to find a fast method for motion estimation and motion segmentation. We chose to compute the motion on basis functions. That allows us to compute the velocity in each point of the image by only solving a linear system. The motion segmentation is carried out using a markovian formulation. We minimize the associated energy by a determinist algorithm. But these choices induce two drawbacks: a temporal aliasing problem and a sensitivity of the optimization algorithm with respect to the initialization. To overcome these two problems, we use a multiresolution method that computes successively the motion and the segmentation at each scale by motion compensation. Thanks to this procedure, we obtain a fast method of motion estimation and segmentation which does not require neither initial spacial segmentation nor dominant motion in the sequence.

Menu