Adaptative Multigrid and Variable Parameterization for Optical-flow Estimation
Abstract
We investigate the use of adaptative multigrid minimization algorithms for the estimation of the apparent motion field. The proposed approach provides a coherent and efficient framework for estimating piecewise smooth flow fields for different parameterizations relative to adaptative partitions of the image. The performances of the resulting algorithms are demonstrated in the difficult context of a non convex global energy formulation.