Estimating multiple independent motions in segmented images using parametric models with local deformations

This paper presents a new model for optical flow based on the motion of planar regions plus local deformations. The approach exploits brightness information to organize and constrain the interpretation of the motion by using segmented regions of piecewise smooth brightness to hypothesize planar regions in the scene. Parametric flow models are fitted to these regions an a two step process which first computes a coarse fit and then refines it using a generalization of the standard area-based regression approaches. Since the assumption of planarity is likely to be violated, we allow local deformations from the planar assumption. This parametric+deformation model exploits the strong constraints of parametric approaches while retaining the adaptive nature of regularization approaches.<>