Section: New Results

Adaptive global and local motion estimation

Participants : Noémie Debroux, Charles Kervrann.

The design of data costs is one of the main research issue for variational optical flow estimation. The aim is to improve discriminative power by integrating appropriate neighborhood information, while preserving computational efficiency. Most previous works define features on patches with predefined sizes and shapes, or filter pixelwise costs with fixed filtering parameters. We proposed a novel approach estimating spatially varying parameters of filters used to define the data term [8] . More specifically, our model considers Gaussian filtering of the pixelwise brightness constancy equation and imposes smoothness constraints on motion and convolution filter size (bandwith). The energy encoding these assumptions is alternatively minimized over flow field and the spatially varying bandwidth in a variational framework. Experimental results on the Middlebury database demonstrated clear improvements yielded by our method over the spatially constant case of [32] (see Fig. 11 ).

Collaborator: Denis Fortun (UMR 144 CNRS-Institut Curie, STED team, Paris)

                                               (EPFL, Lausanne, Switzerland)

Figure 11. Comparison on a sequence of the Middlebury benchmark. Top from left to righ: input image and spatially filter bandwidth estimation. Bottom from left to right: velocity field computed by [32] (endpoint error = 0.143) and by our method (endpoint error = 0.126).