Computing Multiple Image Motions

The computation of image motion for the purposes of determining egomotion is a challenging task as image motion includes discontinuities and multiple values mostly due to scene geometry, surface translucency and various photometric effects such as surface reflectance. We present algorithms for computing multiple image motions arising from occlusion and translucency which are capable of extracting the information-content of occlusion boundaries and distinguish between those and additive translucency phenomena. Sets of experimental results obtained on synthetic images are presented. These algorithms are based on recent theoretical results on occlusion and translucency in Fourier space. 1 Introduction The importance of motion in image processing cannot be understated: in particular, approximations to image motion may be used to estimate 3D scene properties and motion parameters from a moving visual sensor, to perform motion segmentation, to compute the focus of expansion and time-to-coll..