Motion from occlusions

In computer vision, occlusions are almost always seen as undesirable singul arities that pose difficult challenges to image motion analysis problems, s uch as optic flow computation, motion segmentation, disparity estimation, o r egomotion estimation. However, it is well known that occlusions are extre mely powerful cues for depth or motion perception, and could be used to imp rove those methods. In this paper, we propose to recover camera motion information based unique ly on occlusions, by observing two specially useful properties: occlusions are independent of the camera rotation, and reveal direct information about the camera translation. We assume a monocular observer, undergoing general rotational and translati onal motion in a static environment. We present a formal model for occlusio n points and develop a method suitable for occlusion detection. Through the classification and analysis of the detected occlusion points, we show how to retrieve information about the camera translation (FOE). Experiments wit h real images are presented and discussed in the paper. (C) 2001 Elsevier S cience B.V. All rights reserved.