caa a22 4500 467914443 CHVBK 20180406152858.0 cr unu---uuuuu 170328e20060401xx s 000 0 eng 10.1007/s11263-006-3859-2 doi (NATIONALLICENCE)springer-10.1007/s11263-006-3859-2 Segmentation of Planar Objects and Their Shadows in Motion Sequences [Elektronische Daten] [Yaser Yacoob, Larry Davis] We investigate segmentation of planar objects and their cast shadows in image sequences. Given two moving image regions in an image sequence we present an algorithm for determining if the two moving regions can be interpreted as a planar object and its cast shadow. Projective geometry and motion properties are employed to directly recover a homology that constrains point correspondences of the outlines of the image regions and determine if they obey an object/shadow relationship. This homology is derived directly from the motions of the regions and therefore is easier to accomplish than determining point-to-point correspondences between candidate object-shadow pairs. Several experiments under approximate point light source illumination are presented. Springer Science + Business Media, Inc., 2006 motion analysis nationallicence shadow nationallicence planar objects nationallicence projective geometry nationallicence homology nationallicence Yacoob Yaser Computer Vision Laboratory, University of Maryland, 20742, College Park, MD aut Davis Larry Computer Vision Laboratory, University of Maryland, 20742, College Park, MD aut International Journal of Computer Vision Kluwer Academic Publishers 67/1(2006-04-01), 53-69 0920-5691 67:1<53 2006 67 11263 https://doi.org/10.1007/s11263-006-3859-2 text/html Onlinezugriff via DOI 1 research-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/s11263-006-3859-2 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Yacoob Yaser Computer Vision Laboratory, University of Maryland, 20742, College Park, MD aut NATIONALLICENCE 700 1- Davis Larry Computer Vision Laboratory, University of Maryland, 20742, College Park, MD aut NATIONALLICENCE 773 0- International Journal of Computer Vision Kluwer Academic Publishers 67/1(2006-04-01), 53-69 0920-5691 67:1<53 2006 67 11263 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer