ANNOUNCEMENT AND PRELIMINARY PROGRAM: INTERNATIONAL WORKSHOP ON OBJECT REPRESENTATION FOR COMPUTER VISION April 13-14 1996 - University of Cambridge, England (Preceeding ECCV'96) Co-Chairs and Organizers: Jean Ponce Martial Hebert Andrew Zisserman University of Illinois Carnegie-Mellon University University of Oxford (ponce@cs.uiuc.edu) (Martial.Hebert@cs.cmu.edu) (az@robots.ox.ac.uk) Program Committee: Brady, M., Oxford University, UK Cipolla, R., University of Cambridge, UK Forsyth, D., U.C. Berkeley, USA Huttenlocher, D., Cornell University, USA Ikeuchi, K., Carnegie Mellon, USA Lowe, D.G., University of British Columbia, Canada Mohr, R., LIFIA-INRIA, France Mundy, J., General Electric, USA Nayar, S., Columbia University, USA Rothwell, C., INRIA, France Shashua, A., Hebrew University of Jerusalem, Israel Sugihara, K., Tokyo University, Japan Taylor, C.J., University of Manchester, UK Van Gool, L., Catholic University of Leuven, Belgium Yuille, A., Harvard University, USA Zerroug, M., Adept Technologies, USA An international workshop on object representation for computer vision will be held at the University of Cambridge, England, on April 13-14, preceeding ECCV'96 which begins on April 15. For registration information, please consult the ECCV Web site: http://svr-www.eng.cam.ac.uk/Research/Vision/ECCV/ECCV.html and look up the Early Registration info. (Note: early registration is up to Feb. 29. Registration fee goes up afterwards.) BACKGROUND: In December of 1994, a workshop co-sponsored by NSF and ARPA was held in New York City. Its goal was to evaluate current approaches to object representation and to identify important issues and promising research directions. The 1996 Cambridge workshop is a follow-up to the 1994 New York workshop. Its goals are: o to present a state of the art of the research on object representation for object recognition; o to assess the progress achieved in key areas identified during the first workshop, e.g., part decomposition and quasi invariants; and, o to explore the representational issues involved in applications that go beyond traditional object recognition, e.g., image databases, manufacturing, medical imaging, or virtual reality. FORMAT: The workshop will consist of three half-day sessions, each of them including an invited lecture, presentations of accepted papers, and a panel. The invited speakers are: o Takeo Kanade, Carnegie-Mellon University o Jan Koenderink, Utrecht University o Ram Nevatia, University of Southern California To facilitate discussions, the attendance to the workshop will be limited to a hundred participants. The proceedings (including the contributed papers and a summary of the panel discussions) will be published after the workshop by Springer-Verlag in the Lecture Notes in Computer Science series. The workshop will be held just prior to ECCV'96 at the University of Cambridge, and accommodation will be at Robinson College. PRELIMINARY PROGRAM: SATURDAY APRIL 13 -- AFTERNOON o Welcome and introduction o Invited talk: Ram Nevatia, University of Southern California Break o Geometric and topological representations I 1. Sayd, Dhome & Lavest: Using deformable model to recover generalized cylinders. 2. Carlsson: Combinatorial geometry for shape representation and indexing. 3. Rothwell, Mundy & Hoffman: Representing objects using topology. Break o Geometric and topological representations II 1. Angelopoulou, Williams & Wolff: Curvature based signatures for object description and recognition. 2. Forsyth, Malik, Fleck, Leung, Bregler, Carson & Greenspan: Finding objects by grouping. Break o Panel 1. SUNDAY APRIL 14 -- MORNING o Invited Talk: Jan Koenderink, Utrecht University o Appearance-based representations I 1. Nayar & Murase: Dimensionality of illumination manifolds in eigenspace. 2. Belhumeur, Yuille & Epstein: Learning and recognising objects using illumination subspaces. 3. Pope & Lowe: Learning appearance models for object recognition. Break o Appearance-based representations II - hybrid approaches 1. Schmid, Bobet, Lamiroy & Mohr: An image-oriented CAD approach. 2. Liu, Mundy, Zisserman, Pillow & Rothwell: An experimental comparison of appearance models and geometric class models represented by invariants. Break o Panel 2. SUNDAY APRIL 14 -- AFTERNOON o Invited talk: Takeo Kanade, Carnegie-Mellon University Break o 3D representations and applications I 1. Francois & Medioni: Generic shape learning and recognition. 2. Ayoung-Chee, Dudek & Ferrie: Enhanced 3D representation using multiple models. 3. Shum, Hebert & Ikeuchi: On 3D shape synthesis. Break o 3D representations and applications II 1. Wright, Fitzgibbon, Giblin & Fisher: Beyond the Hough transform: further properties of the R, theta mapping and their applications. 2. Kutulakos & Vallino: Non-Euclidean object representation for calibration-free video overlay. Break o Panel 3.