Automatic Reconstruction of Silhouettes Using B-Splines

  • Sonja Glas
  • Gabriel Recatalá
  • Michael Sorg
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2749)


A complete method for automated 2D reconstruction of object silhouettes by B-splines is proposed. The first step is the automated initialisation of the B-spline. A new method is developed that uses the shape information of the object’s outline. The second and third steps are the methods for controlling (corresponding) and fitting (improving) the initial B-spline. These three steps complete the automated 2D representation. To show the quality of the automatically generated final B-spline a benchmark is presented where the spline approximation is compared to the ideal case in which the outline of the object is a B-spline. It will be shown that the presented method produces a highly accurate B-spline approximation of the silhouette.


  1. 1.
    R. Bartels, J. Beatty and B. Barsky: An Introduction to Splines for use in Computer Graphics and Geometric Modeling. Morgan Kaufmann, California, 1987zbMATHGoogle Scholar
  2. 2.
    T. Cham, R. Cipolla: Automated B-spline Curve Representation Incorporating MDL and Error-Minimizing Control Point Intersection Strategies. IEEE Trans. Pattern Analysis and Machine Intelligence, January 1999, Vol. 21, Nr. 1, 49–53CrossRefGoogle Scholar
  3. 3.
    A. Galton and R. Meathrel: Qualitative Outline Theory. Proc.16th Int. Joint Conference on Artificial Intelligence, Thomas Dean, 1999, Stockholm Sweden, 1061–66Google Scholar
  4. 4.
    S. Glas: Automatic Initialisation of B-Splines and Fitting Them to 2D Object Silhouettes. Master Thesis, Institute for Real-Time Computer Systems, TU Munich, 2002Google Scholar
  5. 5.
    A. Blake and M. Isard: Active Contours. Springer, London, 1998, 115–134, 289–293Google Scholar
  6. 6.
    T. Melen, T. Ozanian: A fast Algorithm for Dominant Point Detection on Chain-Coded Contours. Computer Analysis of Images and Patterns, Springer, 1993, 245–253Google Scholar
  7. 7.
    P. R. Mendonça: Multiview Geometry: Profiles and Self-Calibration. PhD Thesis, Jesus College University of Cambridge, 2001Google Scholar
  8. 8.
    I. Pavlidis, N. P. Papanikolopoulos: Automatic Selection of Control Points for Deformable-Model-Based Target Tracking. Proc. Intl. Conf. Robotics and Automation, Minneapolis, Minnesota, April 1996, 2915–2920Google Scholar
  9. 9.
    G. Recatalá, M. Sorg, J. Leupold, P.J. Sanz, A.P. del Pobil: Visual grasp determination and tracking in 2D dynamic scenarios. In Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems, Lausanne, Switzerland, 2002, 25–30Google Scholar
  10. 10.
    J. Sporring and R. Arps: Representing Contours as Sequence of One Dimensional Functions. In Proceedings of the Fourth Asian Conference on Computer Vision (ACCV’00), Taipei, Taiwan, January 2000, 742–746.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Sonja Glas
    • 1
  • Gabriel Recatalá
    • 2
  • Michael Sorg
    • 1
  1. 1.Institute for Real-Time Computer SystemsTechnische Universität MünchenMunichGermany
  2. 2.Robotic Intelligence Lab.Universitat Jaume ICastellónSpain

Personalised recommendations