Shape and correspondence

https://doi.org/10.1016/S0734-189X(83)80032-2Get rights and content

Previous work in three-dimensional motion interpretation has assumed global rigid motion of each object whose shape is to be determined. It is shown how this assumption can be relaxed for textured objects. Using assumptions about the lighting conditions of the scene, the shape of the object can be recovered using only local information and only local rigidity. The technique developed gives redundant information and can be used to guide the recovery of the correspondence of points on the objects in the scene.

References (14)

  • C.L. Fennema et al.

    Velocity determination in scenes containing several moving objects

    Computer Graphics and Image Processing

    (1979)
  • B.K. P. Horn

    Understanding image intensities

    Artif. Intell.

    (1977)
  • Applied Math Series

  • S. Barnard et al.

    Disparity analysis of images

    IEEE Trans. Pattern Anal. Mach. Intell.

    (1980)
  • L. Dreschler et al.

    Volumetric Model and 3D-Trajectory of a Moving Car Derived from Monocular TV-Frame Sequences of a Street Scene

    7th International Joint Conference on Artificial Intelligence, Vancouver, B.C.

    (August 1981)
  • B.K. P. Horn et al.

    Determining optical flow

    Artif. Intell.

    (1981)
  • T. Kanade et al.

    Mapping Image Properties into Shape Constraints: Skewed Symmetry and Affine-Transformable Patterns

There are more references available in the full text version of this article.

Cited by (22)

  • The Incremental Approximation of Nonrigid Motion

    1994, Computer Vision and Image Understanding
  • Non-rigid motion analysis: Isometric motion

    1992, CVGIP: Image Understanding
View all citing articles on Scopus

This research was supported in part by the Air Force Office of Science Research under Grant AFOSR 77-3190.

Current address: Department of Computer Science, Carnegie-Mellon University, Pittsburgh, Penn. 15213.

View full text