Hide menu

Two-Dimensional Channel Representation for Multiple Velocities

Hagen Spies, Per-Erik Forssén
SCIA 2003, Gothenburg, Sweden
Proceedings of the 13th Scandinavian Conference on Image Analysis
Pages 356-362
June-July 2003


We present a two-dimensional information representation, where small but overlapping Gaussian kernels are used to encode the data in a matrix. Apart from points we apply this to constraints that restrict the solution to a linear subspace. A localised decoding scheme accurately extracts multiple solutions together with an estimate of the covariances. We employ the method in optical flow computations to determine multiple velocities occurring at motion discontinuities.

Full Paper

On-line proceedings available at the Springer website.

Bibtex entry

  author = 	 {Hagen Spies and Per-Erik Forss\'en},
  title = 	 {Two-Dimensional Channel Representation for Multiple
  booktitle = 	 {Proceedings of the 13th Scandinavian Conference on
Image Analysis},
  pages = 	 {356--362},
  year = 	 {2003},
  series = 	 {{LNCS} 2749},
  isbn =         {978-3-540-40601-3},
  address = 	 {Gothenburg, Sweden},
  month = 	 {June-July}

Per-Erik Forssén

Per-Erik Forssén


Computer Vision Laboratory
Department of Electrical Engineering
Building B
Room 2D:521
SE-581 83 Linköping, Sweden
+46(0)13 285654

< >

My pages:

Page responsible: Per-Erik Forss&eacute;n
Last updated: 2023-09-06