next up previous contents
Next: Curvature Estimation Up: Analysis of Local Orientation Previous: Quadrature Filter Set Methods

Tensor Field Filtering

Researchers in the computer vision laboratory at Linköping University in Sweden have coined the term tensor field filtering [GK95] to describe a broad class of ``reconvolutions'' on the orientation map representation for a given image or, in higher dimensions, a given image sequence. The term ``tensor field'' is used as a generalisation of the concept of an orientation field (or map) for higher dimensions. In this context, OTC could be regarded as a special case of 2D tensor field filtering. Some of the applications of tensor field filtering that have been explored recently include curvature estimation and tensor field controlled image and image sequence enhancement. We will now give a brief overview of the basic ideas behind curvature estimation on the orientation map representation. We will also briefly investigate a set of operators which are designed to detect rotational symmetry in the orientation field.

Markus Weber
Tue Jan 7 15:44:13 PST 1997