Paper
23 June 1993 Common image energy map of corner and edge points for active contour models
Eberhard Guelch
Author Affiliations +
Abstract
The idea presented here is to use the features of the Forstner interest operator to address problem areas. The Forstner interest operator can localize point type features (corners and intersections) as well as edge type features. A common frame is at hand that is based on the estimate of the position accuracy and the significance of a point or an edge element (edgel). The derived energy values based on these measures can be geometrically interpreted. Breakpoints (corners) are introduced via image energies. A theoretically well understood method is at hand that combines point and edge energies. In order to increase the area of attraction in one level of a resolution hierarchy an energy pit around each point/edgel is created, based on the interest measures. By placing the center of the area of attraction at the sub-pixel position it is to a certain extent possible to spatially refine the energy map. To avoid the setting of a single specific window size for the interest operator a sequence of windows is applied to the image. Image energies are derived that are summed up, weighted by the window size. This final energy map is used as an input for the active contour models. This report gives the necessary framework and describes the principle ideas. Some alternatives for the design of the energy map and ranges for the few parameters required are discussed. Examples on some real images show the potential of this approach which is going to be used for the application of snakes for automatic and semi-automatic cartographic feature extraction in aerial images.
© (1993) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Eberhard Guelch "Common image energy map of corner and edge points for active contour models", Proc. SPIE 2031, Geometric Methods in Computer Vision II, (23 June 1993); https://doi.org/10.1117/12.146632
Lens.org Logo
CITATIONS
Cited by 1 scholarly publication.
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Chemical elements

Computer vision technology

Machine vision

Image segmentation

Feature extraction

Image resolution

Visual process modeling

RELATED CONTENT


Back to Top