Abstract
We address the problem of adaptive segmentation of images of objects with smooth surfaces. The images are composed of regions of slowly varying intensities that may be corrupted by additive noise. The underlying field is modelled by a Markov random field that consists of both a label process which contains the classification of each pixel in the image and intensity functions which contain the possible grey levels that each pixel may take. The algorithm iteratively repeats two steps; (a) the parameter estimation step, in which the ML estimates of the associated parameters are obtained, and (b) the restoration step, in which the underlying field is estimated through the MAP method. The major contribution of this paper is the idea of allowing the pixel grey values to vary across the image regions. These values are estimated by using windows on the observed data and, as the algorithm progresses, the window size is decreased so that the algorithm adapts to the characteristics of each region.
Original language | English |
---|---|
Title of host publication | IEEE Computer Vision and Pattern Recognition |
Editors | Anon |
Publisher | Publ by IEEE |
Pages | 772-773 |
Number of pages | 2 |
ISBN (Print) | 0818638826 |
Publication status | Published - 1993 |
Event | Proceedings of the 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - New York, NY, USA Duration: 15 Jun 1993 → 18 Jun 1993 |
Other
Other | Proceedings of the 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition |
---|---|
City | New York, NY, USA |
Period | 15/06/93 → 18/06/93 |