Decision Rules for Choice of Neighbors in Random Field Models of Images.
Abstract
Consider the parametric representations of a homogeneous (not necessarily isotropic) random field by a set of neighborhood models. Using Bayesian methods and spectral representation of random fields, this paper develops decision rules for choosing a model from a set of r such random field models which differ in their neighborhoods. The theory developed here will find applications in image modeling and texture characterization. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1979
- Accession Number
- ADA078369
Entities
People
- N. Ahuja
- R. Chellappa
- R. L. Kashyap
Organizations
- University of Maryland