Fitting Random Field Models to Images.
MARYLAND UNIV COLLEGE PARK COMPUTER SCIENCE CENTER
Pagination or Media Count:
This paper deals with fitting two-dimensional stationary random field RF models to images. We assume that the given image is represented on a torus lattice, obeying an R.F. model driven by uncorrelated noise. The stochastic model is characterized by a set of unknown parameters. We describe two sets of experimental results. First, by assigning values to parameters in the stationary range, two-dimensional patterns are generated. It appears that quite a variety of patterns can be generated. Next we consider the problem of estimating the parameters, given an arbitrary image. By assuming a Gaussian structure for the noise, we given an iterative scheme to estimate the unknown parameters. We also implement a decision rule to choose an appropriate set of neighbors for the image. The theory is illustrated by applying it to synthetic patterns. Author
- Human Factors Engineering and Man Machine Systems