Title :
Markov image modeling
Author_Institution :
Rensselaer Polytechnic Institute, Troy, NY, USA
fDate :
10/1/1978 12:00:00 AM
Abstract :
The theory of two-dimensional spectral factorization is reviewed in the context of recursive modeling. The role of the Markov random field in recursive image modeling is then presented. Since spectral factorization in two or higher dimensions generally results in infinite-order factors, it is necessary to perform Markov modeling after spectral factorization. The above concepts are then applied to the problem of Kalman filtering of images.
Keywords :
Image processing; Kalman filtering; Markov processes; Recursive estimation; Spectral factorizations; Context modeling; Data compression; Filtering; Image processing; Kalman filters; Markov random fields; Maximum likelihood detection; Nonlinear filters; Signal processing; Two dimensional displays;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.1978.1101866
