Abstract
One of the most powerful uses for Markov random fields is in the area of image analysis, where the (noisy) image is observed on a rectangular lattice. In Bayesian approaches, Markov chain Monte Carlo (McMC) algorithms are usually suggested as a means to obtain a maximum a pos.. teriori (MAP) prediction. The particular problem we consider here is that of contextual image segmentation. In practice, approximations to theoretically optimal McMC algorithms are necessary but these algorithms tend to restrict movement through the space of potential segmentations. In this paper, efficient multi-resolution techniques are used to obtain a good initial labeling and to allow more movement through the label configuration space. Examples of both natural images and synthetic images are presented.
ArticleFootnote
The research was supported by the Office of Naval Research (N00014-93-1-001) and an Iowa State University Research Grant (Carver Grant). The authors are grateful to Michail A. Esterman for providing the glomerulus image.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Y. Amit and U. Grenander (1991). Comparing sweep strategies for stochas-tic relaxation. Journal of Multivariate Analysis, 37:197–222.
J. Besag (1986). On the statistical analysis of dirty pictures. Journal of the oyal Statistical Society, Series B, 48:259–279.
J. Besag and P. Green (1993). Spatial statistics and Bayesian computation. Journal of the Royal Statistical Society, Series B, 55:25–37.
J. Banfield and A. Raftery (1992). Ice floe identification in satellite images using mathematical morphology and clustering about principle curves. Journal of the American Statistical Association, 87:7–16.
N. Cressie (1993). Statistics for Spatial Data, revised edition. Wiley, New York.
N. Cressie and J. Davidson (1997). Image processing, entry in Encyclopedia Statistical Sciences (Update), eds S. Kotz, C.B. Read, and D.L. Banks. Wiley, New York, forthcoming.
R. Cristi (1990). Markov and recursive least squares methods for the estimation of data with discontinuities. IEEE Transactions on Acoustics, Speech, and Signal Processing, 38:1972–1980.
D. Cooper, J. Subrahmonia, Y. Hung, and B. Cernuschi-Frias (1993). The use of Markov random fields in estimating and recognizing objects in 3D space, in Markov Random Fields: Theory and Applications, eds R. Chellappa and A. Jain. Academic Press, New York, 335–367.
H. Derin and H. Elliot (1987). Modeling and segmentation of noisy and textured images using Gibbs random fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 9:39–55.
P.A. Devijver (1988). Image segmentation using causal Markov random fields models. Pattern Recognition, 301:131–143.
H. Elliot, H. Derin, R. Cristi, and D. Geman (1986). Application of the Gibbs distribution to image segmentation, in Statistical Image Processing and Graphics, eds E.J. Wegman and D.J. DePriest. Marcel Dekker, New York, 3–24.
S. Geman and D. Geman (1984). Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6:721–741.
D. Geman, S. Geman, C. Graiigne, and P. Dong (1990). Boundary detection by constrained optimization. IEEE Transactions on Pattern Analysis and Machine Intelligence, 12:609–628.
U. Grenander and M. Miller (1994). Representations of knowledge in complex systems. Journal of the Royal Statistical Society, Series B, 56:549–581.
J. Helterbrand (1996). One-pixel-wide closed boundary identification. IEEE Transactions on Image Processing, 5:780–783.
T. Hong and A. Rosenfeld (1984). Compact region extraction using weighted pixel linking in a pyramid. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6:222–229.
V. Johnson (1994). A model for segmentation and analysis of noisy images. Journal of the American Statistical Association, 89:230–241.
J. Kay (1994). Statistical models for PET and SPECT data.Statistical Methods in Medical Research, 3:5–21.
R. Kindermann and J. Snell (1980). Markov Random Fields and Their Applications, volume 1. American Mathematical Society, Providence, RI.
I. Pitas (1988). Markovian image models for image labeling and edge detection. Signal Processing, 15:365–374.
J. Russ (1992). The Image Processing Handbook. CRC Press, Inc., Boca Raton, FL.
R. Schallkoff (1989). Digital Image Processing and Computer Vision. Wiley, New York.
T. Short (1993). An algorithm for the detection and measurement of rail surface defects. Journal of the American Statistical Association, 88:436–440.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1997 Springer Science+Business Media New York
About this paper
Cite this paper
Helterbrand, J.D., Cressie, N. (1997). Object Identification Using Markov Random Field Segmentation Models at Multiple Resolutions of a Rectangular Lattice. In: Gregoire, T.G., Brillinger, D.R., Diggle, P.J., Russek-Cohen, E., Warren, W.G., Wolfinger, R.D. (eds) Modelling Longitudinal and Spatially Correlated Data. Lecture Notes in Statistics, vol 122. Springer, New York, NY. https://doi.org/10.1007/978-1-4612-0699-6_14
Download citation
DOI: https://doi.org/10.1007/978-1-4612-0699-6_14
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-98216-8
Online ISBN: 978-1-4612-0699-6
eBook Packages: Springer Book Archive