Abstract
The aura matrix of an image indicates how much of each gray level is present in the neighborhood of each other gray level and generalizes the popular texture-analysis tool, the co-occurrence matrix. In this paper we show that interesting structure appears in both the aura and co-occurrence matrices for textures that are synthesized from Gibbs random-field models. We derive this structure by characterizing configurations of the distribution that are most likely to be synthesized when the Gibbs energy is minimized. This minimization is an important part of applications that use the Gibbs model within a Bayesian estimation framework for maximum a posteriori (MAP) estimation. In particular, we show that the aura matrix will become tridiagonal for an attractive autobinomial field when suitable constraints exist on the histogram, neighborhood, and image sizes. Under the same constraints, but where the field is repulsive instead of attractive, the matrix will become antitridiagonal. The interpretation of this structure is especially significant for modeling textures with minimum-energy configurations: zeros in the matrix prohibit certain colors from occurring next to each other, thus prohibiting large classes of textures from being formed.
Similar content being viewed by others
References
J. Besag, “Spatial Interaction and the Statistical Analysis of Lattice Systems (with Discussion), J. Roy. Statist. Soc. Ser. B, vol. 36, 1974, pp. 192–236.
G.R. Cross and A.K. Jain, “Markov Random Field Texture Models,” IEEE Trans. Patt. Anal. Mach. Intell., vol. PAMI 5, 1983, pp. 25–39.
S. Geman and D. Geman, “Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images. IEEE Trans. Patt. Anal. Mach. Intell., vol. PAMI 6, 1984, pp. 721–741.
I.M. Elfadel and R.W. Picard, “Miscibility Matrices Explain the Behavior of Grayscale Textures Generated by Gibbs Random Fields, in Proc. SPIE Conference on Intelligent Robots and Computer Vision, Boston, November 1990, pp. 524–535.
S.W. Zucker and D. Terzopoulos, “Finding Structure in Co-occurrence Matrices for Texture Analysis, Computer Vis. Graph. and Image Process., vol. 12, 1980, pp. 286–308.
G.R. Cross, “Markov Random Field Texture Models,” Ph.D. thesis, Michigan State University, East Lansing, MI, 1980.
L. Garand and J.A. Weinman, “A Structural-Stochastic Model for the Analysis and Synthesis of Cloud Images,” J. Climate Appl. Meteorol., vol. 25, 1986, pp. 1052–1068.
I.M. Elfadel and R.W. Picard, “Auras Part I: Theory and Relation to Co-occurrence and Morphology,” Vision and Modeling Group Tech. Report 149, MIT Media Lab., Massachusetts Institute of Technology, Cambridge, MA, 1991, also submitted for publication.
R.W. Picard, I.M. Elfadel, and A.P. Pentland, “Markov/ Gibbs Texture Modeling: Aura Matrices and Temperature Effects,” in Proc. IEEE Conference on Computer Vision and Pattern, Recognition, Maui, HI, 1991.
G. Strang, Introduction to Applied Mathematics, Cambridge, MA: Wellesley-Cambridge Press, 1986.
I.M. Herstein, Topics in Albebra, New York: John Wiley, 1975.
R.W. Picard and A.P. Pentland, “Markov/Gibbs Image Modeling: Temperature and Texture,” in Proc. SPIE Conference on Intelligent Robots and Computer Vision, Boston, November 1991, pp. 15–26.
H. Derin and H. Elliott, “Modeling and Segmentation of Noisy and Textured Images using Gibbs Random Fields,” IEEE Trans. Patt. Anal. Mach. Intell., vol. PAMI 9, 1987, pp. 39–55.
R.J. Baxter, Exactly Solved Models in Statistical Mechanics, London: Academic Press, 1982.
R. Haralick, “Statistical and Structural Approaches to Texture,” Proc. IEEE, vol. 67, 1979, pp. 786–804.
H. Wechsler, “Texture Analysis—A Survey,” Signal Process, vol. 2, 1980, pp. 271–282.
L.Van Gool, P. Dewaele, and A. Oosterlinck, “Texture Analysis Anno 1983,” Comput. Vis., Graph. Image Process, vol. 29, 1985, pp. 336–357.
Author information
Authors and Affiliations
Additional information
This work was supported by the National Science Foundation and the Defense Advanced Research Projects Agency under grant MIP-88-14612, the National Science Foundation under grant IRI-8719920, and the Rome Air Development Center of the U.S. Air Force Systems Command and the Defense Advanced Research Projects Agency under contract F30602-89-C-0022.
Rights and permissions
About this article
Cite this article
Picard, R.W., Elfadel, I.M. Structure of aura and co-occurrence matrices for the Gibbs texture model. J Math Imaging Vis 2, 5–25 (1992). https://doi.org/10.1007/BF00123878
Issue Date:
DOI: https://doi.org/10.1007/BF00123878