Segmentation of Multisensor Images using Evidential Combination in a Markovian Environment
Segmentation d'Images Multisenseur par Fusion Évidentielle dans un Contexte Markovien
OPEN ACCESS
The Dempster-Shafer combination rule turns out to be quite efficient in the segmentation of multisensor images in numerous situations. On the other hand, in a Bayesian framework the Hidden Markov Random Fields have been of interest for some twenty years. The aim of our work is to propose some methods capable of merging both evidential and Markovian field advantages. The interest of the methods proposed and the differences in their behaviour are studied through simulations on synthetic images.
Résumé
The Dempster-Shafer combination rule turns out to be quite efficient in the segmentation of multisensor images in numerous situations. On the other hand, in a Bayesian framework the Hidden Markov Random Fields have been of interest for some twenty years. The aim of our work is to propose some methods capable of merging both evidential and Markovian field advantages. The interest of the methods proposed and the differences in their behaviour are studied through simulations on synthetic images.
Theory of evidence, Hidden Markov fields, Dempster-Shafer combination rule, Image segmentation.
Mots clés
Théoriedel'évidence, champs de Markov cachés,fusion de Dempster-Shafer,segmentationd'images.
[1] A. Appriou: Probabilitéset incertitude en fusion dedonnéesmufti-senseurs, Revue Scientifiqueet Technique de la Defence,n° 11, pp. 27-40, 1991.
[2] J.A. Barnett :Computational Methods for a mathematical theory of evidence, proc. 7th Int. joint conf. Artificial Intell., Vancouver, BC, pp. 868-875, 1985.
[3] J. Besag : On the statistical analysis of dirty pictures, Journal of the Royal Statistical Society,Series B,vol. 48, pp. 259-302, 1986.
[4] 1 .Bloch : Some aspect of Dempster-Shafer evidence theory for classification of multi-modality medical images taking partial volume effect into account, Pattern Recognition Letters, n° 17, pp. 905-919, 1996.
[5] I. Bloch,H.Maître :Fusion dedonnées entraitement desimages : Modèles d'information et décision, TraitementDu Signal, vol. 11, n°6, pp. 435-446, 1994.
[6] M. Broniatowski, G. Celeux, J. Diebolt : Reconnaissance de mélanges de densitésparun algorithme d'apprentissage probabiliste.In Data Analysis and Informatics 3, E. Diday (Ed.) North-Holland, Amsterdam, 1983.
[7] G. Casella, E.I. George : Explaining the Gibbs Sampler, The American Statistician, vol. 46, n°3, 1992.
[8] J. Van Cleynenbreugel, S. A. Osinga, F. Fierens, P. Suetens, A. Oosterlinck Road Extraction from Multi-temporel Satellite Images by an Evidentiel Reasoning Approach,Pattern Recognition Letters n° 12, pp. 371-380, 1991.
[9] Y. Delignon, A. Marzouki, W.. Pieczynski:Estimation of Generalized Mixture and Its Application in Image Segmentation, IEEE Transactions on Image Processing, vol. 6, n° 10, 1997.
[10] A.P. Dempster : Upper and Lower probabilities induced by a multivalued mapping,Annals of Mathematical statistics, AMS-38, pp. 325-339, 1967.
[11] D. Dubois, H. Prade: Théorie des possibilités,application à lareprésentation des connaissancesenInformatique, Masson, Paris, 1988.
[12] R. Duda and P. Hart : Pattern Classification and Scene Analysis, New-York, Willey, 1973.
[13] S. Geman, G. Geman : Stochastic relaxation, Gibbs distribution and the bayesian restoration of images, IEEE Transactions on PAM], vol . 6, pp. 721741, 1984.
[14] N. Giordana, Pieczynski : Estimation of Generalized Multisensor Hidden Markov Chains and Unsupervised Image Segmentation, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 19, n°5, pp. 465-476,1997.
[15] J. Guan, D.A. Bell : Evidence Theory and its applications, Amsterdam, Nort-Holland, 1991.
[16] G.J. Klir, T.A. Folger :Fuzzy sets, Uncertainty and Information.Englewood Cliffs, NJ: Prentice Hall, 1988.
[17] J. Marroquin, S. Mitter, T. Poggio:Probabilistic solution of ill-posed problems in computational vision, Journal of the American Statistical Association, 82, pp. 76-89, 1987.
[18] J. Pearl:Probabilistic reasoning in intelligent system : networks of plausible inference. MorganKaufmannpub., San Mateo, ca, USA, 1988.
[19] W. Pieczynski : Champs de Markov cachés et estimation conditionnelle itérative,Traitementdu Signal, vol.11, n°2, pp. 141-153, 1994.
[20] G. Shafer :A Mathematical theory of Evidence,Princeton University Press, Princeton, New Jersey, 1976.
[21] G. Shafer : Constructive probability, Synthese,vol. 48, pp. 1-60, 1981.
[22] G. Shafer : Perspective on the theory and practice of belief function, International Journal of Approximate Reasoning, vol.4, pp. 323-362, 1990.
[23] P. Smets : Un modèle mathematico-statistique simulant le processus du diagnostic medical, Doctoral Dissertation at the free Unervsity of Brussels, PressesUniversitairesdeBruxelles, 1978.
[24] P. Smets : Belief functions versus probability functions. In B. Bouchon,L. Saitta and R.R. Yager (eds), Uncertainty and intelligent systems, pp. 17-24. Berlin: Springer-Verlag, 1988.
[25] F. Voorbaark : On the justification of Dempster's rule of combination, Artificial Intelligence, vol. 48, pp. 171-197, 1991.
[26] R.R. Yager : A general approach to decision making with evidential knowledge, Uncertainty in Artificial Intelligence, L.N. Kanalet J.F. Lemmer ed., Elsvier Science Publishers, BV North-Holland, 1986.
[27] P. Walley: Statistical reasoning with imprecise probabilities, Chapmann and Hall,London, 1991.
[28] L. A. Zadeh :Fuzzy sets, Inform. Conte, vol. 8, pp. 338-353, 1965.
[29] L. A. Zadeh: A simple view of the Dempster-Shafer theory of evidence, AI Magazine, Summer, pp. 85-90, 1986.
[30] L. Zhang : Representation, Independence, and Combination of evidence in the Dempster-Shafer theory, Advances in the Dempster-Shafer Theory of Evidence,R. R. Yager, M. Fedrizzi and J. Kacprzyk (eds), Willey, 1994.