Title :
Markovian random fields energy minimization algorithms
Author :
Oukil, A. ; Serir, Amina
Author_Institution :
Ecole Mil. Polytech., Bordj-El-Bahri, Algeria
Abstract :
This paper introduces two frameworks that are modified versions of the iterated conditional modes (ICM) algorithm. These methods provide an efficient minimization of Markovian random fields energy function with substantial reduction in computing time. An application to a multi-resolution motion detection scheme is then presented and the results of the two frameworks are discussed and compared to those obtained by ICM
Keywords :
Markov processes; image sequences; minimisation; motion estimation; Markovian random fields; energy minimization; image sequences; iterated conditional modes; motion detection; multiple-resolution analysis; Bayesian methods; Computational modeling; Data mining; Image motion analysis; Image sequences; Laboratories; Minimization methods; Motion detection; Simulated annealing; Stochastic processes;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.903597
