Title :
A novel pseudo-likelihood equation for Potts MRF model parameter estimation in image analysis
Author :
Levada, Alexandre L M ; Mascarenhas, Nelson D A ; Tannùs, Alberto
Author_Institution :
Inst. de Fis. de Sao Carlos, Univ. de Sao Paulo, Sao Carlos
Abstract :
This paper presents a novel pseudo-likelihood equation for the estimation of the Potts MRF model parameter on second-order neighborhood systems, allowing the modeling of less restrictive contextual systems in a large number of MRF applications in a computationally feasible way. We propose a hypothesis testing approach to validate the obtained results. The test statistic together with the p-values, calculated through our approximation for the asymptotic variance of the derived maximum pseudo-likelihood estimator, provide a complete framework for quantitative analysis in MRF parameter estimation.
Keywords :
Markov processes; image processing; maximum likelihood estimation; Markov random fields; Potts MRF model; hypothesis testing; image analysis; maximum pseudo-likelihood estimator; parameter estimation; pseudo-likelihood equation; Analysis of variance; Computer applications; Context modeling; Density functional theory; Equations; Image analysis; Maximum likelihood estimation; Parameter estimation; Statistical analysis; Testing; Markov Random Fields; Potts model; image analysis; maximum pseudo-likelihood estimation;
Conference_Titel :
Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4244-1765-0
Electronic_ISBN :
1522-4880
DOI :
10.1109/ICIP.2008.4712133