Publication details

A Compound MRF Texture Model

Conference Paper (international conference)

Haindl Michal, Havlíček Vojtěch

serial: 20th International Conference on Pattern Recognition, p. 1792-1795

action: 20th International Conference on Pattern Recognition ICPR 2010, (Istanbul, TR, 23.08.2010-26.08.2010)

research: CEZ:AV0Z10750506

project(s): 1M0572, GA MŠk, GA102/08/0593, GA ČR, 2C06019, GA MŠk

keywords: compound Markov random field, bidirectional texture function

preview: Download

abstract (eng):

This paper describes a novel compound Markov random field model capable of realistic modelling of multispectral bidirectional texture function, which is currentl the most advanced representation of visual properties of surface materials. The proposed compound Markov random field model combines a non-parametric control random field with analytically solvable widesense Markov representation for single regions and thus allows to avoid demanding Markov Chain Monte Carlo methods for both parameters estimation and the compound random field synthesis.