IAM   02674
INSTITUTO ARGENTINO DE MATEMATICA ALBERTO CALDERON
Unidad Ejecutora - UE
artículos
Título:
"Model Distribution Dependant Complexity Estimation on Textures". DOI: 10.1007/978-3-642-17277-9_28
Revista:
LECTURE NOTES IN COMPUTER SCIENCE
Editorial:
Springer
Referencias:
Lugar: Heidelberg; Año: 2010 vol. 6455 p. 271 - 271
ISSN:
0302-9743
Resumen:
On this work a method for the complexity of a texturedimage to be estimated is presented. The method allow to detect changeson its stationarity by means of the complexity with respect to a givenmodel set (distribution dependant). That detection is done in such away that also allows to classify textured images according to the wholetexture complexity. When di®erent models are used to model data, themore complex model is expected to ¯t it better because of the higherdegree of freedom. Thus, a naturally-arisen penalization on the modelcomplexity is used in a Bayesian context. Here a nested models schemeis used to improve the robustness and e±ciency on the implementation.Even when MRF models are used for the sake of clarity, the procedureit is not subject to a particular distribution.