Publication details

Two Compound Random Field Texture Models

Conference Paper (international conference)

Haindl Michal, Havlíček Vojtěch

serial: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 21st Iberoamerican Congress, CIARP 2016, p. 44-51 , Eds: Beltran-Castanon C., Nystrom I., Famili F.

action: CIARP 2016 - 21st Iberoamerican Congress 2016, (Lima, PE, 20161108)

project(s): GA14-10911S, GA ČR

keywords: Texture, texture synthesis, compound random field model, CAR model, two-dimensional Bernoulli mixture, two-dimensional Gaussian mixture, bidirectional texture function

preview: Download

abstract (eng):

Two novel models for texture representation using parametric compound random field models are introduced. These models consist of a set of several sub-models each having different characteristics along with an underlying structure model which controls transitions between them. The structure model is a two-dimensional probabilistic mixture model either of the Bernoulli or Gaussian mixture type. Local textures are modeled using the fully multispectral three-dimensional causal auto-regressive models. Both presented compound random field models allow to reproduce, compress, edit, and enlarge a given measured color, multispectral, or bidirectional texture function (BTF) texture so that ideally both measured and synthetic textures are visually indiscernible.