by Zoltan Kato, Ting Chuen Pong
Abstract:
We propose a Markov random field (MRF) image segmentation model which aims at combining color and texture features. The theoretical framework relies on Bayesian estimation via combinatorial optimization (Simulated Annealing). The segmentation is obtained by classifying the pixels into different pixel classes. These classes are represented by multi-variate Gaussian distributions. Thus, the only hypothesis about the nature of the features is that an additive Gaussian noise model is suitable to describe the feature distribution belonging to a given class. Here, we use the perceptually uniform CIE-L*u*v* color values as color features and a set of Gabor filters as texture features. Gaussian parameters are either computed using a training data set or estimated from the input image. We also propose a parameter estimation method using the EM algorithm. Experimental results are provided to illustrate the performance of our method on both synthetic and natural color images.
Reference:
Zoltan Kato, Ting Chuen Pong, A Markov Random Field Image Segmentation Model for Color Textured Images, In Image and Vision Computing, volume 24, no. 10, pp. 1103-1114, 2006, Elsevier.
Bibtex Entry:
@string{ivc="Image and Vision Computing"}
@string{elsevier="Elsevier"}
@Article{Kato-Pong2006a,
author = {Kato, Zoltan and Pong, Ting Chuen},
title = {A {M}arkov Random Field Image Segmentation Model for
Color Textured Images},
journal = ivc,
year = 2006,
volume = 24,
number = 10,
pages = {1103--1114},
month = oct,
publisher = elsevier,
pdf = {papers/ivc2001.pdf},
abstract = {We propose a Markov random field (MRF) image
segmentation model which aims at combining color and
texture features. The theoretical framework relies
on Bayesian estimation via combinatorial
optimization (Simulated Annealing). The segmentation
is obtained by classifying the pixels into different
pixel classes. These classes are represented by
multi-variate Gaussian distributions. Thus, the only
hypothesis about the nature of the features is that
an additive Gaussian noise model is suitable to
describe the feature distribution belonging to a
given class. Here, we use the perceptually uniform
CIE-L*u*v* color values as color features and a set
of Gabor filters as texture features. Gaussian
parameters are either computed using a training data
set or estimated from the input image. We also
propose a parameter estimation method using the EM
algorithm. Experimental results are provided to
illustrate the performance of our method on both
synthetic and natural color images.}
}