scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

A multilevel Gibbs model for texture image analysis

26 Sep 2005-pp 810-813
TL;DR: A texture image model based on Gibbs distribution is suggested, meant for texture image segmentation and analysis and some results of experiments are presented.
Abstract: A texture image model based on Gibbs distribution is suggested. The model contains three levels - one observable (pixel) level corresponding to the texture image and two unobservable (hidden) levels. One hidden level represents a map of texture (i.e. a field of texture domain labels) and the other consists of contour elements and conforms to line process. The model is meant for texture image segmentation and analysis. Some results of experiments are presented.
Citations
More filters
Proceedings ArticleDOI
Xiaoting Shi1, Hai Huang1, Bo Wang1, Shuo Pang1, Hongde Qin1 
01 Jul 2019
TL;DR: An automatic Identifying cage boundary technology based on a Gray Level Co-occurrence Matrix by Support Vector Machine Classifier, which shows classification accuracy rate was enough high if there are enough training samples for building training model.
Abstract: Underwater vehicle plays an indispensable role in marine observation and biological fishing operations for cage culture. If underwater vehicle could identify the cage boundary autonomously, it can plan the route ahead of time to avoid colliding with the cage, which greatly improves the efficiency of the vehicle and ensure its safety. This paper proposes an automatic Identifying cage boundary technology based on a Gray Level Co-occurrence Matrix by Support Vector Machine Classifier. Based on rich textural features, Gray Level Co-occurrence Matrix (GLCM) was extracted from the cage image and then calculate GLCM features including energy, contrast, entropy, inverse different moment, correlation, and Homogeneity. Support Vector Machine (SVM) classifier is trained with these features and then the classified results were obtained for the query images. The experiments show classification accuracy rate was enough high if there are enough training samples for building training model.

5 citations


Cites methods from "A multilevel Gibbs model for textur..."

  • ...Since the 1980s, the Gibbs model [8], the Gauss Markov random field (GMRF) model [9], and the Markov Random Field model (HMRF) [10] have emerged....

    [...]

References
More filters
Journal ArticleDOI
TL;DR: The analogy between images and statistical mechanics systems is made and the analogous operation under the posterior distribution yields the maximum a posteriori (MAP) estimate of the image given the degraded observations, creating a highly parallel ``relaxation'' algorithm for MAP estimation.
Abstract: We make an analogy between images and statistical mechanics systems. Pixel gray levels and the presence and orientation of edges are viewed as states of atoms or molecules in a lattice-like physical system. The assignment of an energy function in the physical system determines its Gibbs distribution. Because of the Gibbs distribution, Markov random field (MRF) equivalence, this assignment also determines an MRF image model. The energy function is a more convenient and natural mechanism for embodying picture attributes than are the local characteristics of the MRF. For a range of degradation mechanisms, including blurring, nonlinear deformations, and multiplicative or additive noise, the posterior distribution is an MRF with a structure akin to the image model. By the analogy, the posterior distribution defines another (imaginary) physical system. Gradual temperature reduction in the physical system isolates low energy states (``annealing''), or what is the same thing, the most probable states under the Gibbs distribution. The analogous operation under the posterior distribution yields the maximum a posteriori (MAP) estimate of the image given the degraded observations. The result is a highly parallel ``relaxation'' algorithm for MAP estimation. We establish convergence properties of the algorithm and we experiment with some simple pictures, for which good restorations are obtained at low signal-to-noise ratios.

18,761 citations

Book
01 Feb 1995
TL;DR: In this paper, the mathematical foundations of Bayesian image analysis and its algorithms are discussed, and the necessary background from imaging is sketched and illustrated by a number of concrete applications like restoration, texture segmentation and motion analysis.
Abstract: The book is mainly concerned with the mathematical foundations of Bayesian image analysis and its algorithms. This amounts to the study of Markov random fields and dynamic Monte Carlo algorithms like sampling, simulated annealing and stochastic gradient algorithms. The approach is introductory and elementary: given basic concepts from linear algebra and real analysis it is self-contained. No previous knowledge from image analysis is required. Knowledge of elementary probability theory and statistics is certainly beneficial but not absolutely necessary. The necessary background from imaging is sketched and illustrated by a number of concrete applications like restoration, texture segmentation and motion analysis.

614 citations

BookDOI
01 Jan 1995
TL;DR: The book is mainly concerned with the mathematical foundations of Bayesian image analysis and its algorithms, which amounts to the study of Markov random fields and dynamic Monte Carlo algorithms like sampling, simulated annealing and stochastic gradient algorithms.

486 citations

01 Jan 1995
TL;DR: A new algorithm for segmentation of textured images using a multiresolution Bayesian approach and proposes new approaches for the extension of binary and grayscale morphological operations to color imagery.
Abstract: Comer, Mary L. Ph.D., Purdue University, December 1995. Multiresolution Image Processing Techniques with Applications in Texture Segmentation and Nonlinear Filtering. Major Professor: Edward J. Delp. We present a new algorithm for segmentation of textured images using a multiresolution Bayesian approach. The algorithm uses a multiresolution Gaussian autoregressive (MGAR) model for the pyramid representation of the observed image, and assumes a multiscale Markov random eld model for the class label pyramid. Unlike other approaches, which have either used a single-resolution representation of the observed image or implicitly assumed independence between di erent levels of a multiresolution representation of the observed image, the models used in this thesis incorporate correlations between di erent levels of both the observed image pyramid and the class label pyramid. The criterion used for segmentation is the minimization of the expected value of the number of misclassi ed nodes in the multiresolution lattice. The estimate which satis es this criterion is referred to as the \multiresolution maximization of the posterior marginals" (MMPM) estimate, and is a natural extension of the single-resolution maximization of the posterior marginals (MPM) estimate. The parameters of the MGAR model | the means, prediction coe cients, and prediction error variances of the di erent textures | are unknown. The expectation-maximization (EM) algorithm is used to estimate these parameters while simultaneously performing the segmentation. Analysis and experimental results demonstrating the performance of the algorithm are presented. We also propose new approaches for the extension of binary and grayscale morphological operations to color imagery. We investigate two approaches for \color

9 citations

Proceedings ArticleDOI
01 Oct 2006
TL;DR: This paper is concerned with Gibbs fields taking on values from finite sets, which allows to overcome difficulties in estimating Gibbs distribution parameters and to synthesize some useful algorithms of image processing.
Abstract: Gibbs (Markov) random fields are used as stochastic picture models in image processing because of their conceptual simplicity and due to the fact that Gibbs models are fit to synthesize algorithms based on Bayes approach. In this paper, we are concerned with Gibbs fields taking on values from finite sets. This restriction allows to overcome difficulties in estimating Gibbs distribution parameters and to synthesize some useful algorithms of image processing.

3 citations