R. Paget, "Nonparametric Markov random field models for natural texture images," Ph.D. dissertation, University of Queensland, St Lucia, QLD Australia, Dec. 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

No context found.

R. Paget, "Nonparametric Markov random field models for natural texture images," Ph.D. dissertation, University of Queensland, St Lucia, QLD Australia, Dec. 1999.

No context found.

R. Paget, "Nonparametric Markov random field models for natural texture images," Ph.D. dissertation, University of Queensland, St Lucia, QLD Australia, Dec. 1999.

....theorem[18] giving, log #(x) 9) From Eq. 9) Moussouris [12] gives the following decomposition, #(x) #(x nSC where n SC = 1) C (10) is performed over the sets C # . A similar decomposition can be obtained for the LCPDF of Eq. 2) as shown by Paget [13], giving, s) s) where n (11) III. STRONG MARKOV RANDOM FIELD THEORY The decomposition formula of Eq. 11) is tantalizing in the fact that it gives a factorization of the neighborhood probability into clique probabilities, but unfortunately these clique ....

....marginal distributions are desired, then an iterative proportional fitting technique may be used. Fienberg [6] and Bishop et al. 3] described the iterative proportional fitting technique for a distribution defined in three dimensions. A generalized version of the technique is presented by Paget [13]. A problem with the technique is that it is memory intensive and computationally expensive. V. SYNTHESIS To synthesis a texture we need to sample from the LCPDF P (x s s ) to update a single site of the synthetic texture. This is repeated iteratively over the synthetic texture until the ....

R. Paget. Nonparametric Markov random field models for natural texture images. PhD thesis, University of Queensland, St Lucia, QLD Australia, Dec. 1999.

.... over a homogeneous textured image [7] When the sample data is sparsely dispersed over the multi dimensional histogram domain (as in our case) nonparametric estimates of the LCPDF tend to be more reliable than their parametric counterparts if the underlying true distribution is unknown [8] In [6] we showed that we can estimate the LCPDF as a function of its marginal distributions by assuming that there is conditional independence between nonneighbouring sites for any subset of the image lattice. This is a much stronger assumption than made for a normal MRF which defines a site as being ....

....MeasTex Test Suite [9] required a probability associated with the classification. As the Kruskal Wallis hypothesis test returned a value that was chi squared distributed with one degree of freedom, the probability we returned was the probability of recording a larger chi squared distributed value [6]. 4.1 Comparative assessment of performance In Table 1, a list of summary scores for a suite of nonparametric MRF models are presented. The key to the MRF model names is: n1 refers to a nearest neighbour neighbourhood, n3 refers to a neighbourhood, n5 refers to a 523 neighbourhood. The ....

R. Paget. Nonparametric Markov random field models for natural texture images. PhD thesis, University of Queensland, St Lucia, QLD Australia, December 1999.

....been shown that by restricting the standard constant false alarm rate receiver (CFAR) detectors to homogeneous texture regions, detection performance can be improved [6] One possibility for improving the false detection rates is to use a better texture model. We believe that our texture model [7] will do just that by giving a better statistical understanding of the background texture. In [7, 8] we demonstrated the ability of the texture model to fully characterise a multitude of different textures by using the model to synthesise visually similar texture with regard to a set of training ....

....to homogeneous texture regions, detection performance can be improved [6] One possibility for improving the false detection rates is to use a better texture model. We believe that our texture model [7] will do just that by giving a better statistical understanding of the background texture. In [7, 8], we demonstrated the ability of the texture model to fully characterise a multitude of different textures by using the model to synthesise visually similar texture with regard to a set of training textures, as shown in Fig. 1. For a texture model to detect anomalies in different types of ....

[Article contains additional citation context not shown here]

Rupert Paget, Nonparametric Markov random field models for natural texture images, Ph.D. thesis, University of Queensland, St Lucia, QLD Australia, Dec. 1999.

....hypothesis test [7] it is also possible to attain a goodness of fit measure. As the Kruskal Wallis hypothesis test returns a value that is chisquared distributed with one degree of freedom, the goodnessof fit is given by the probability of recording a larger chisquared distributed value [9]. Fig. 2 shows various probability maps for a texture mosaic and a training texture. 5. PRACTICAL APPLICATION The practical application of terrain mapping a SAR image of Cultana, Fig. 3, shows the two results if: 1) the training class was a patch of trees from the bottom left corner, Fig. 3(b) ....

Rupert Paget, Nonparametric Markov random field models for natural texture images, Ph.D. thesis, University of Queensland, St Lucia, QLD Australia, December 1999.

No context found.

Rupert D. Paget. Nonparametric Markov random field models for natural texture images. PhD thesis, University of Queensland, February 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC