Unsupervised texture segmentation with Markov random field models and self organizing maps

Download
056887.pdf
1996
Göktepe, Mesut
Show full item record
42
views
0
downloads
Markov random fields., Self-organizing systems.
https://hdl.handle.net/11511/1019
Graduate School of Natural and Applied Sciences, Thesis

Suggestions

Unsupervised segmentation of gray level Markov model textures with hierarchical self organizing maps
Goktepe, Mesut; Yalabik, Nese; Atalay, Mehmet Volkan (1996-01-01)
Segmentation of gray level images into regions of uniform texture is investigated. An unsupervised approach through the use of Kohonen's self organizing map (SOM) and a multilayer version of it, the hierarchical self organizing map (HSOM), is employed to find the regions in an image composed of textures from different classes. For testing, gray level artificial textured images modeled as Markov random fields are used as the input. No parameter estimation is done. The size and the topology of SOM and HSOM ar...
Arbitrarily Long Factorizations in Mapping Class Groups
DALYAN, ELİF; Korkmaz, Mustafa; Pamuk, Mehmetcik (2015-01-01)
On a compact oriented surface of genus g with n= 1 boundary components, d1, d2,..., dn, we consider positive factorizations of the boundary multitwist td1 td2 tdn, where tdi is the positive Dehn twist about the boundary di. We prove that for g= 3, the boundary multitwist td1 td2 can be written as a product of arbitrarily large number of positive Dehn twists about nonseparating simple closed curves, extending a recent result of Baykur and Van Horn- Morris, who proved this result for g= 8. This fact has immed...
Random walk particle modelling of polymer injection using matlab reservoir simulation toolbox
Mamak, Gökhan; Durgut, İsmail; Department of Petroleum and Natural Gas Engineering (2017)
Enhanced oil recovery (EOR) is essential to increase the maximum recoverable oil by natural means of production. Having chosen an EOR method, the effectiveness of the method should be analyzed before applying to a reservoir since the methods are generally costly. Polymer injection is a chemical EOR process, where the injected polymer with water increases the water viscosity, and help increasing the sweep efficiency in the reservoir. In order to model the effects of polymer injection, the random-walk particl...
Time memory trade off attack on symmetric ciphers
Saran, A. Nurdan; Doğanaksoy, Ali; Department of Cryptography (2009)
Time Memory Trade O (TMTO) is a cryptanalytic method that aims to develop an attack which has a lower memory complexity than lookup table and a lower online time complexity than exhaustive search. TMTO methods are widely studied in the literature and used for inverting various cryptosystems. We focus on the design and the analysis of TMTO on symmetric ciphers in this thesis. Firstly, the summary of the random mapping statistics from the view point of TMTO is presented. We also recalculate some expected valu...
Low-Level Hierarchical Multiscale Segmentation Statistics of Natural Images
Akbaş, Emre (2014-09-01)
This paper is aimed at obtaining the statistics as a probabilistic model pertaining to the geometric, topological and photometric structure of natural images. The image structure is represented by its segmentation graph derived from the low-level hierarchical multiscale image segmentation. We first estimate the statistics of a number of segmentation graph properties from a large number of images. Our estimates confirm some findings reported in the past work, as well as provide some new ones. We then obtain ...
Citation Formats
M. Göktepe, “Unsupervised texture segmentation with Markov random field models and self organizing maps,” Ph.D. - Doctoral Program, Middle East Technical University, 1996.
METU IIS - Integrated Information Service open@metu.edu.tr