Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: texture images

Sortuj według:

Ogranicz wyniki do:

Computer analysis of images of worn surfaces.

Oleksiak Z., Rudnicki Z.

Zagadnienia Eksploatacji Maszyn

2002

Vol. 37, nr 3

73-84

Digitized images of tribology surfaces belong to a large class of texture images which are frequently analyzed in many areas, mainly for diagnostic and quality control tasks. The paper describes some methods of texture image analysis, discrimination and numerical description on examples of two classes of worn surface images. The considered problems are dealing with the texture image feature extraction and selection of the best features to describe and to discriminate classes of images. The two approaches are presented: (1) extraction of large, redundant feature set and automatic selection of the best features; (2) intuitive feature definition and image processing to enhance the best discriminating features - proposed by authors. The methods were compared by use of Fisher's coefficient.

Cyfrowe obrazy powierzchni badanych w tribologii należą do obszernej klasy obrazów teksturowych analizowanych w wielu dziedzinach, głównie dla celów diagnostyki i kontroli jakości. Artykuł opisuje pewne metody analizy, rozróżniania i opisu liczbowego obrazów teksturowych zrealizowane dla dwu klas obrazów powierzchni o różnym stopniu zużycia. Rozważane są problemy ekstrakcji cech obrazów teksturowych i wyboru najlepszych cech dla opisu i rozróżniania badanych klas obrazów. Zaprezentowano dwa podejścia (1) automatycznej selekcji najlepszych cech z obszernego zbioru ekstrahowanych cech, (2) intuicyjnego określenia najlepszych cech i odpowiedniego ich uwydatnienia przez przetwarzanie obrazów - wg propozycji autorów. Metody porównano przy użyciu współczynnika Fishera.

Pairwise Markov random fields and segmentation of textured images

Pieczynski W., Tebbache A. N.

Machine Graphics and Vision

2000

Vol. 9, No. 3

705-718

The use of random fields, which allows one to take into account the spatial interaction among random variables in complex systems, becomes a frequent tool in numerous problems of statistical mechanics, spatial statistics, neural network modelling, and others. In particular, Markov random field based techniques can be of exceptional efficiency in some image processing problems, like segmentation or edge detection. In statistical image segmentation, that we address in this work, the model is generally defined by the propability distributions of the observations field conditional to the class field. Under some hypotheses, thr a posteriori distribution of the class field, i.e. conditional to the observations field, a still a Markov distribution and the latter property allows one to apply different bayesian methods of segmentation like Maximum a Posteriori (MAP) or Maximum a Posterior MOde (MPM). However, in such models the segmentation of textured images is difficult to perform and one has to resort to some model approximations. The originality of our contribution is to consider the markovianity of the couple (class field, observations field). We obtain a different model; in particular, the class field is not necessarily a Markov field. However, the posterior distribution of the class field is a Markov distribution, which makes possible bayesian MAP and MPM segmentations. Furthermore, the model proposed makes possible textured image segmentation with no approximations.