Changes between Version 28 and Version 29 of SandBoxPac

Timestamp:

01/27/09 11:54:29 (16 years ago)

Author:

pacitu

Comment:

--

SandBoxPac

@@ -24 +24 @@
 == Setting up Postgresql ==
 
+INTRODUCTION
+Mathematical morphology (MM) offers a variety of tools for texture char-
+acterization, such as granulometry, morphological covariance, orientation
+maps, etc. The 
+rst two in particular have been employed successfully in a
+number of texture analysis applications [3, 7, 22, 23].
+More precisely, granulometry is a powerful tool based on the "sieving"
+principle, implemented by means of successive openings and/or closings with
+structuring elements (SE) of various sizes, hence it is capable of extracting
+shape and size characteristics from textures. Morphological covariance on
+the other hand, is based on erosions with pairs of points separated by vectors
+of various lengths, and provides information on the coarseness, anisotropy
+as well as periodicity of its input.
+In this paper, we concentrate on these two operators, and speci
+cally on
+the combined exploitation of their SE variables: size, distance and direction.
+Since the original size-only de
+nition of pattern spectra [13], these operators
+have been extended in various ways (e.g., color, multivariate, attribute based
+versions, etc.). Relatively recent applications have explored for instance
+the combination of SE shape and size as far as granulometry is concerned
+[24,25], hence leading to a feature matrix rather than a vector, that describes
+Proceedings of the 8th International Symposium on Mathematical Morphology,
+the combined size and shape distribution of its input. As to covariance, the
+coupled use of SE pair distance and direction makes it possible to exploit the
+anisotropic properties of textures additionally to their periodicity [12, 23].
+Here we investigate the ways of combining the complementary infor-
+mation extracted by these two operators (e.g., concatenation, dimension
+reduction, etc.), and propose a hybrid of the two, where SE couples are
+varied in terms of size, direction as well as distance. The proposed combi-
+nation scheme is compared in terms of classi
+cation accuracy, against the
+standard de
+nitions, using the publically available Outex13 color texture
+database. The so far obtained experimental results show that it leads to an
+improvement over the usual concatenation of feature vectors.
+Furthermore, as far as the extension of this operator to color images is
+concerned, since MM is based on complete lattice theory, a vector ordering
+mechanism becomes necessary. Hence, we propose a weight based reduced
+vector ordering, de
+ned on the improved HLS (IHLS) color space, designed
+speci
+cally for the purpose of color texture classi
+cation. This approach
+makes it possible to optimize, for instance through genetic algorithms, the
+weight of each component adaptively, according to the training set under
+consideration.
+The rest of the paper is organized as follows. Section 2 introduces brie
+y
+granulometry and covariance, and then elaborates on the combination of
+their variables. In Section 3, the problem of extending morphological op-
+erators to multivariate images is discussed, and the proposed ordering is
+detailed. Next, Section 4 presents the experimental results that have been
+obtained with the Outex13 database. Finally, Section 5 is devoted to con-
+cluding remarks.
 
-AsdAsd
-SandBox
+