Comparing large graphs efficiently by margins of feature vectors

Measuring the structural similarity of graphs is a challenging and outstanding problem. Most of the classical approaches of the so-called exact graph matching methods are based on graph or subgraph isomorphic relations of the underlying graphs. In contrast to these methods in this paper we introduce a novel approach to measure the structural similarity of directed and undirected graphs that is mainly based on margins of feature vectors representing graphs. We introduce novel graph similarity and dissimilarity measures, provide some properties and analyze their algorithmic complexity. We find that the computational complexity of our measures is polynomial in the graph size and, hence, significantly better than classical methods from, e.g. exact graph matching which are NP-complete. Numerically, we provide some examples of our measure and compare the results with the well-known graph edit distance. (c) 2006 Elsevier Inc. All rights reserved.

Deep-sea chitons from sunken wood in the West Pacific (Mollusca: Polyplacophora: Lepidopleurida): taxonomy, distribution, and seven new species

Natural deposits of sunken wood provide an important habitat for deep-sea invertebrates. Deep-sea chitons in the primitive order Lepidopleurida are typically collected rarely and as single specimens. However, these animals have been recovered in large densities associated with sunken wood in the tropical West Pacific, in groups of up to 50 individuals. Four deep- sea expeditions in the West Pacific, to the Philippines, Solomon Islands, and Vanuatu, recovered a large number of poly- placophorans. We have examined the morphology as well as the range and distribution of these species, based on the larg- est collection ever examined (more than 1300 individuals). These species show potentially adapted characters associated with exploitation of sunken wood as habitat, such as protruding caps on sensory shell pores (aesthetes) and large interseg- mental bristles with potential sensory function. In this study we investigated the twenty-two species recovered, including seven newly described here (Leptochiton consimilis n. sp., L. angustidens n. sp., L. dykei n. sp., L. samadiae n. sp., L. longisetosus n. sp., L. clarki n. sp., L. schwabei n. sp.), and provide the first identification key to the 34 lepidopleuran chitons known from sunken wood worldwide.

Taxonomy and phylogeny of Osmundea (Rhodomelaceae, Rhodophyta) in Atlantic Europe

Two species of Osmundea Stackhouse (Rhodomelaceae, Rhodophyta) that occur in Atlantic Europe have been confused under the names Osmundea ramosissima (Oeder) Athanasiadis and Osmundea truncata (Kutzing) Nam et Maggs, regarded until now as a synonym of O. ramosissima, An epitype from its type locality (Stavanger, Norway) is selected for Osmundea ramosissima Athanasiadis, recognized here as a valid name for Fucus ramosissimus Oeder, nom. illeg. Details of vegetative and reproductive morphology of O. ramosissima are reported, based on material from France, the British Isles, and Helgoland. Osmundea ramosissima resembles other species of Osmundea in its vegetative axial segments with two pericentral cells and one trichoblast, spermatangial development from apical and epidermal cells (filament type), the formation of five pericentral cells in the procarp-bearing segment of the female trichoblast, and tetrasporangial production from random epidermal cells. Among the species of Osmundea, O. ramosissima is most similar to O. truncata. Both species have discoid holdfasts, secondary pit connections between epidermal cells, and cup-shaped spermatangial pits. They differ in that: (a) O. ramosissima lacks lenticular wail thickenings and refractive needle-like inclusions in medullary cells, both of which are present in O. truncata; (b) O. ramosissima has branched spermatangial filaments that terminate in a cluster of several cells, whereas in O. truncata the unbranched spermatangial filaments have a single large terminal sterile cell; and (c) cystocarps of O. ramosissima lack protuberant ostioles but ostioles are remarkably protuberant in o. truncata. Phylogenetic analyses of rbcL sequences of Laurencia obtusa (Hudson) Lamouroux and all five Atlantic European species of Osmundea, including the type species, strongly support the generic status of Osmundea. Osmundea ramosissima and O. truncata are closely related (5.2% sequence divergence) and form a well-supported clade sister to a clade consisting of O. pinnatifida (Hudson) Stack-house, O. osmunda Stackhouse and O. hybrida (A. P. de Candolle) Nam. The formation of secondary pit connections between epidermal cells is a synapomorphy for the O. ramosissima + O. truncata clade. The close relationship between species with cup-shaped spermatangial pits (Osmundea hybrida) and urn-shaped pits (Osmundea pinnatifida and Osmundea osmunda) shows that spermatangial pit shape is not an important phylogenetic character. Parsimony analysis of a morphological data set also supports the genus Osmundea but conflicts with the molecular trees in infrageneric relationships, placing O. hybrida basal within the Osmundea clade and grouping O. osmunda and O. pinnatifida but not O. truncata and O. ramosissima. A key to Osmundea species is presented.

AdaBoost Multiple Feature Selection and Combination for Face Recognition

Gabor features have been recognized as one of the most successful face representations. Encouraged by the results given by this approach, other kind of facial representations based on Steerable Gaussian first order kernels and Harris corner detector are proposed in this paper. In order to reduce the high dimensional feature space, PCA and LDA techniques are employed. Once the features have been extracted, AdaBoost learning algorithm is used to select and combine the most representative features. The experimental results on XM2VTS database show an encouraging recognition rate, showing an important improvement with respect to face descriptors only based on Gabor filters.

Max separation clustering for feature extraction from Optical Emission Spectroscopy Data

This paper proposes max separation clustering (MSC), a new non-hierarchical clustering method used for feature extraction from optical emission spectroscopy (OES) data for plasma etch process control applications. OES data is high dimensional and inherently highly redundant with the result that it is difficult if not impossible to recognize useful features and key variables by direct visualization. MSC is developed for clustering variables with distinctive patterns and providing effective pattern representation by a small number of representative variables. The relationship between signal-to-noise ratio (SNR) and clustering performance is highlighted, leading to a requirement that low SNR signals be removed before applying MSC. Experimental results on industrial OES data show that MSC with low SNR signal removal produces effective summarization of the dominant patterns in the data.