Fast Learning VIEWNET Architectures for Recognizing 3-D Objects from Multiple 2-D Views

The recognition of 3-D objects from sequences of their 2-D views is modeled by a family of self-organizing neural architectures, called VIEWNET, that use View Information Encoded With NETworks. VIEWNET incorporates a preprocessor that generates a compressed but 2-D invariant representation of an image, a supervised incremental learning system that classifies the preprocessed representations into 2-D view categories whose outputs arc combined into 3-D invariant object categories, and a working memory that makes a 3-D object prediction by accumulating evidence from 3-D object category nodes as multiple 2-D views are experienced. The simplest VIEWNET achieves high recognition scores without the need to explicitly code the temporal order of 2-D views in working memory. Working memories are also discussed that save memory resources by implicitly coding temporal order in terms of the relative activity of 2-D view category nodes, rather than as explicit 2-D view transitions. Variants of the VIEWNET architecture may also be used for scene understanding by using a preprocessor and classifier that can determine both What objects are in a scene and Where they are located. The present VIEWNET preprocessor includes the CORT-X 2 filter, which discounts the illuminant, regularizes and completes figural boundaries, and suppresses image noise. This boundary segmentation is rendered invariant under 2-D translation, rotation, and dilation by use of a log-polar transform. The invariant spectra undergo Gaussian coarse coding to further reduce noise and 3-D foreshortening effects, and to increase generalization. These compressed codes are input into the classifier, a supervised learning system based on the fuzzy ARTMAP algorithm. Fuzzy ARTMAP learns 2-D view categories that are invariant under 2-D image translation, rotation, and dilation as well as 3-D image transformations that do not cause a predictive error. Evidence from sequence of 2-D view categories converges at 3-D object nodes that generate a response invariant under changes of 2-D view. These 3-D object nodes input to a working memory that accumulates evidence over time to improve object recognition. ln the simplest working memory, each occurrence (nonoccurrence) of a 2-D view category increases (decreases) the corresponding node's activity in working memory. The maximally active node is used to predict the 3-D object. Recognition is studied with noisy and clean image using slow and fast learning. Slow learning at the fuzzy ARTMAP map field is adapted to learn the conditional probability of the 3-D object given the selected 2-D view category. VIEWNET is demonstrated on an MIT Lincoln Laboratory database of l28x128 2-D views of aircraft with and without additive noise. A recognition rate of up to 90% is achieved with one 2-D view and of up to 98.5% correct with three 2-D views. The properties of 2-D view and 3-D object category nodes are compared with those of cells in monkey inferotemporal cortex.

imátgenes 1, 2 y 3

Esta sección encontrará sus lectores más inmediatos entre las personas vinculadas al mundo académico y matemático, pero lo ideal sería que estas páginas pudieran llegar más allá. Tú, lector, podrías ser quien concretara el sentido educativo de la sección invitando a tus alumnos, familiares y amigos a relacionar lo que ven con las matemáticas. Por el nivel de matemáticas necesario no deberían preocuparse, ya que se restringirá al de la E.S.O. y el Bachillerato. Todo lo que vemos son imágenes. Pensando en ellas buscamos en nuestro conocimiento modos de interpretarlas y entenderlas. Ahora se propone la reflexión sobre imágenes no con la intención de efectuar una descripción matemática gratuita de lo que se ve mediante la aplicación de conocimiento matemático, sino más bien al contrario: observar cómo las matemáticas pueden resultar imprescindibles para comprender lo que vemos. La idea es desvelar el trasfondo matemático subyacente en las imágenes, de ahí el título de la sección: imátgenes. Una iMATgen será una imagen portadora de matemáticas esenciales para su comprensión. Nada impide realizar un juego de palabras con un cariz biológico. Puesto que ante una misma imagen dos personas pueden dar interpretaciones diferentes, una imagen puede admitir dos iMÁTgenes distintas. Por eso ofrecemos la posibilidad de participar en la sección mediante el correo electrónico: "imatgenes.suma@fespm.org". Cualquier comentario, sugerencia o ayuda será bien recibida. Muchas gracias. ¡Que lo veáis bien!

Expression of gentisate 1,2-dioxygenase (gdoA) genes involved in aromatic degradation in two haloarchaeal genera.

Gentisate-1,2-dioxygenase genes (gdoA), with homology to a number of bacterial dioxygenases, and genes encoding a putative coenzyme A (CoA)-synthetase subunit (acdB) and a CoA-thioesterase (tieA) were identified in two haloarchaeal isolates. In Haloarcula sp. D1, gdoA was expressed during growth on 4-hydroxybenzoate but not benzoate, and acdB and tieA were not expressed during growth on any of the aromatic substrates tested. In contrast, gdoA was expressed in Haloferax sp. D1227 during growth on benzoate, 3-hydroxybenzoate, cinnamate and phenylpropionate, and both acdB and tieA were expressed during growth on benzoate, cinnamate and phenylpropionate, but not on 3-hydroxybenzoate. This pattern of induction is consistent with these genes encoding steps in a CoA-mediated benzoate pathway in this strain.

The stereoselective coordination chemistry of the helicating ligand N,N '-bis(-2,2 '-dipyridyl-5-yl)carbonyl-(S/R,S/R)-1,2-diphenylethylenediamine

Enantiomerically pure N,N'-bis(-2,2'-dipyridyl-5-yl)carbonyl-(S/R,S/R)-1,2-diphenylethylenediamine has been synthesised by linking two 2,2'-bipyridine units by (R,R)- and (S,S)-1,2-diphenylethylenediamine. The ligands possess a hindered rotation between the bipyridine chromophores, which are held together by intramolecular hydrogen bonds. ES mass spectroscopy confirmed that reaction with Fe(II), Co(III) and Cd(II) afforded dinuclear complexes. CD spectroscopy implied that enantiopure ligands conferred helicity to the metals centre giving a dominant triple helicate diastereoisomer (with the RR isomer giving a P helicate). H-1 NMR spectroscopy of the cadmium complex confirmed the presence of a single diastereoisomer. (C) 2003 Elsevier B.V. All rights reserved.

1,2-Cyclic sulfite and sulfate furanoside diesters: improved syntheses and stability

The facile syntheses of 1,2- and 3,5-cyclic sulfite and sulfate furanoside diesters were conducted in molecular solvents and ionic liquids in the presence of immobilised morpholine. Molecular solvents and ionic liquids performed similarly with regards to overall yields. However, the use of ILs allowed for the reactions to be carried out under atmospheric conditions and showed good recyclability. Additionally, increases in product stability was achieved in ILs over organic solvents, in particular, in bis{(trifluoromethanesulfonyl)imide) and trispentafluoro-ethyltrifluorophosphate-based ionic liquids, which were also excellent media to control the hydrolysis of thionyl chloride and sulfuryl chloride. (C) 2009 Elsevier Ltd. All rights reserved.

beta-1,2,3-Triazolyl-Nucleosides as Nicotinamide Riboside Mimics

The synthesis of a series of pyridine- and piperidine-substituted 1,2,3-triazolides linked to a riboside moiety is described. The presence of a triazolide substituent on the pyridine moiety permitted the facile reduction of the latter under mild hydrogenation conditions. These analogues were modelled as to define their similarity to nicotinamide riboside and quantify their ability to bind NAD-dependent protein deacetylases.

The solution structure of 1:2 phenol/N-methylpyridinium bis{(trifluoromethyl)sulfonyl}imide liquid mixtures

Neutron diffraction has been used to investigate the liquid structure of a 1:2 solution of phenol in the ionic liquid N-methylpyridinium bis{(trifluoromethyl)sulfonyl}imide at 60 ◦C, using the empirical potential structure refinement (EPSR) process to model the data obtained from the SANDALS diffractometer at ISIS. Addition of phenol results in suppression of the melting point of the pyridinium salt and formation of a room temperature solution with aromatic phenol–cation and phenol-OH to anion hydrogen-bonding interactions.

Preparation of Dess-Martin periodinane - The role of the morphology of 1-hydroxy-1,2-benziodoxol-3(1H)-one 1-oxide precursor

1-Hydroxy-1,2-benziodoxol-3(1H)-one I-oxide prepared by oxidation of o-iodobenzoic acid with potassium bromate forms either a microcrystalline powder, a macrocrystalline material, or a mixture of both forms. This difference in physical form is the source of the difficulty in reproducibly converting 1-hydroxy-1,2-benziodoxol-3(1H)-one 1-oxide to the corresponding I-triacetoxy derivative. A simple method is given for conversion of crystalline 1-hydroxy-1,2-benziodoxol-3(1H)-one 1-oxide to the more reactive powder form, The microcrystalline powder and macrocrystalline material are characterised by X-ray diffraction.

LL-37 quantification in gingival crevicular fluid (GCF) from periodontitis patients

Background: LL-37, an anti-microbial peptide belonging to the cathelicidin family, derives its name from two N-terminal leucine residues and the 37 amino acids comprising the peptide. LL-37 is the only known cathelicidin to exist in humans. It exhibits both anti-bacterial and immunomodulatory properties. Objectives: In the current study, LL-37 was quantified in GCF from periodontitis patients. Previous studies have relied on qualitative results from Western blotting to detect LL-37 in GCF. This study aims to quantitatively determine LL-37 levels in GCF. Methods: GCF and bacterial plaque samples, pre- and post non-surgical periodontal treatment, were collected from 4 sites in 12 patients presenting with advanced periodontitis. Plaque samples were analysed by QPCR for the presence or absence of the periodontopathic bacterium Porphyromonas gingivalis (P. gingivalis). The concentrations of LL-37 in patient samples pre- and post-treatment were deduced by indirect enzyme linked immunosorbent assay (ELISA). Results: Concentrations of LL-37 in samples varied between a minimum and maximum of 1 and 40 ng/ml. LL-37 levels were shown, pre-treatment, to be higher in deep pockets (6-9 mm) compared with shallower pockets (3-5 mm) and highest in those sites which were positive for P. gingivalis. Non-surgical therapy resulted in a significant improvement in clinical indices while expression levels of P. gingivalis were reduced. Following treatment, LL-37 levels in GCF decreased from an average of 6.5 ± 1 - 5.8 ± 1.2 ng/ml. The most interesting observation however was the reduction in LL-37 levels, from an average of 7 ± 1.3 – 2.5 ± 1.1 ng/ml in those sites where P. gingivalis infection was eradicated post-treatment. Conclusions: LL-37 levels are increased at sites showing advanced periodontal disease, reduce following treatment and appear to be linked to the presence of P. gingivalis. This study will further our knowledge of host defence in chronic diseases such as periodontitis.