For some new multimedia applications using Mpeg-4 or Mpeg-7 video coding standards, it is important to find the main objects in a video frame. In this thesis, we propose a neuro-fuzzy modeling approach to the detection of human face and body. Firstly, a fuzzy clustering technique is performed to segment a video frame into clusters to generating several fuzzy rules. Secondly, chrominance and motion features are used to roughly classify the clusters into foreground and background, respectively. Finally, the fuzzy rules are refined by a fuzzy neural network, and the ambiguous regions between foreground and background are further distinguished by the fuzzy neural network. Our method improves the correctness of human face and body detection by getting training data more precisely. Besides, we can extract the VOs correctly even the VOs have no obvious motion in the video sequence.

none

A gauged SU(4) flavor symmetric hadronic Lagrangian with empirical hadron masses is constructed to study charmonium absorption and charmed hadron production in hadronic reactions. For the coupling constants, empirical values are used if available. Otherwise, they are determined from known coupling constants using the SU(4) relations. To take into account the finite sizes of hadrons, form factors are introduced at strong interaction vertices with empirical cutoff parameters. For J/ψabsorption by nucleons, we have included both two-and three-body final states and find that with a cutoff parameter of 1 GeV at interaction vertices involving charm hadrons, the cross section is at most 5 mb and is consistent with that extracted from J/ψproduction from both photo-and proton-nucleus reactions. We have also evaluated the cross sections for charmed hadron production from pion and rho meson interactions with nucleons. With the same cutoff parameter of 1 GeV at interaction vertices, we find that these cross sections have values of a few tenths of mb and are dominated bythe s-channel nucleon pole diagram. For charmed hadron production from proton-proton reactions, their cross sections including bothtwo-andthree-body final states are about 1 batcenter-of-mass energyof 11.5 GeV, which is comparable to the measured inclusive cross section in these reactions.Including photon as a U(1) gauge particle, we have extended the model to study charmed hadron production in photon-proton reactions with both two-and three-body final states included. For form factors, an overall one is introduced in each processin order to maintain the gauge invariance of the total amplitude. Fitting the cutoff parameter in the form factor to the measured total cross section for charmed hadron production in photon-proton reactions at a center-of-mass energy of 6 GeV, the ratio of the cross sections for two-body and three-body final states is consistent with available experimental data.This result is further compared with predictions from the leading-order perturbative QCD calculation. Knowledge of the cross sections for charmonium absorption byhadrons and for charmed hadron production in hadronic reactions is essential for understanding charm production in heavy ion collisions at the Relativistic Heavy Ion Collider (RHIC), where a quark-gluon plasma is expected to be formed during the initial hot dense stage.

Soft corals belonging to the genus Nephthea have been found to be a rich source of terpenoids and steroids. Chromatographic separation of acetone and CH2Cl2/MeOH extracts of Formosan soft coral Nephthea chabrolii resulted in the isolation of thirteen new steroids, including nine 4£\-methylated steroids (1−6, 8−10), three 19-oxygenated steroids (11−13), and a novel skeleton 19-norsteroid (7). The structures of these metabolites were elucidated by extensive spectral analyses and anti-inflammatory activities were measured in vitro. Compounds 1−2, 4−5, 7−11 and 13 at a concentration of 10 £gM significantly reduce the levels of iNOS (inducible nitric oxide synthase ). Compounds 2, 4, 6−7 and 10−11 at a concentration of 10 £gM also significantly reduce the levels of COX-2 (cyclooxygenase-2). Compounds 2, 4, 7, 10 and 11 can reduce the levels of both iNOS and COX-2. All compounds at a concentration of 10 £gM did not affect the housekeeping £]-actin protein expression.

Traditional Mobile IP supports only one tunnel between HA (Home Agent) and FA (Foreign Agent) for MN (Mobile Node) to transmit data from CN (Corresponding Node) to MN during handoff. To alleviate traffic load, in this Thesis, we propose a multiple-tunnel approach by setting up a secondary tunnel in addition to the primary tunnel. HA can split the video stream from the primary tunnel to the secondary tunnel when the former encounters high traffic load. Thus, quality of video stream from CN to MN can be guaranteed through the proposed multiple-tunnel schemes. For the purpose of validation, we implement the multiple-tunnel schemes on Linux platform for HA, FA, and MN to support video stream splitting. Specifically, if HA detects that high system load occurs on FA, an MPEG-4 packet filter is activated on HA to split some portions of video packets from the primary tunnel to the secondary tunnel based on I/B/P frame types. Video receiver on MN can then successfully recover these two divided video packets and playback the complete stream smoothly. Three experiments for validation are developed. They are (i) a simple Mobile IP handoff, (ii) Different bit rates to activate the stream splitting mechanism, and (iii) automatic adjustment of stream splitting according to system load. The experimental results have proved that the proposed multiple-tunnel schemes can substantially improve the quality of video streams.

No description available.

<p>Syftet med detta arbete är söka svar på hur elever i de tidiga skolåren löser praktiska problem inom matematikområdena, mätning och vägning. Det har gjorts en jämförande studie av elevernas lösningsstrategier för att undersöka huruvida det sker någon förändring med stigande ålder och därmed förväntade ökade kunskaper i matematik.</p><p>Genom undervisningsförsök med kompletterande löpande observationer har vi sökt svar på hur förskoleklasselevernas lösningsstrategier skiljer sig i jämförelse med elevernas i skolår 4. Försöken har genomförts av elever i förskoleklass och skolår 4. Eleverna har varit indelade i grupper om 4-6 i varje och alla undervisningsförsök har genomförts utomhus. Undersökningen har bestått av två övningar, ett försök i mätning och ett försök i vägning.</p><p>Resultatet visar att eleverna i de olika åldrarna uppskattar att lösa matematiska problem, på ett praktiskt sätt. Det finns märkbara skillnader mellan de yngre och de äldre elevernas sätt att lösa de olika praktiska problemen. De yngre eleverna har ännu inte tillägnat sig formella kunskaper i matematik i lika stor utsträckning som de äldre eleverna har.</p>

Epstein-Barr Virus (EBV) has two states of infection, latent and lytic. During the latent state the viral genome remains stable in cells as episomes and replicates with cellular DNA. During the lytic cycle the viral DNA becomes amplified and packaged in newly formed virions. An unsolved problem is whether newly replicated EBV DNA produced upon lytic cycle activation is associated with histones, and if so, whether these histones are acetylated. This question has biological significance as knowing the chromatin structure of genes is important in determining their function and expression profile. Our hypothesis is that newly synthesized EBV lytic DNA is associated with histones and the histone tails are selectively acetylated. To investigate our hypothesis we performed chromatin immunoprecipitation (ChIP) in HH514-16 cells, a Burkitts Lymphoma cell line, during latent and lytic replication. We used quantitative PCR (qPCR) to detect the relative concentration of DNA among the different samples. We tested three different variables: type of inducing agent, duration of treatment, and different regulatory regions in the genome of Epstein-Barr Virus. We found that in cells induced into the lytic cycle with Trichostatin A (TSA), a histone deacetylase inhibitor (HDACi), association of newly replicated EBV DNA with acetylated histone 3 (H3) increased ~ 6-10 fold. This increase in association was greatest 72 hrs after treatment. Furthermore, activation of lytic viral replication in HH514-16 cells using a different inducing agent, Azacytidine (AZC), which is known to function as a DNA methyltransferase inhibitor, increased binding of H3 with viral DNA ~8 fold. However, unlike TSA, AZC increased the acetylation state of histones bound to newly synthesized viral DNA only ~ 2 fold. Changing the regulatory region of the EBV genome analyzed in qPCR did not affect our results. Our results suggest that newly replicated viral DNA is associated with histones, a fraction of which are acetylated. The degree of acetylation likely depends on the agent used to induce the lytic cycle. H3 is highly acetylated when an HDACi is used and less acetylated when AZC is used. Our study provides new insight on the epigenetic profile of newly replicated viral DNA during the lytic cycle. It remains to be determined whether histones are packaged together with viral genomes into virions and whether the chromatin state of virion DNA affects gene expression after the virus enters uninfected cells.

The glucose transporter 4 (GLUT4) is the major insulin-responsive glucose transporter in adipose and muscle tissues. Although the early steps in the insulin signaling pathway governing translocation of GLUT4 to the plasma membrane are well understood, the final steps in this pathway are not. TUG is a protein which has been shown to affect trafficking of GLUT4 both in the basal state and in response to insulin. One protein-protein interaction between TUG and the large cytosolic loop of GLUT4 has previously been identified. Based on reports of the requirement of the GLUT4 N-terminal domain for its proper targeting to the plasma membrane, we postulated that an interaction might also exist between TUG and the N-terminal domain of GLUT4, and we tested this hypothesis using two sets of pull-down experiments. In the first set, using the N-terminal domain of GLUT4 fused with glutathione S-transferase (GST), we were able to pull TUG down from the lysates of TUG-transfected HEK 293 cells. TUG was also pulled down by the GLUT4 cytosolic loop and, to a much lesser extent, its C-terminal domain. However, there was no specific interaction between these fusion proteins and the lysates of cells transfected with a truncated form of TUG lacking its own N-terminal domain. In the second set of experiments, using a biotinylated synthetic GLUT4 N-terminal peptide, we pulled down a protein detected by an anti-TUG antibody and running at ~64 kDa, a slightly higher molecular weight than wild-type TUG. We believe that this band represents modified full-length TUG. This interaction was not seen using synthetic GLUT4 N-terminal peptide mutated at 4 amino acids previously identified as necessary for proper GLUT4 retention and insulin-responsive trafficking. We conclude that TUG interacts not only with the large cytosolic loop of GLUT4, but also with the N-terminal domain of GLUT4, and that this latter interaction can be disrupted by mutations in GLUT4 that cause defective trafficking, suggesting that this interaction is critical for GLUT4 intracellular retention and insulin-responsive GLUT4 trafficking.

Thèse doctorat : Sciences de la vie et de la santé : Paris 12 : 2003. / Titre provenant de l'écran-titre.