No description available.

This thesis proposes an approach to segmenting and identifying mixed-language speech. Automatic LID can be divided into four steps, feature extraction, segmentation, segment clustering, and re-labeling. In feature extraction, we compare the group delay feature (GDF) with MFCC feature. Unlike the traditional feature from Fourier trans-form magnitude, GDF uses the phase spectrum. In segmentation, we compare delta Bayesian information criterion (delta-BIC) with support vector machines (SVMs). A delta-BIC is applied to segment the input speech utterance into a sequence of lan-guage-dependent segments using acoustic features. The segments are clustered using the K-means algorithm. Finally, re-labeling is used to determine the language of the clusters. SVMs proceed to segment and identify automatically after model training. Considering the effect of the accent issue, we use the corpus English Across Taiwan (EAT) to perform our system. The experimental results show that the system can reach 78.13% in the frame hit rate under the baseline 57.77%.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-04-10T18:04:35Z No. of bitstreams: 1 rafaelsilvagama.pdf: 1774054 bytes, checksum: b50899947c904ace42800056fa0d9afc (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-04-11T15:19:43Z (GMT) No. of bitstreams: 1 rafaelsilvagama.pdf: 1774054 bytes, checksum: b50899947c904ace42800056fa0d9afc (MD5) / Made available in DSpace on 2018-04-11T15:19:43Z (GMT). No. of bitstreams: 1 rafaelsilvagama.pdf: 1774054 bytes, checksum: b50899947c904ace42800056fa0d9afc (MD5) Previous issue date: 2017-10-30 / O diagnóstico da hanseníase é eminentemente clínico, podendo ser complementado com baciloscopia, histopatologia e testes imunológicos. Métodos bacteriológicos clássicos para identificação de bactérias patogênicas não podem ser aplicados para o diagnóstico de hanseníase, sobretudo pela impossibilidade de cultivo in vitro do M. leprae. O exame histopatológico e a baciloscopia têm sido utilizados como métodos auxiliares para a classificação clínica dos casos. Técnicas moleculares e sorológicas têm sido avaliadas como ferramentas de diagnóstico na hanseníase. Ao contrário da baciloscopia, que requer cerca de 104 organismos por grama de tecido para detecção real, a PCR é uma técnica de alta especificidade e sensibilidade, capaz de detectar 25 fg (10-15g) de DNA de M. leprae. Além disso, a possibilidade de sua utilização em quase todos os tipos de amostras clínicas confere a este método um alto potencial para o diagnóstico diferencial. Os testes sorológicos têm como alvo a detecção de anticorpos específicos contra o M. leprae que indicam infecção. Esses testes podem ser úteis no monitoramento da eficácia da terapia, na determinação da prevalência da doença e na avaliação da distribuição da infecção entre contatos domiciliares. Considera-se que o indivíduo que reside ou tenha residido com o doente de hanseníase apresenta maior risco de adoecimento em relação à população em geral, pelo fato de estarem expostos ao M. leprae. Familiares de pacientes Multibacilar e Paucibacilar devem ser examinados, independente do tempo de convívio. Sugere-se avaliar anualmente, durante cinco anos, todos os contatos não doentes, quer sejam familiares ou sociais. Neste estudo utilizou-se o suporte da Inteligência Artificial (Random Forest) para análise integrada de métodos sorológicos e moleculares, no diagnóstico de novos casos de hanseníase e monitoramento de contatos domiciliares por um período de cinco anos. O estudo foi desenvolvido em Governador Valadares – MG, considerada área endêmica de hanseníase. O desenho proposto é do tipo longitudinal, com coleta de dados cadastrais de todos os casos diagnosticados em 2011 e seus respectivos contatos domiciliares registrados no período de 2011, 2012 e 2016. Um total de 196 indivíduos, sendo 43 casos, 113 contatos domiciliares e 40 indivíduos considerados controles endêmicos que relataram não ter convívio com pacientes com hanseníase nem tão pouco histórico de hanseníase na família foi incluído no estudo. Foram coletadas amostras sangue e de raspado intradérmico dos casos de hanseníase e seus respectivos contatos domiciliares, para análise por qPCR (16S rRNA) e ELISA (anti NO-OLID e LID-1). A análise integrada dos dados foi realizada por meio da Random Forest com o objetivo de melhorar o desempenho dos testes para o diagnóstico de hanseníase. Isoladamente, a qPCR apresentou sensibilidade de 48,8% e especificidade de 100% no diagnóstico de casos de hanseníase. No ensaio de ELISA anti-ND-O-LID a sensibilidade alcançada foi de 57,9% e especificidade de 97,5%, enquanto que no ensaio de anti-LID-1, a sensibilidade e especificidade foram de 63,2% e 92,5%, respectivamente. Entretanto, a análise integrada dos dados por Random Forest, utilizando 10.000 árvores de decisão, com um erro modal de 12,8%, obteve-se uma taxa de sensibilidade de 81,6% e especificidade de 92,5% na predição de novos casos de hanseníase. O modelo de Random Forest foi utilizado para o monitoramento de contatos domiciliares no período de 05 anos. Esta ferramenta de análise identificou entre os contatos, 02 doentes, mesmo antes do diagnóstico clínico. Ao final do período de acompanhamento, 03 contatos domiciliares foram notificados como casos novos de hanseníase. Desta forma, o modelo proposto pela análise Random Forest permitiu diagnosticar casos de hanseníase com alta sensibilidade e especificidade e identificar precocemente novos casos entre os contatos domiciliares durante o monitoramento. / The diagnosis of leprosy is eminently clinical and may be supplemented with bacilloscopy, histopathology and immunological tests. Classical bacteriological methods for the identification of pathogenic bacteria can not be applied for the diagnosis of leprosy, mainly due to the impossibility of in vitro culture of M. leprae. Histopathological examination and smear microscopy have been used as auxiliary methods for the clinical classification of cases. Molecular and serological techniques have been evaluated as diagnostic tools in leprosy. Unlike smear microscopy, which requires about 104 organisms per gram of tissue for actual detection, PCR is a technique of high specificity and sensitivity, capable of detecting 25 fg (10-15g) of M. leprae DNA. In addition, a possibility of its use in almost all types of clinical specimens gives this method a high potential for differential diagnosis. Serological tests are aimed at detecting specific antibodies against M. leprae that indicate infection. These tests may be useful in monitoring therapy efficacy, in determining disease prevalence, and in assessing the distribution of infection between household contacts. It is considered that the individual who resides or has lived with the leprosy patient has a higher risk of becoming ill in relation to the general population due to the fact that they are exposed to M. leprae. Relatives of patients Multibacillary and Paucibacillary should be examined, regardless of the time of conviviality. It is suggested to evaluate annually, for five years, all non-sick contacts, whether family or social. In this study we used the support of Artificial Intelligence (Random Forest) for integrated analysis of serological and molecular methods, in the diagnosis of new cases of leprosy and monitoring of household contacts for a period of five years. The study was developed in Governador Valadares - MG, considered an endemic area of leprosy. The proposed design is of the longitudinal type, with the collection of cadastral data of all cases diagnosed in 2011 and their respective household contacts registered in the period of 2011, 2012 and 2016. A total of 196 subjects, including 43 cases, 113 household contacts and 40 individuals considered endemic controls who reported not having lived with patients with leprosy or a history of leprosy in the family were included in the study. Blood samples and intradermal scrapings were collected from leprosy cases and their respective household contacts for analysis by qPCR (16S rRNA) and ELISA (anti-ND-O-LID and LID-1). The integrated analysis of the data was performed through Random Forest with the objective of improving the performance of the tests for leprosy diagnosis. In isolation, the qPCR showed sensitivity of 48.8% and specificity of 100% in the diagnosis of leprosy cases. In the anti-NDO- LID ELISA the sensitivity reached was 57.9% and specificity was 97.5%, whereas in the anti-LID-1 assay the sensitivity and specificity were 63.2% and 92.5%, respectively. However, the integrated analysis of the data by Random forest, using 10,000 decision trees, with a modal error of 12.8%, obtained a sensitivity rate of 81.6% and specificity of 92.5% in the prediction of new cases of leprosy. The Random Forest model was used to monitor household contacts within a period of 5 years. This analysis tool identified between the contacts, 02 patients, even before the clinical diagnosis. At the end of the follow-up period, 03 household contacts were reported as new cases of leprosy. Thus, the model proposed by the Random Forest analysis allowed the diagnosis of leprosy cases with high sensitivity and specificity and early identification of new cases among household contacts during monitoring.

In this thesis the solutions of the two-dimensional (2D) and three-dimensional (3D) lid-driven cavity problem are obtained by solving the steady Navier-Stokes equations at high Reynolds numbers. In 2D, we use the streamfunction-vorticity formulation to solve the problem in a square domain. A numerical method is employed to discretize the problem in the x and y directions with a spectral collocation method. The problem is coded in the MATLAB programming environment. Solutions at high Reynolds numbers are obtained up to $Re=25000$ on a fine grid of 131 * 131. The same method is also used to obtain the numerical solutions for the steady separated corner flow at high Reynolds numbers are generated using a for various domain sizes, at various Reynolds number which are much higher than those obtained by other researchers.Finally, the numerical solutions for the three-dimensional lid-driven cavity problem are obtained by solving the velocity-vorticity formulation of the Navier-Stokes equations for various Reynolds numbers. A spectral collocation method is employed to discretize the y and z directions and finite difference method is used to discretize the x direction. Numerical solutions are obtained for Reynolds number up to 200.

This thesis is focused on folkloristic work in Domažlicko, Klatovsko and Sušicko at the end of especially in journal Český lid, published 1891- - estate of Čeněk Zíbrt, the editor of the journal. folklore, ethnographers, Domažlicko, Klatovsko, Sušicko, Český lid, Čeněk Zíbrt

In Drosophila oocytes, chromosomes undergo dynamic reorganisation during the prophase of the first meiotic division. This is essential to prepare chromatin for synapsis, recombination and consequent chromosome segregation. The progression of meiotic prophase I is well described, while the molecular mechanisms and regulation of these dramatic chromosomal reorganisations are not well understood. Histone modifying enzymes are major regulators of chromatin structure, however, our knowledge of their roles in meiotic prophase I is still limited. In this work, I investigated the role of the histone demethylase Kdm5/Lid, which removes one of the trimethyl groups at Lys4 of Histone 3 (H3K4me3). I showed that Kdm5/Lid is important for the assembly of the synaptonemal complex, pairing of homologous centromeres, and the karyosome formation. Additionally, Kdm5/Lid promotes crossing over and therefore ensures accurate chromosome segregation. Although loss of Kdm5/Lid dramatically increased the level of H3K4me3 in oocytes, catalytically inactive Kdm5/Lid rescued the above cytological defects. Thereby, I found that Kdm5/Lid regulates chromatin architecture in meiotic prophase I oocytes independently of its demethylase activity. To further identify the regulators of meiotic chromatin organisation during prophase I, I carried out a small-scale RNAi screen for karyosome defects. I found that depletion of ubiquitin ligase components, SkpA, Cul-3 and Ubc-6, disrupted the karyosome formation and the assembly of the synaptonemal complex. The success of the small-scale screen motivated me to initiate the genome-scale RNAi screen for karyosome defects. I found 40 new genes that, when depleted, strongly impaired karyosome morphology. Further studies are required to confirm and elucidate their role in chromatin organisation in oocytes. Overall, my findings have advanced our understanding of the regulation of chromatin reorganisation during oocyte development. Because of the conservation between Drosophila and human meiosis, this study provides novel insights into the regulation of meiotic progression in human oocytes.

The present work deals with the development of a direct simulation strategy for solving the motion of spherical particles in non-Newtonian liquids. The purely viscous (non-elastic) non-Newtonian liquids are described by Bingham and thixotropy models. Validation of the strategy is performed for single phase (lid driven cavity flow) and two phase flows (sphere sedimentation). Lid driven cavity flow results illustrate the flow evolution of thixotropic liquid and subtle differences between thixotropic rheology and pseudo Bingham rheology. Single sphere sedimentation in Bingham liquid is shown to be influenced by the yield stress of the liquid. Time-dependent properties such as aging prominently affect the settling of a sphere in thixotropic liquid. The hydrodynamic interactions between two spheres are also studied at low and moderate Reynolds numbers. In thixotropic liquid, an intriguing phenomenon is observed where the separation distance between the spheres first increases and then rapidly decreases. / Chemical Engineering

No description available.

The cornea is the most powerful refractive component of the eye and as such, subtle changes in corneal shape can cause substantial changes in the optical characteristics of the eye. Monocular diplopia has previously been linked to corneal distortion following near work in various studies but has not been investigated in detail. The work reported in this thesis has investigated the optical effects of corneal distortions caused by eyelid forces and demonstrated that several corneal higher and lower order Zernike wavefront aberrations can change following reading. Measuring subtle changes in corneal topography requires the highest possible instrument accuracy, while software analysis tools should be able to detect and highlight those subtle changes with high reliability. The effect of ocular microfluctuations on the qualitative and quantitative analysis of corneal topography was investigated. A technique was developed to measure tilt, displacement, and cyclotorsion in multiple videokeratographs from the same cornea. This information was used to reposition each videokeratograph according to the average position of a sample of multiple measurements. The corneal topography of ten subjects was measured 20 times each, using videokeratoscopy. The RMSE calculated from difference between single videokeratographs and the average videokeratograph decreased by an average of 24.6 % for the ten subjects' data. The method can improve the precision performance of videokeratoscopy in multiple measurements of corneal topography. A study was undertaken, to investigate whether there are significant changes in corneal topography during accommodation in normal corneas and corneas that are pathologically thinner due to keratoconus. This was done to eliminate the possibility that changes in corneal aberrations associated with near work could be at least partly due to corneal changes caused by the effects of accommodation. A videokeratoscope was modified to present an accommodation stimulus that was coaxial with the instrument's measurement axis. Six subjects with normal corneas and four subjects with keratoconus were studied. In the initial analysis it was found that a number of the subjects showed significant changes in corneal topography as accommodation changed. However further analysis showed a significant group mean excyclotorsion of the topography maps for both accommodation stimuli compared with the 0 D stimulus. When the excyclotorsion was accounted for, no clear evidence of statistically significant changes in corneal topography as a result of accommodation were found. A small ocular excyclotorsion typically accompanies accommodation and this changes the relative orientation of the topography of the cornea. To investigate the effects of eyelid pressure on corneal shape and corneal aberrations during reading, twenty young subjects with normal ocular health were recruited. Cornea1 topography of one eye was measured with a videokeratoscope prior to reading and then again after a 60 minute reading task. Twelve of the twenty corneas showed significant changes in central topography immediately following reading. The location of the changes corresponded closely to the position and angle of the subject's eyelids during reading. Within the central region of the cornea there were significant changes in corneal wavefront Zernike coefficients, the root-mean-square error, overall refractive power and astigmatism. The changes observed in corneal topography appear to be directly related to the force exerted by the eyelids during reading. These findings may have important implications for the definition of refractive status and may also aid in the understanding of the relationship between reading and the development of refractive errors. To study whether corneal distortions after reading significantly differ between refractive error groups, corneal aberrations were measured before and after a period of reading, for a group of ten young progressing myopes and a group of ten young stable emmetropes. The major difference between the two groups was the location and magnitude of the corneal distortions, which had a significantly larger effect on central corneal optics in the myopic group compared to the emmetropic group. A significantly smaller palpebral aperture for the myopic group in the reading gaze position was the cause of this difference. The experiments described in this thesis have shown that numerous corneal characteristics can change due to eyelid forces during near work. The eye was shown to undergo a small cyclotorsion during higher levels of accommodation. There was a shift in direction of against the rule astigmatism of the cornea following reading and a change was found for primary vertical coma and trefoil. The changes in corneal shape following reading appear to be different in myope versus emmetropic refractive error groups. These findings are important for our understanding of the stability of the refractive error of the eye and could have important implications for refractive error development.

The change in the hydrological cycle due to climate change entails more frequent and intense rainfall. As a result, urban water systems will be disproportionately affected by the climate change, especially in such urban areas as Las Vegas, which concentrates its population, infrastructure, and economic activity. Understanding the proper management of urban stormwater in the changing climate is becoming a critical concern to the water resources managers. Proper design and management of stormwater facilities are needed to attenuate the severe effects of extreme rainfall events. In an effort to develop better management techniques and understanding the probable future scenario, this study used the high-resolution climate model data conjunction with advanced statistical methods and computer simulation. Las Vegas Valley which has unique climatic condition and is surrounded by the mountains in every direction was chosen for the study. The North American Regional Climate Change Assessment Program is developing multiple high-resolution projected-climate data from different combinations of regional climate models and global climate models. First, the future design depths was calculated using generalized extreme value method with the aid of L-moment regionalization technique. The projected climate change was incorporated into the model at the 100 year return period with 6h duration depths. Calculation showed that, the projection from different sets of climate model combinations varied substantially. Gridded reanalysis data were used to assess the performance of the climate models. This study used an existing Hydrologic Engineering Center’s Hydrological Modeling System (HEC-HMS) model and Storm Water Management Model (SWMM) developed by the Environmental Protection Agency (EPA) were implemented in the hydrological simulation. Hydrological simulation using HEC-HMS showed exceedances of existing stormwater facilities that were designed under the assumption of stationarity design depth. Low Impact Developments such as permeable pavement and green roof were found to be effective in the attenuation of climate change induced excess surface runoff. The primary purpose of this study is understanding of proper designing, planning and management of the urban stormwater system in the predicted climate scenarios.