Este trabalho apresenta um estudo sobre a utilização da Transformada de Wavelet para o realce de contraste de estruturas de interesse mamográfico (nódulos e microcalcificações), visando o desenvolvimento de um sistema computacional para auxílio ao diagnóstico do câncer de mama. Este sistema compõe-se basicamente de 3 etapas: pré-processamento, detecção e classificação, sendo que este trabalho insere-se na etapa de pré-processamento. Nessa etapa as estruturas de interesse foram realçadas de forma a propiciar um melhor desempenho nas etapas seguintes. No trabalho aqui descrito as estruturas de interesse contidas em mamogramas digitalizados foram realçadas pela Transformada de Wavelet através de dois métodos distintos: Ganho Adaptativo Multiescala e Reconstrução Parcial. Na técnica Ganho Adaptativo Multiescala a imagem é decomposta em sub-imagens pela Transformada de Wavelet. As sub-imagens decompostas são realçadas por uma função não linear e então a imagem realçada é obtida pela aplicação da Transformada Inversa de Wavelet. Na técnica Reconstrução Parcial, a imagem é decomposta em sub-imagens pela Transformada de Wavelet. São selecionadas as sub-imagens onde as estruturas de interesse são mais representativas e então a imagem realçada é obtida pela aplicação da Transformada Inversa de Wavelet à estas sub-imagens. A avaliação dos resultados é feita por um quantificador da melhoria de contraste das imagens e pela análise do histograma. / In the present work, we presents a study about using of Wavelet Transform in contrast enhancement of mammographic features(mass and microcalcifications), as part of developing a Computerized System proposed to aid in diagnosis of breast cancer. This system is composed of 3 main stages: early processing, detection and classification, and this work is into the early processing stage. In this stage structures of interest have been enhanced to improve performance at following stages. In this work, mammographic features have been enhanced by application of 2 distinct methods using Wavelet Transform: Multiscale Adaptive Gain and Partial Reconstruction. Multiscale Adaptive Gain decomposes the original image in subimages by application of Wavelet Transform. These subimages are enhanced throw the application of a non linear function and the enhanced image is obtained by application of Inverse Wavelet Transform. Partial Reconstruction decomposes the original image in subimages by application of Wavelet Transform. The subimages more representative for mammographic features are selected and the enhanced image is obtained by application of lnverse Wavelet Transform in these subimages. The final results obtained are evaluated by a Contrast Improvement Quantifier and histogram analysis.

Este trabalho apresenta uma nova aplicação de wavelets em sistemas de espalhamento espectral. A aplicação apresentada foi realizada no contexto de detecção multiusuário e pode ser empregada em sistemas de acesso múltiplo por divisão de código. A detecção e a estimação são realizadas através de um algoritmo de inferência bayesiana, empregando o Amostrador de Gibbs e utilizando as wavelets para melhorar a relação sinal ruído na recepção. Na transmissão, um código específico, baseado na wavelet de Haar, modula o sinal contendo a informação preparando-o contra alterações que possam ocorrer devido a características intrínsecas do canal de ruído branco, gaussiano e aditivo. Esta técnica foi aqui denominada de modulação wavelet. Na recepção, uma transformada wavelet de Haar é aplicada ao sinal, gerando coeficientes wavelets. A ortogonalidade da wavelet de Haar torna possível executar um procedimento de cancelamento de ruído, em inglês denoising, que melhora a qualidade da recepção. Para validar a abordagem apresentada neste trabalho, foi executada uma série de simulações. Os resultados mostraram um melhor desempenho sistêmico com o emprego de wavelets. Nas simulações foram observados vários parâmetros de sistema, dentre eles e amplitude do sinal, a taxa de erro de bits (BER, bit error rate) e a variância do sinal. Os resultados foram bastante satisfatórios, indicando uma redução do ruído da ordem de 3dB em relação ao valor sem cancelamento de ruído. Este trabalho abre novas oportunidades de pesquisa para o uso de wavelets em sistemas de espalhamento espectral e telecomunicações em geral. / This work presents a new application of wavelets to spread spectrum systems. The presented application was made in the context of multiuser detection and can be used in code-division multiple access systems. Detection and estimation are performed through a Bayesian inference algorithm and, employing the Gibbs sampler and utilizing wavelets to improve the signal-to-noise ratio at the reception. At the transmission, a specific code based on the Haar wavelet modulates the signal containing the information, preparing it to alterations that may occur due to the intrinsic characteristics of the additive Gaussian noise channel. This technique was here denominated wavelet modulation. At the reception, a Haar wavelet transform is applied to the signal, generating wavelet coefficients. The orthogonality of the Haar wavelet makes it possible to perform a noise cancelling process, called denoising, which enhances the receiving quality. To validate the approach presented in this work, a number of simulations were performed. The results showed a better system performance with the use of wavelets. Several parameters were monitored in the simulations, among them the signal amplitude, de BER (bit error rate), and the signal variance. The results were very satisfactory, indicating a noise reduction of approximately 3dB referred to the value without noise cancelling. This work opens new research opportunities for the use of wavelets in spread spectrum systems and Telecommunications in general.

In the field of machine prognostics, vibration analysis is a proven method for detecting and diagnosing bearing faults in rotating machines. One popular method for interpreting vibration signals is envelope demodulation, which allows a technician to clearly identify an impulsive fault source and its severity. However incipient faults -faults in early stages - are masked by in-band noise, which can make the associated impulses difficult to detect and interpret. In this thesis, Wavelet De-Noising (WDN) is implemented after envelope-demodulation to improve accuracy of bearing fault diagnostics. This contrasts the typical approach of de-noising as a preprocessing step. When manually measuring time-domain impulse amplitudes, the algorithm shows varying improvements in Signal-to-Noise Ratio (SNR) relative to background vibrational noise. A frequency-domain measure of SNR agrees with this result. / Includes bibliography. / Thesis (M.S.)--Florida Atlantic University, 2014. / FAU Electronic Theses and Dissertations Collection

Aims: To explore and inform current water immersion recovery practice of high performance athletes; and to compare recovery interventions of 5 minutes cold water immersion, warm water immersion and passive rest, in trained subjects, following intense exercise replicating the demands of game-sports. Methods: Study 1: In a repeated measures design, a measurement approach for use in the evaluation of water immersion efficacy was piloted. The within-day and between-day reliability of surface electromyelography (sEMG), particularly functional wavelet analysis, was evaluated in human lower limb muscles. Functional wavelet analysis provides the opportunity to measure neuromuscular function at the greatest level of detail by differentiating the relative intensity of low and high frequency motor unit recruitment. On 2 consecutive days (Trial 1 & Trial 2), 12 participants performed 3x5 second isometric 80% maximal voluntary contractions (MVC) on a Biodex® dynamometer in each of 150 ankle plantarflexion, 200 knee extension and 200 knee flexion. sEMG was obtained from the medial gastrocnemius (MG), vastis medialis (VM), vastis lateralis (VL) and biceps femoris (BF) muscles. Joint position and force production were controlled. Electrodes remained in situ during each trial. Electrodes were removed upon completion of Trial 1 and replaced in the same position the next day for Trial 2. Simultaneous sEMG metrics for intervals of consistent force production were compared between contractions in Trial 1 and Trial 2 (between-day) and contractions within Trial 2 (within-day). Study 2: 11 trained participants completed the 90 minute Loughborough Intermittent Shuttle Test (LIST). Five minutes of COLD water immersion (8.8 ± 0.30C), WARM water immersion (35.1 ± 1.80C) and REST were compared in a repeated measures randomised cross over design. Recovery was evaluated at 2, 4 and 24 hours post exercise using circulating markers of muscle damage, muscle dynamometry, drop jump and repeated single leg hop performance tests and perceived recovery. Study 3: Current water immersion practice of high performance athletes, practice implications stemming from this study’s findings, and the rationale were explored. In a purposive, theoretical sampling approach of expert consultation, 8 professionals advising internationally competing athletes on water immersion recovery practice were provided with a research brief of this project in advance of a scribed, semi-structured interview. Participants were of Sports Coach, Strength & Conditioning Coach and Sports Physiotherapist professions with a minimum of 5 years’ experience working with internationally competing athletes; and differed in international location and sporting disciplines.

System identification (SI) for constructed structural systems has received a lot of attention with the continuous development of modern technologies. This thesis proposes a new nonlinear time series model for use in system identification (SI) of smart structures. The proposed model is implemented by the integration of a wavelet transform (WT) and nonlinear autoregressive moving average (NARMA) time series model. The approach demonstrates the efficient and accurate nonlinear SI of smart structures subjected to both ambient excitation and high impact load. To demonstrate the effectiveness of the wavelet-based NARMA modeling (WNARMA), smart structures equipped with magnetorheological (MR) dampers are investigated. The simulation results show that the computation of the WNARMA model is faster than that of the NARMA model without sacrificing the modeling accuracy. In addition, the WNARMA model is robust against noise in the data since it inherently has a denoising capacity.

As HPC systems move towards exascale, the discrepancy between computational power and I/O transfer rate is only growing larger. Lossy in situ compression is a promising solution to address this gap, since it alleviates I/O constraints while still enabling traditional post hoc analysis. This dissertation explores the viability of such a solution with respect to a specific kind of compressor — wavelets. We especially examine three aspects of concern regarding the viability of wavelets: 1) information loss after compression, 2) its capability to fit within in situ constraints, and 3) the compressor’s capability to adapt to HPC architectural changes. Findings from this dissertation inform in situ use of wavelet compressors on HPC systems, demonstrate its viabilities, and argue that its viability will only increase as exascale computing becomes a reality.

nÃo hà / This dissertation describes a system for analysis of heart rate variability through metrics on time and frequency domain and by non-linear methodology, which is initiated by the process of segmentation of the QRS complex of the electrocardiogram signal. The motivation for this work is the analysis of the influence from the algorithms of beat segmentation and selection of valid cardiac cycles for the variability analysis, which were developed in the research process, over the computation of the metrics. After determining the intervals between QRS complexes (RR), the cardiac cycles with ectopic beats, resultant of arrhythmic events or detection fails (false-positive or false negative) are excluded. Then, the available metrics of heart rate variability found on literature are computed over the time series of intervals between normal beats (NN): on time domain (statistical and geometrical methods), on frequency domain (VLF - Very Low Frequency, LF - Low Frequency and HF â High Frequency) and by non-linear methodology (Poincarà plot). The QRS detection and segmentation results are validated through simulation tests over exams from Arrhythmia Database and QT database of the MIT-BIH database. The manual annotations of the QRS fiducial points and QRS onset and offset are compared with the automatic detections. The results related to heart rate variability metrics are validated through the manual selection of beats, and consequently of the intervals between them, pertaining to exams selected from Arrhythmia Database and the computation of the referred metrics over them, comparing with those ones automatically generated by the proposed method. The system, which provides averages of positive predictivity as 99.35% and sensitivity as 99.02%, and averages of deviations between automatic and manual analysis of heart rate variability metrics varying from 0.05% to 5.24%, can be carried into several platforms, making possible its production in commercial scale. / Esta dissertaÃÃo descreve um sistema de analise da variabilidade da frequÃcia cardÃaca atravÃs de mÃtricas no domÃnio do tempo, da freqÃÃncia e por mÃtodo nÃo-linear a partir do processo de segmentaÃÃo do complexo QRS do sinal eletrocardiograma. O trabalho à motivado pela influÃncia dos algoritmos de segmentaÃÃo de batimentos e de seleÃÃo dos ciclos cardÃacos vÃlidos para anÃlise da variabilidade, esenvolvidos para este fim, na determinaÃÃo das mÃtricas de interesse. ApÃs a determinaÃÃo dos intervalos entre os complexos QRS (RR), sÃo excluÃdos os ciclos cardÃacos com batimentos ectÃpicos, resultantes de arritmia ou de falhas de detecÃÃo (falso-positivo ou falso-negativo). Em seguida, sobre a sÃrie temporal de intervalos entre batimentos normais NN sÃo calculadas as mÃtricas de variabilidade da freqÃÃncia cardÃaca disponÃveis na literatura: no domÃnio do tempo (mÃtodos estatÃsticos e geomÃtricos), no domÃnio da freqÃÃncia (componentes VLF - Very Low Frequency, LF - Low Frequency e HF - High Frequency) e por mÃtodo nÃo-linear (mapa de PoincarÃ). Os resultados de deteccÃo e segmentacÃo do QRS sÃo validados atravÃs de testes experimentais sobre exames das bases Arrhythmia Database e QT database do MIT-BIH, em que as marcaÃÃes manuais dos picos e das bordas dos batimentos sÃo comparadas com as detecÃÃes automÃticas. Os resultados obtidos quanto as mÃtricas de variabilidade sÃo validados atravÃs da seleÃÃo manual de batimentos e, por conseguinte, dos intervalos entre os mesmos, a partir de exames selecionados da base Arrhythmia Database por cardiologistas do Hospital UniversitÃrio Walter Cantidio (HUWC), e do cÃlculo das referidas mÃtricas, comparando-se com aquelas geradas automaticamente pelo mÃtodo proposto. O sistema, que apresenta taxas mÃdias de99,35% de preditividade positiva e 99,02% de sensibilidade, para detecÃÃo do QRS, e mÃdias de erros entre a anÃlise automÃtica e a anÃlise manual das mÃtricas de variabilidade variando entre 0,05% e 5,24%, pode ser embutido em diversas plataformas, viabilizando sua producÃo em escala comercial.

Sabe-se que folgas e impactos de partes internas reduzem o desempenho da transmissão, do sistema de controle e a qualidade global do mecanismo. Folgas causam instabilidade, mas impactos também geram ruídos estruturais que perturbam medições e causam ruído acústico que reduz o conforto. Sinais de movimento são frequentemente usados para realimentação e podem ser perturbados por ruídos de impacto. Para avaliar este problema, um procedimento capaz de detectar e quantificar a severidade e a frequência de ocorrência de impactos é proposto. Desta forma, o desempenho de diferentes sistemas em relação à sua robustez ao impacto pode ser comparado. Sinais de vibração na estrutura do sistema são medidos e o sinal é processado digitalmente e decomposto em múltiplas escalas pela Transformada de Wavelet. As wavelets Harmônicas e de Daubechies são comparadas. Escalas que correspondem aos níveis de detalhamento característicos dos impactos são extraídas do sinal de vibração e a severidade do impacto é avaliada. Após o processamento do sinal e a retirada do ruído, o sinal original é disponibilizado para posterior processamento e controle. Foi estudado o ruído de impacto em caixas de direção conhecido como knock noise. Ao final, tentou-se comparar a técnica de Transformada de Wavelet com a técnica de Emissão Acústica. Na sessão Transformada de Wavelets é feito um breve estudo de alguns conceitos necessários para o entendimento de wavelets e é explicada a técnica utilizada para a detecção dos ruídos de impacto. Na sessão Emissão Acústica é explicada a técnica utilizada para comparar com a técnica de wavelets. Na sessão Direção Hidráulica é explicado o mecanismo de direção estudado. Na sessão Resultados, são mostrados os sinais de ruídos de impacto detectados pela Transformada de Wavelet e Emissão Acústica e o método adotado para mensurar a severidade dos impactos e a frequência de ocorrência dos mesmos. / It is well known that backlash and impact of inner parts reduce the performance of transmission, control systems and overall quality of mechanism. Clearance causes instability, but impact also generates structural borne noise that disrupts measurements and cause acoustic noise that reduces the comfort. Motion signals are often used for feedback and can be disrupted by impact noise. In order to overcome this problem a procedure able to detect and quantify the severity and rate of occurrence of impacts is proposed. Thence performance of different systems in regard to their robustness to impact can be compared. A procedure to detect and evaluate impacts of inner parts of mechanical mechanism is presented. Therefore, vibration signals on the structure of the system are measured and the signal is digitally processed and decomposed into multiscales by the Wavelet transform. Both Daubechies and Harmonic Wavelet are compared. Scales that correspond to the impact response are extracted from the overall vibration signal, and the impact severity is rated. After signal processing and de-noising, the original signal without the impact is also available for further processing and control. The analysis technique was applied to study the impact noise in steering boxes known as knock noise. Finally, it was tried to compare the technique of wavelet transform with the Acoustic Emission. First, a brief study is presented of some concepts necessary to understand wavelets and Wavelet Transform technique is explained and the approach used for detection of impact noise is presented. In the following section the technique of Acoustic Emission is explained and how it was compared to the wavelet technique. The Hydraulic Steering System is explained in the next section. In the final section the Results of the study are show, the signal with impact noise, the impact detected by the Wavelet Transform and Acoustic Emission and the method used to measure the severity of impacts and to measure the frequency of their occurrence.

Este trabalho apresenta uma nova aplicação de wavelets em sistemas de espalhamento espectral. A aplicação apresentada foi realizada no contexto de detecção multiusuário e pode ser empregada em sistemas de acesso múltiplo por divisão de código. A detecção e a estimação são realizadas através de um algoritmo de inferência bayesiana, empregando o Amostrador de Gibbs e utilizando as wavelets para melhorar a relação sinal ruído na recepção. Na transmissão, um código específico, baseado na wavelet de Haar, modula o sinal contendo a informação preparando-o contra alterações que possam ocorrer devido a características intrínsecas do canal de ruído branco, gaussiano e aditivo. Esta técnica foi aqui denominada de modulação wavelet. Na recepção, uma transformada wavelet de Haar é aplicada ao sinal, gerando coeficientes wavelets. A ortogonalidade da wavelet de Haar torna possível executar um procedimento de cancelamento de ruído, em inglês denoising, que melhora a qualidade da recepção. Para validar a abordagem apresentada neste trabalho, foi executada uma série de simulações. Os resultados mostraram um melhor desempenho sistêmico com o emprego de wavelets. Nas simulações foram observados vários parâmetros de sistema, dentre eles e amplitude do sinal, a taxa de erro de bits (BER, bit error rate) e a variância do sinal. Os resultados foram bastante satisfatórios, indicando uma redução do ruído da ordem de 3dB em relação ao valor sem cancelamento de ruído. Este trabalho abre novas oportunidades de pesquisa para o uso de wavelets em sistemas de espalhamento espectral e telecomunicações em geral. / This work presents a new application of wavelets to spread spectrum systems. The presented application was made in the context of multiuser detection and can be used in code-division multiple access systems. Detection and estimation are performed through a Bayesian inference algorithm and, employing the Gibbs sampler and utilizing wavelets to improve the signal-to-noise ratio at the reception. At the transmission, a specific code based on the Haar wavelet modulates the signal containing the information, preparing it to alterations that may occur due to the intrinsic characteristics of the additive Gaussian noise channel. This technique was here denominated wavelet modulation. At the reception, a Haar wavelet transform is applied to the signal, generating wavelet coefficients. The orthogonality of the Haar wavelet makes it possible to perform a noise cancelling process, called denoising, which enhances the receiving quality. To validate the approach presented in this work, a number of simulations were performed. The results showed a better system performance with the use of wavelets. Several parameters were monitored in the simulations, among them the signal amplitude, de BER (bit error rate), and the signal variance. The results were very satisfactory, indicating a noise reduction of approximately 3dB referred to the value without noise cancelling. This work opens new research opportunities for the use of wavelets in spread spectrum systems and Telecommunications in general.

Cette thèse en algorithmique du texte étudie la compression, l'indexation et les requêtes sur un texte annoté. Un texte annoté est un texte sur lequel nous ajoutons des informations. Ce peut être par exemple une recombinaison V(D)J, un marqueur de globules blancs, où le texte est une séquence ADN et les annotations sont des noms de gènes. Le système immunitaire d'une personne se représente par un ensemble de recombinaisons V(D)J. Avec le séquençage à haut débit, on peut avoir accès à des millions de recombinaisons V(D)J qui sont stockées et doivent pouvoir être retrouvées et comparées rapidement.La première contribution de ce manuscrit est une méthode de compression d'un texte annoté qui repose sur le principe du stockage par références. Le texte est découpé en facteurs pointant vers les séquences annotées déjà connues. La seconde contribution propose deux index pour un texte annoté. Ils utilisent une transformée de Burrows-Wheeler indexant le texte ainsi qu'un Wavelet Tree stockant les annotations. Ces index permettent des requêtes efficaces sur le texte, les annotations ou les deux. Nous souhaitons à terme utiliser l'un de ces index pour indexer des recombinaisons V(D)J obtenues dans des services d'hématologie lors du diagnostic et du suivi de patients atteints de leucémie. / This thesis in text algorithm studies the compression, indexation and querying on a labeled text. A labeled text is a text to which we add information. For example: a V(D)J recombination, a marker for lymphocytes, where the text is a DNA sequence and the labels are the genes' names. A person's immune system can be represented with a set of V(D)J recombinations. With high-throughput sequencing, we have access to millions of V(D)J recombinations which are stored and need to be recovered and compared quickly.The first contribution of this manuscript is a compression method for a labeled text which uses the concept of storage by references. The text is divided into sections which point to pre-established labeled sequences. The second contribution offers two indexes for a labeled text. Both use a Burrows-Wheeler transform to index the text and a Wavelet Tree to index the labels. These indexes allow efficient queries on text, labels or both. We would like to use one of these indexes on V(D)J recombinations which are obtained with hematology services from the diagnostic or follow-up of patients suffering from leukemia.