Before launching into ultrasound research, it is important to recall that the ultimate goal is to provide the clinician with the best possible information needed to make an accurate diagnosis. Ultrasound images are inherently affected by speckle noise, which is due to image formation under coherent waves. Thus, it appears to be sensible to reduce speckle artifacts before performing image analysis, provided that image texture that might distinguish one tissue from another is preserved. The main goal of this thesis was the development of novel speckle suppression methods from medical ultrasound images in the multiscale wavelet domain. We started by showing, through extensive modeling, that the subband decompositions of ultrasound images have significantly non-Gaussian statistics that are best described by families of heavy-tailed distributions such as the alpha-stable. Then, we developed Bayesian estimators that exploit these statistics. We used the alpha-stable model to design both the minimum absolute error (MAE) and the maximum a posteriori (MAP) estimators for alpha-stable signal mixed in Gaussian noise. The resulting noise-removal processors perform non-linear operations on the data and we relate this non-linearity to the degree of non-Gaussianity of the data. We compared our techniques to classical speckle filters and current state-of-the-art soft and hard thresholding methods applied on actual ultrasound medical images and we quantified the achieved performance improvement. Finally, we have shown that our proposed processors can find application in other areas of interest as well, and we have chosen as an illustrative example the case of synthetic aperture radar (SAR) images. / Ο απώτερος σκοπός της έρευνας που παρουσιάζεται σε αυτή τη διδακτορική διατριβή είναι η διάθεση στην κοινότητα των κλινικών επιστημόνων μεθόδων οι οποίες να παρέχουν την καλύτερη δυνατή πληροφορία για να γίνει μια σωστή ιατρική διάγνωση. Οι εικόνες υπερήχων προσβάλλονται ενδογενώς από θόρυβο, ο οποίος οφείλεται στην διαδικασία δημιουργίας των εικόνων μέσω ακτινοβολίας που χρησιμοποιεί σύμφωνες κυματομορφές. Είναι σημαντικό πριν τη διαδικασία ανάλυσης της εικόνας να γίνεται απάλειψη του θορύβου με κατάλληλο τρόπο ώστε να διατηρείται η υφή της εικόνας, η οποία βοηθά στην διάκριση ενός ιστού από έναν άλλο. Κύριος στόχος της διατριβής αυτής υπήρξε η ανάπτυξη νέων μεθόδων καταστολής του θορύβου σε ιατρικές εικόνες υπερήχων στο πεδίο του μετασχηματισμού κυματιδίων. Αρχικά αποδείξαμε μέσω εκτενών πειραμάτων μοντελοποίησης, ότι τα δεδομένα που προκύπτουν από τον διαχωρισμό των εικόνων υπερήχων σε υποπεριοχές συχνοτήτων περιγράφονται επακριβώς από μη-γκαουσιανές κατανομές βαρέων ουρών, όπως είναι οι άλφα-ευσταθείς κατανομές. Κατόπιν, αναπτύξαμε Μπεϋζιανούς εκτιμητές που αξιοποιούν αυτή τη στατιστική περιγραφή. Πιο συγκεκριμένα, χρησιμοποιήσαμε το άλφα-ευσταθές μοντέλο για να σχεδιάσουμε εκτιμητές ελάχιστου απόλυτου λάθος και μέγιστης εκ των υστέρων πιθανότητας για άλφα-ευσταθή σήματα αναμεμειγμένα με μη-γκαουσιανό θόρυβο. Οι επεξεργαστές αφαίρεσης θορύβου που προέκυψαν επενεργούν κατά μη-γραμμικό τρόπο στα δεδομένα και συσχετίζουν με βέλτιστο τρόπο αυτή την μη-γραμμικότητα με τον βαθμό κατά τον οποίο τα δεδομένα είναι μη-γκαουσιανά. Συγκρίναμε τις τεχνικές μας με κλασσικά φίλτρα καθώς και σύγχρονες μεθόδους αυστηρού και μαλακού κατωφλίου εφαρμόζοντάς τες σε πραγματικές ιατρικές εικόνες υπερήχων και ποσοτικοποιήσαμε την απόδοση που επιτεύχθηκε. Τέλος, δείξαμε ότι οι προτεινόμενοι επεξεργαστές μπορούν να βρουν εφαρμογές και σε άλλες περιοχές ενδιαφέροντος και επιλέξαμε ως ενδεικτικό παράδειγμα την περίπτωση εικόνων ραντάρ συνθετικής διατομής.

La problématique de cette thèse est l'étude de la localisation d'un condensat de Bose-Einstein confiné harmoniquement et quasi-1D à travers lequel différents potentiels désordonnés sont transportés. Cette problématique qui se veut pleinement pertinente pour les expérimentalistes est à priori difficile à traiter. Cela est dû au caractère non-linéaire, inhomogène et hors-équilibre du système. De ce fait, la plage des vitesses du désordre est limitée d'une part par la vitesse critique de superfluidité et d'autre part par la configuration inhomogène du système. Des notions habituelles de localisation telles que transmission ou exposant de Lyapunov ne sont plus applicables. Donc, il a fallu apporter une nouvelle mesure de localisation pour notre problématique: le ratio du déplacement du centre de masse du condensat au déplacement du désordre qu'on a identifié à la fraction d'atomes localisés. De plus, nous avons des corrélations dans le désordre qui introduisent l'effet d'un comportement non-monotone de l'efficacité de la localisation du potentiel désordonné en fonction de l'énergie. Ainsi, les corrélations peuvent être un moyen pour mettre en évidence la nature quantique de la localisation. Chose que nous avont fait dans un premier temps avec du désordre de type Modèle d'Edwards et dans une seconde partie avec du désordre de type speckle, qu'on nomme le Random Dimer speckle. Pour ce deuxième cas, nous avons proposé une procédure pour contrôler les corrélations et introduire un pic de localisation dans une certaine région énergétique. Cette configuration pourrait être vérifié dans les expériences à l'aide d'un modulateur spatial de lumière.

Background: Aortic stenosis (AS) is the most common heart valve disease in Europe and North America. Age-related calcification of the valve is the commonest cause of acquired AS, especially in patients older than 70 years.Conventional surgical aortic valve replacement (SAVR) and the novel, minimally invasive transcatheter aortic valve implantation (TAVI), effectively preserve left ventricular (LV) function, relieve symptoms and improve survival in patients with severe symptomatic AS. However, patients with impaired LV function may carry significant operative risk, and long recovery time. In addition, such patients might have other comorbidities, and hence adding another challenge. Thus evaluation of ventricular function before and after AVR, as well as critical evaluation of TAVI patients should contribute to better clinical outcome. Methods: We studied LV function by conventional echocardiography before and after SAVR in the following groups; (I) 86 patients (aged 71±10 years) with severe AS and LV dysfunction; (II) 112 consecutive elderly AS patients (aged 77±2 years) and compared them with 72 younger patients (aged 60±1 years); (III)66 patients (age 70±2 years, 53 male) who underwent AVR for severe AS with concurrent LV dysfunction; (IV) 89 consecutive patients with symptomatic severeAS who underwent successful TAVI, 45 of whom received trans-apical TAVI (TA)(age 80.8±4.9 year, 26 male) and 44 trans-femoral TAVI (TF) (age 82.9±5.8 year,22 male).The conventional echocardiographic measurements were made according to the guidelines. Severe AS was identified by aortic valve mean pressure gradient >40mmHg or valve area <1.0 cm2. LV systolic dysfunction was identified as ejection fraction (EF) <50%. LV long-axis function was presented by mitral annular plane systolic excursion ( MAPSE ) at lateral wall and septal wall, which were measured from apical four-chamber view. Also from the same view, LV septal and lateral wall deformation using STE as well as global longitudinal systolic strain. The LV systolic twist as the net difference between apical rotation and basal rotation was measured from the parasternal apical and basal short-axis views in the TAVI patients. Results: Study I: In the low flow and high gradient group, operative (30-day) mortality was 10%, and peri-operative mortality was associated with lower mean LVEF, higher mitral E:A ratio, peak systolic pulmonary artery pressure (PSPAP), and higher serum creatinine (all p<0.001), NYHA class III–IV, concomitant coronary artery bypass graft (CABG), urgent surgery, and longer bypass-time (all p< 0.05). Mortality at 4 years was 17%. Univariate predictors of 4-year mortality were: lower EF (p<0.001), presence of restrictive LV filling (p<0.001), raised PSPAP (p<0.001) and CABG (p=0.037). However, only EF<40 % (p=0.03), the presence of restrictive LV filling (p=0.033) and raised PSPAP (p<0.01)independently predicted mortality in this group.Study II: Elderly patients had higher NYHA class, more frequent atrial fibrillation (AF), coronary artery disease (CAD), emergency operation and use of bioprosthetic valves. They also had shorter E-wave deceleration time (DT) and larger left atria (LA) (p<0.05 for all). 30-day mortality was 12% vs 4 % (Log Rank x2=3.02, p=0.08) and long term mortality was 18% vs 7% (Log Rank x2=4.38,p=0.04) in the two groups, respectively. Age was not related to mortality after adjustment for other variables. Among all variables, anemia (OR 4.20, CI:1.02–6.86, p=0.04), cardiopulmonary bypass (CPB) time (OR 1.02, CI 1.01–1.04,p<0.01), significant patient prosthesis mismatch (PPM) (OR 5.43, CI 1.04–18.40,p<0.05) were associated with 30-day mortality in elderly patients. Their long-term mortality was related to CBP time (OR 1.02, CI 1.00–1.05, p=0.04),PPM (OR 4.64, CI 1.33–16.11, p=0.02) and raised LA pressure: DT (OR 0.94, CI0.84–0.99, p=0.03) and pulmonary artery systolic pressure (PASP) (OR 1.12, CI1.03–1.19, p<0.001).STUDY III: Following SAVR peak aortic pressure gradient (AOPG) decreased and indexed valve area increased (64±3 to 19±1 mmHg and 0.30±0.01 to 0.89±0.03 cm2/m2, p<0.001 for both). LVEF increased (from 45±1 to 54±2%;p<0.001), LV end diastolic and end-systolic dimensions fell (LVEDD index: from 33±1 to 30±1 mm/m2; and LVESD index: from 27±1 to 20±1 mm/m2; (p<0.01 forboth). LV diastolic dysfunction improved as evidenced by the fall in E/A ratio (from 2.6±0.2 to 1.9±0.4) and prolongation of total filling time; (from 29.2±0.6 to31.4±0.5 s/min, p=0.01 for both). Among all echocardiographic variables, LV dimensions (LVEDD index, OR 0.70, CI 0.52–0.97, p<0.05; LVESD index, OR 0.57, CI 0.40–0.85, p=0.005) were the two independent predictors of post-operative LV functional recovery on multivariate analysis. A cut-off value ofpre-operative LVESD index<=27.5 mm/m2 was 85% sensitive and 72% specific inpredicting intermediate-term recovery of LV function after AVR (AUC, 0.72, p=0.002). STUDY IV: Before TAVI, there was no difference between the two patient groups in gender, age, body surface area (BSA) and baseline LV function. However, left ventricular mass index (LVMi), left atrial volume index (LAVi) and tricuspid regurgitation pressure drop (TRPdrop) were increased in the TA group (p<0.05).One week after TAVI, aortic pressure gradient (AOPG) markedly dropped in thetwo groups (both p<0.001), LVEDD index and LVESD index fell but EF andmyocardial strain remained unchanged. Overall cavity twist reduced (p<0.048).Significant LVESD index reduction was only seen in TF group (p=0.02) with a slight increase in LVEF (p=0.04). Lateral MAPSE increased only in the TF group(p=0.02). LV longitudinal systolic strain remained unchanged in TA patients while apical lateral strain increased in TF group. LV apical rotation fell in the two groups but basal rotation increased only in the TA patients (p=0.02). LAVi reduced in bothgroups and to a greater extent in TF TAVI (p=0.006), as did TRPdrop (p<0.001). Conclusion: SAVR and TAVI are two effective treatments for severe AS patients.The severity of pre-operative systolic and diastolic LV dysfunction is the major predictor of mortality following SAVR for low-flow and high gradient AS.Peri-operative AVR survival is encouraging in the elderly. Long term mortality in the elderly is related to PPM, LV diastolic dysfunction and secondary pulmonary hypertension. LV functional recovery was evident in most patients with LV dysfunction after SAVR. A lower prevalence of LV functional recovery in patients with large pre-operative LVESD index might signify the loss of contractile reserveand thus predict post-operative functional recovery. TAVI results in significant early improvement of segmental and overall ventricular function, particularly in patients receiving the trans-femoral approach. The delayed recovery of the trans-apical TAVI group, we studied, might reflect worse pre-procedural diastolic cavity function.

L'utilisation des techniques de mesure optiques impose un choix, souvent difficile, entre la résolution temporelle et la résolution spatiale. L'interferométrie Speckle (IS) fournit des mesures des champs d'amplitudes de vibration de grande résolution spatiale, mais la résolution temporelle des mesures est réduite par les cadences d'acquisition des détecteurs. La thèse est dédiée au développement d'un système interférométrique adapté à la mesure des vibrations et des phénomènes dynamiques des structures, basé sur les principes de l'IS, mais ayant une résolution temporelle accrue. Le système a été testé à des cadences de plus de 26000 images/s pour les cas suivants : sollicitation harmonique mono-fréquence ou par une forme d'onde arbitraire ; sollicitation dynamique par une séquence d'échelons de force ; choc mécanique. Les essais ont démontré que le système est capable de dépasser les systèmes classiques d'IS à laser continu en ce qui concerne la mesure des phénomènes dynamiques.

Background：Aortic stenosis (AS) is the commonest valve disease in the West. Aortic valve replacement (AVR) remains the only available management for AS and results in improved symptoms and recovery of ventricular functions. In addition, it is well known that AVR results in disruption of LV function mainly in the form of reversal of septal motion as well as depression of right ventricular (RV) systolic function. The aim of this thesis was to study, in detail, the early and mid-term response of ventricular function to AVR procedures (surgical and TAVI) as well as post operative patients’ exercise capacity. Methods：We studied LV and RV function by Doppler echocardiography and speckle tracking echocardiography (STE) in the following 4 groups; (1) 30 severe AS patients (age 62±11 years, 19 male) with normal LV ejection fraction (EF) who underwent AVR, (2) 20 severe AS patients (age 79±6 years, 14 male) who underwent TAVI, (3) 30 healthy controls (age 63±11 years, 16 male), (4) 21 healthy controls (age 57±9 years, 14 male) who underwent exercise echocardiography. Results: After one week of TAVI, the septal radial motion and RV tricuspid annulus peak systolic excursion (TAPSE) were not different from before, while surgical AVR had significantly reversed septal radial motion and TAPSE dropped by 70% compared to before. The extent of the reversed septal motion correlated with that of TAPSE (r=0.78, p<0.001) in the patients as a whole after AVR and TAVI (Study I). Compared with controls, the LV twist function was increased in AS patients before and normalized after 6 months of surgical AVR. In controls, the LV twist correlated with LV fractional shortening (r=0.81, p<0.001), a relationship which became weak in patients before (r=0.52, p<0.01) and after AVR (r=0.34, p=ns) (Study II). After 6 months of surgical AVR, the reversed septal radial motion was still significantly lower than before. The septal peak displacement also decreased and its time became prolonged. In contrast, the LV lateral wall peak displacement increased and the time to peak displacement was early. The accentuated lateral wall peak displacement correlated with the septal peak displacement time delay (r=0.60, p<0.001) and septal-lateral time delay (r=0.64, p<0.001) (Study III). In 21 surgical AVR patients who performed exercise echocardiography, the LV function was normal at rest but different from controls with exercise. At peak exercise, oxygen consumption (pVO2) was lower in patients than controls. Although patients could achieve cardiac output (CO) and heart rate (HR) similar to controls at peak exercise, the LV systolic and early diastolic myocardial velocities and strain rate as well as their delta changes were significantly lower than controls. pVO2 correlated with peak exercise LV myocardial function in the patients group only, and the systolic global longitudinal strain rate (GLSRs) at peak exercise was the only independent predictor of pVO2 in multivariate regression analysis (p=0.03) (Study IV). Conclusion: Surgical AVR is an effective treatment for AS patients, but results in reversed septal radial motion and reduced TAPSE. The newly developed TAVI procedure maintains RV function which results in preservation of septal radial motion. In AS, the LV twist function is exaggerated, normalizes after AVR but loses its relationship with basal LV function. While the reversed septal motion results in decreased and delayed septal longitudinal displacement which is compensated for by the accentuated lateral wall displacement and the time early. These patients remain suffering from limited exercise capacity years after AVR.

The definition of optimal selection criteria for maximizing the response rate to Cardiac Resynchronization Therapy (CRT) is still an issue under active debate. Recent clinical approaches propose a classification of patients into classes of mechanisms that could lead to heart failure and study their response to the therapy. In this line of research, the computation of a metric between the motion and deformation patterns of a given subject and well identified classes of CRT responders is considered in this thesis, as the basis of a new strategy to compute patient selection indexes. The thesis proposes first an improved design for the construction of statistical atlases of myocardial motion and deformation, and applies it to the characterization of populations of patients involved in CRT. The added-value of our approach is highlighted in a clinical study, applying the methodology to a large population of patients with a given pattern of dyssynchrony (septal flash) and understanding the link between its correction and CRT response. Finally, we propose a method to extend the analysis to the comparison of individuals to reference populations, either healthy or pathological, using manifold learning techniques to model a disease as progressive deviations from normality along a manifold structure, and demonstrate the potential of our method for inter-subject comparison in CRT patients. / La definición de un criterio óptimo para mejorar la respuesta a la Terapia de Resincronización Cardíaca (TRC) sigue siendo un debate abierto. Estudio clínicos recientemente publicados proponen clasificar pacientes según diversos mecanismos patofisiológicos que pueden inducir insuficiencia cardíaca y estudian su respuesta a la terapia. Siguiendo esta línea de investigación, esta tesis considera el cálculo de una distancia entre los patrones de movimiento y deformación de un individuo y las clases de respondedores a la TRC, siendo la base de una nueva estrategia para calcular índices para seleccionar pacientes. Esta tesis presenta primero un método para construir un atlas estadístico de movimiento y deformación miocárdica, y su aplicación posterior a la caracterización de poblaciones de potenciales candidatos a la TRC. El valor añadido de nuestro método se enfatiza en un estudio clínico, en el cual se aplica la metodología a una gran población de pacientes con un patrón específico de disincronía cardíaca (llamado septal flash), y se relaciona su corrección y la respuesta a la TRC. Finalmente, se extiende el método para comparar individuos a una población de referencia, sana o patológica, usando técnicas de manifold learning para representar una patología como una desviación progresiva de la normalidad, con una estructura no lineal específica, y se demuestra el potencial de nuestro método para comparar entre sí candidatos a la TRC.

There has been widespread interest in corneal biomechanics over recent years, driven largely by the advancements in, and the popularity of refractive surgery techniques and subsequent concerns over their safety. Lately there has been interest into whether crosslinking, which is currently used for the treatment of keratoconus, could be developed as a minimally invasive technique to change the refractive power of the cornea by selectively changing the corneal biomechanics in specific regions to induce a shape change. Successful application of this technique requires a detailed understanding of corneal biomechanics and so far, little is known about the biomechanics of this complex tissue. The current lack of understanding can be mostly attributed to the absence of a suitable measurement technique capable of examining the dynamic behaviour of the cornea under physiological loading conditions. This thesis describes the development of a novel full-field, ex vivo, measurement method incorporating speckle interferometric techniques, to examine the biomechanics of the cornea before and after crosslinking in response to hydrostatic pressure fluctuations representative of those that occur in vivo during the cardiac cycle. The eventual measurement system used for the experiments detailed in this thesis incorporated; an Electronic Speckle Pattern Interferometer (ESPI), a Lateral Shearing Interferometer (LSI) and a fringe projection shape measurement system. The combination of these systems enabled the 3-dimensional components of surface displacement and the 1st derivative of surface displacement to be determined in response to small pressure fluctuations up to 1 mmHg in magnitude. The use of both ESPI and LSI together also enabled the applicability of LSI for measurement of non-flat surfaces to be assessed, and limitations and error sources to be identified throughout this work. To enable the measurement of corneal biomechanics, part of this thesis was concerned with the design of a bespoke loading rig. A chamber was designed that could accommodate tissue of both porcine and human origin. This chamber was linked to a hydraulic loading rig, whereby the cornea could be held at a baseline pressure representative of normal intraocular pressure and small pressure variations could be introduced by the automated vertical movement of the reservoir supplying the chamber. Experiments were conducted on a range of non-biological samples with both flat and curved surface topography, and both uniform and non-uniform mechanical properties, to determine if the measurement configuration was giving the expected measurement data and the loading rig was stable and repeatable. Following experiments on non-biological samples, a range of experiments were conducted on porcine corneas to develop a suitable testing methodology and address some of the challenges associated with corneal measurement, including transparency and hydration instability. During these investigations, a suitable surface coating was identified to generate an adequate return signal from the corneal surface, while not interfering with the response. Alongside this, the natural variation in the response of the cornea was investigated over the total experimental time, and a range of data was presented on corneas before and after crosslinking, which confirmed the suitability of the measurement methods for the assessment of crosslinking. Ultimately, a small sample size of six human corneas were investigated before and after crosslinking in specific topographic locations. From the experiments on human and porcine corneas, full-field maps of surface deformation have been presented, and a compliant region incorporating the peripheral and limbal areas has been identified as being fundamental to the response of the cornea to small pressure fluctuations. In addition to this, the regional effects of crosslinking in four different topographic locations on corneal biomechanics have been evaluated. From this, it has been demonstrated that crosslinking in specific regions in isolation can influence the way the cornea deforms to physiological-scale fluctuations in hydrostatic pressure and this could have implications for refractive correction.

In this work, it was studied theoretically and experimentally biospeckle technique for analysis of rough surface with application to the investigation of biological activity. The theoretical study was conducted using light scattering theory, in particular a first order perturbation method, seeking to obtain the scattering cross-section per unit area of rough surface. Once given the cross-section, it was possible to obtain the distribution of intensities in a certain region of space as a function of time, allowing the comparison of different light distributions in different instants. The experimental investigation was made from the comparison of speckle patterns recorded (by a CCD camera) at different time instants produced from scattering of coherent light at 635 nm from a freshly harvested plant leaf. The parameter of comparison used was the Pearson correlation coefficient. The experimental result obtained from the correlation coefficient versus time showed a temporal evolution type exponential decay, indicating a degradation behavior of the leaf chemically active and regulated by equations rate. The theoretical result of the correlation function of time, obtained by simulation, also showed a behavior type exponential decay indicating a good agreement with experimental results. / Nesta dissertação, foi estudada teórica e experimentalmente a técnica de biospeckle para análise de superfície rugosa com aplicação na investigação de atividade biológica. O estudo teórico foi feito usando teoria de espalhamento da luz, em particular usando método perturbativo de primeira ordem, buscando obter a seção de choque de radiação espalhada por unidade de área da superfície rugosa. Uma vez determinada a seção de choque, foi possível obter a distribuição de intensidades numa certa região do espaço, em função do tempo, permitindo a comparação das diferentes distribuições de luz em instantes distintos. A investigação experimental foi feita a partir da comparação de padrões de speckles gravados (por uma câmera CCD) em diferentes instantes de tempo originados do espalhamento de luz coerente em 635 nm, a partir de uma folha vegetal recém colhida. O parâmetro de comparação usado foi o coeficiente de correlação de Pearson. O resultado experimental obtido do coeficiente de correlação em função do tempo mostrou uma evolução temporal do tipo exponencial decrescente, indicando um comportamento de degradação da folha quimicamente ativo e regido por equações de taxas. O resultado teórico do coeficiente de correlação em função do tempo, obtido por meio de simulação, também evidenciou o comportamento exponencial decrescente indicando uma boa concordância com o resultado experimental.

Made available in DSpace on 2014-06-12T15:59:32Z (GMT). No. of bitstreams: 2 arquivo5160_1.pdf: 4852221 bytes, checksum: 00af0b636289f3973abc1f5ce9bce923 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2004 / O ru´ıdo speckle aparece em imagens com iluminac¸ ao coerente, tais como, sonar, laser, ultra-sonografia e radar de abertura sint´etica (Synthentic Aperture Radar SAR). Existem diversas t´ecnicas para reduc¸ ao de ru´ıdo speckle. Este trabalho propoem uma metodologia que permite a utilizac¸ ao destas t´ecnicas a fim de produzir uma imagem filtrada. Esta imagem filtrada ´e obtida combinando pontualmente de forma linear as t´ecnicas j ´a dispon´ıveis. Os coeficientes da combinac¸ ao linear s ao encontrados tal que satisfac¸am aproximadamente as expectativas do usu´ario. A t´ecnica ´e denominada Filtragem Interagente, desde que as expectativas do usu´ario s ao expressas atrav´es das especificac¸ oes das regi oes de interesse e dos valores desejados associados a cada regi ao de interesse. A t´ecnica ´e implementada na linguagem IDL e integrada ao ENVI, uma plataforma de processamento de imagens. A proposta ´e avaliada em imagens sint´etica e real SAR utilizando duas medidas de qualidade, o erro m´edio quadr ´atico normalizado (NMSE - Normalized Mean Square Error) e o n ´umero equivalente de visadas (ENL - Equivalent Number of Looks), sob o crit´erio da relac¸ ao sinal-ru´ıdo (SNR Signal-to-Noise Ratio). Os resultados quantitativos e qualitativos obtidos s ao superiores aos obtidos com filtros cl ´assicos redutores de ru´ıdo speckle e as imagens obtidas satisfazem as expectativas do usu´ario.

L’évolution du champ de localisation des déformations plastiques des métaux ductiles a été suivie par interférométrie de granularité laser (ESPI) au cours d’une sollicitation de traction uniaxiale. Il a été montré que le modèle mathématique utilisé, croisement de deux bandes rectilignes, décrit bien l’évolution du champ de vitesse de déformation depuis la striction diffuse jusqu'à rupture. Les caractéristiques physiques de la localisation, telles que la largeur des bandes, leur orientation et leur vitesse de déformation maximale ont été identifiées quantitativement. Leur évolution a été suivie et analysée sous plusieurs facteurs d’influence, qui sont à la fois microscopiques, tels que la taille de grains et la structure cristalline, et macroscopiques, comme la géométrie de l’éprouvette et la direction de son prélèvement par rapport à la direction de laminage, et la vitesse de traction. Il a été trouvé que les bandes de localisation rétrécissent au cours de la striction et leur orientation évolue également. Ces évolutions, le mode de striction et le moment de transition entre la striction diffuse et localisée sont influencés différemment par ces facteurs / The evolution of the plastic strain localization field of ductile metals was followed by electronic speckle pattern interferometry (ESPI) during a uniaxial tensile test. It was shown that the mathematical model, a system of two crossing straight bands, describes accurately the evolution of the strain rate field from diffuse necking up to rupture. The physical characteristics of the localization, such as the width of the bands, their orientations and maximum strain rates were identified quantitatively. Their evolutions were followed and analyzed from several influencing factors, which are both microscopic, such as grain size and crystal structure, and macroscopic, like the geometry of the specimen and the direction in which the specimen were cut with respect to the rolling direction, and the pulling speed. It has been found that the bands narrow down during the necking process and their orientations were also changing. These evolutions, the necking mode and the transition between the diffuse and localized necking are influenced differently by these factors