This thesis is a study of perception-driven automatic segmentation of colour images. Despite immediate practical interest for this task, there exist very few reliable algorithms suitable for unsupervised processing. Most of the results presented in this thesis are based on mathematical morphology. This is a relatively new field which explores topological and geometrical properties of images and which has proven to be useful for image processing. The overview of morphological techniques can be found in chapter 2. A brief overview of segmentation methods is presented in, chapter 3. Only a small proportion of the vast number of publications on the subject is reviewed, namely those that are papers directly relevant to the subject of the thesis. Two novel non-parametric algorithms have been developed by the author for processing colour images. The first one is for processing randomly textured images. It uses a bottom-up segmentation algorithm which takes into consideration both colour and texture properties of the image. An "LUV gradient" is introduced which provides both a colour similarity measure and a basis for applying the watershed transform. The patches of watershed mosaic are merged according to their colour contrast until a termination criterion is met. This criterion is based on the topology of a typical processed image. The resulting algorithm does not require any additional information, be it various thresholds, marker extraction rules and suchlike, thus being suitable for automatic processing. The second algorithm deals with non-textured images and takes into consideration the noise that is present during the image acquisition. The watershed algorithm is used to segment either the 2- or 3-dimensional colour histogram of an image. To comply with the way humans perceive colour, this segmentation has to take place in a perceptually uniform colour space such as the Luv space. To avoid over segmentation, the watershed algorithm has to be applied to a smoothed-out histogram. The noise, however, is inhomogeneous in the Luv space and noise analysis for this space based on experimentally justified assumptions is presented. Both algorithms have been extensively tested on real data and were found to give stable results that are in good accord with human perception.

No description available.

A compact low-frequency projector is of crucial importance especially in underwater acoustics due to the frequency dependence of the absorption. To improve the efficiency of an omnidirectional acoustic source at low frequencies, parametric amplification may be used by adding a thin layer of nonlinear medium around a spherical transducer. The parametric effect is based on the interaction of two acoustic waves propagating through a nonlinear medium to produce a difference frequency wave. If both primary frequencies are sufficiently close enough, the result is the creation of a low frequency wave. Investigation is required to find the optimal medium, that is to say, one with a large nonlinear coefficient and a low sound speed. Such a source has already been built using a medium composed of a gel and microsphere mixture. In this case, the nonlinear coefficient is highly pressure dependent reaching a maximum when the microspheres buckle. The need is to optimize the material layer to increase the range of hydrostatic pressures over which the projector is useful.

Indiana University-Purdue University Indianapolis (IUPUI) / Background: Many authors have indicated that HIV-infected patients mortality risk is higher immediately following the start of Antiretroviral Therapy. However, mortality rate of HIV-infected patients is expected to decrease after starting Antiretroviral Therapy (ART) potentially complicating accurate statistical estimation of patient survival and, more generally, effective monitoring of the evolution of the worldwide epidemic. Method: In this thesis, we determine if mortality of HIV-patients increases or decreases after the initiation of ART therapy using flexible survival modelling techniques. To achieve this objective, this study uses semi-parametric statistical models for fitting and estimating survival time using different covariates. A combination of the Weibull distribution with splines is compared to the usual Weibull, exponential, and gamma distribution parametric models, and the Cox semi-parametric model. The objective of this study is to compare these models to find the best fitting model so that it can then be used to improve modeling of the survival time and explore the pattern of change in mortality rates for a cohort of HIV-infected patients recruited in a care and treatment program in Uganda. Results: The analysis shows that flexible survival Weibull models are better than usualoff-parametric and semi-parametric model fitting according to the AIC criterion. Conclusion: The mortality of HIV-patients is high right after the initiation of ART therapy and decreases rapidly subsequently.

published_or_final_version / abstract / Electrical and Electronic Engineering / Master / Master of Philosophy

No description available.

No description available.

We consider the convex parametric quadratic programming problem when the end of the parametric interval is caused by a multiplicity of possibilities ("ties"). In such cases, there is no clear way for the proper active set to be determined for the parametric analysis to continue. In this thesis, we show that the proper active set may be determined in general by solving a certain non-parametric quadratic programming problem. We simplify the parametric quadratic programming problem with a parameter both in the linear part of the objective function and in the right-hand side of the constraints to a quadratic programming without a parameter. We break the analysis into three parts. We first study the parametric quadratic programming problem with a parameter only in the linear part of the objective function, and then a parameter only in the right-hand side of the constraints. Each of these special cases is transformed into a quadratic programming problem having no parameters. A similar approach is then applied to the parametric quadratic programming problem having a parameter both in the linear part of the objective function and in the right-hand side of the constraints.

In this work I illustrate an approach to the development of a library of problem solving components for knowledge modelling. This approach is based on an epistemological modelling framework, the Task/Method/Domain/Application (TMDA) model, and on a principled methodology, which provide an integrated view of both library construction and application development by reuse. The starting point of the proposed approach is given by a task ontology. This formalizes a conceptual viewpoint over a class of problems, thus providing a task-specific framework, which can be used to drive the construction of a task model through a process of model-based knowledge acquisition. The definitions in the task ontology provide the initial elements of a task-specific library of problem solving components. In order to move from problem specification to problem solving, a generic, i.e. taskindependent, model of problem solving as search is introduced, and instantiated in terms of the concepts in the relevant task ontology, say T. The result is a task-specific, but method-independent, problem solving model. This generic problem solving model provides the foundation from which alternative problem solving methods for a class of tasks can be defined. Specifically, the generic problem solving model provides i) a highly generic method ontology, say M; ii) a set of generic building blocks (generic tasks), which can be used to construct task-specific problem solving methods; and iii) an initial problem solving method, which can be characterized as the most generic problem solving method, which subscribes to M and is applicable to T. More specific problem solving methods can then be (re-)constructed from the generic problem solving model through a process of method/ontology specialization and method-to-task application. The resulting library of reusable components enjoys a clear theoretical basis and provides robust support for reuse. In the thesis I illustrate the approach in the area of parametric design.

Several pulsed optical oscillators (OPOs) based on periodically poled lithium niobate (PPLN) and pumped by single longitudinal mode Ti:sapphire lasers have been developed. These OPOs provide access to important spectroscopic regions in the 1 - 5.5 [mu]m region and can be rapidly turned by varying the pump wavelength. Previously many of the OPOs developed to take advantage of PPLN relied on a combination of period selection and temperature tuning and as a result were slow and cumbersome to tune. This problem my be avoided by using tunable pump sources or acoustically induced strain waves. Several candidate OPO pump sources were characterised. These pump sources with themselves pumped by lasers operating at repetition reates of either 1.5 kHz (high repetition rate) or 10Hz (low repetition rate). High repetition rate systems include: a Ti-sapphire laser, injection seeded by a single longitudinal mode diode laser, several coupled cavity Ti:sapphire lasers with bandwidths less than 100 Ghz and Cr:forsterite lasers narrowed by prisms and étalons. The low repetition rate systems were all coupled cavity Ti:sapphire lasers one of which was single and double pass amplified. Of these it was found that only the high repetition rate injection seeded laser and the low repetition Ti:sapphire lasers were suitable as OPO pump sources. OPOs were characterised at high and low repetition rates. The high repetition rate system exhibited a low threshold of oscillation (18.7 [mu]J) and a low overall efficiency (25%) which was thought to be due to the pulse to pulse variability of the Ti:sapphire bandwidth. The tuning range of the OPO was 932 to 1310 nm (signal) and 1.989 [mu]m to 5.281 [mu]m (idler) using multiple poling periods and only 15 nm of pump tuning. OPO oscillation on two separate signals simultaneously was observed. Two separate low repetition rate systems were investigated; the first was tuned from 1200 to 1600 nm (signal) and from 1600 to 2400 nm (idler) on a single poling period with a high absolute efficiency of 35% and a threshold of 180 [mu]J. The second OPO was tuned from 940 to 1220 nm (signal) and 2.2 to 4.3 [mu]m (idler) on a single poling period. The absolute efficiency of the system was 25% and the threshold was 200 [mu]J. OPO oscillation on two separate signals was investigated using an OPO based on grazing incidence configured cavity. It was found that the signals coupled together through Raman transitions present in lithium niobate and that coupling reduced the efficiency of the device as a whole. The affect of an acoustically induced strain field on the optical nonlinearity of tetragonal ferroelectric materials was investigated. It was found that the optical nonlinear coefficient varies linearly with the cell displacement and as the square root of the acoustic power. A crystal designed to implement a quasi phase matched interaction based on this variation is proposed.