Skip directly to content

Coming dissertations at TekNat

  • Application, Optimisation and Evaluation of Deep Learning for Biomedical Imaging

    Author: Håkan Wieslander
    Publication date: 2022-05-13 14:04

    Microscopy imaging is a powerful technique when studying biology at a cellular and sub-cellular level. When combined with digital image analysis it creates an invaluable tool for investigating complex biological processes and phenomena. However, imaging at the cell and sub-cellular level tends to generate large amounts of data which can be difficult to analyse, navigate and store. Despite these difficulties, large data volumes mean more information content which is beneficial for computational methods like machine learning, especially deep learning. The union of microscopy imaging and deep learning thus provides numerous opportunities for advancing our scientific understanding and uncovering interesting and useful biological insights.

    The work in this thesis explores various means for optimising information extraction from microscopy data utilising image analysis with deep learning. The focus is on three different imaging modalities: bright-field; fluorescence; and transmission electron microscopy. Within these modalities different learning-based image analysis and processing techniques are explored, ranging from image classification and detection to image restoration and...

  • Tailoring Gaussian processes and large-scale optimisation

    Author: Carl Jidling
    Publication date: 2022-05-12 14:55

    This thesis is centred around Gaussian processes and large-scale optimisation, where the main contributions are presented in the included papers.

    Provided access to linear constraints (e.g. equilibrium conditions), we propose a constructive procedure to design the covariance function in a Gaussian process. The constraints are thereby explicitly incorporated with guaranteed fulfilment. One such construction is successfully applied to strain field reconstruction, where the goal is to describe the interior of a deformed object. Furthermore, we analyse the Gaussian process as a tool for X-ray computed tomography, a field of high importance primarily due to its central role in medical treatments. This provides insightful interpretations of traditional reconstruction algorithms. 

    Large-scale optimisation is considered in two different contexts. First, we consider a stochastic environment, for which we suggest a new method inspired by the quasi-Newton framework. Promising results are demonstrated on real world benchmark problems. Secondly, we suggest an approach to solve an applied deterministic optimisation problem that arises within the design of electrical circuit...

  • Phase transitions in magnetic metamaterials

    Author: Björn Erik Skovdal
    Publication date: 2022-05-12 10:55

    Magnetic metamaterials consisting of arrays of densely packed, two-dimensional nanoscale magnetic islands have degrees of freedom on two separate length scales: inside the islands, and among them. These degrees of freedom can be tuned by e.g. size, shape, island separation and lattice geometry. The material can thereby be tailored to display behavior corresponding to conventional universality classes, wherein small elongated islands behave like Ising spins and circular ones behave like XY-spins. Making the islands larger promotes inner degrees of freedom in the form of inner magnetic textures. Some of these textures, such as magnetic vortices in circular islands, have a critical impact on the interaction between the islands and therefore also on the global order.

    In this thesis, the interplay between the inner textures and island-island interactions is explored, anticipating the emergence of behavior beyond that of conventional universality classes. A transition temperature between static and dynamic inner textureswas found in systems with elongated islands. In arrays of circular islands, a collapse from metastable collinear islands to vortex islands was observed, with a...

  • Properties of invariant sets in certain two-dimensional dynamical systems : Renormalization and beyond

    Author: Dan Lilja
    Publication date: 2022-05-12 10:37

    This thesis consists of an introduction and four research papers concerning dynamical systems, focusing on renormalization in two dimensions.

    Paper I (joint with Jordi-Lluís Figueras) studies a generalization of the anti-integrable limit of skew-product systems based on a Frenkel-Kontorova model. It is shown that under certain regularity conditions the orbits of such a system has a Cantor set structure and the existence of a non-smooth folding bifurcation is deduced.

    Paper II studies the invariant Cantor sets of period doubling type of infinitely renormalizable area-preserving maps in the universality class of the Eckmann-Koch-Wittwer renormalization fixed point. It is shown that for such maps the invariant Cantor set is always contained in a Lipschitz curve but never in a smooth (meaning C1) curve.

    Paper III (joint with Denis Gaidashev) continues the study of the invariant Cantor sets of infinitely renormalizable area-preserving maps, focusing on their geometry. It is shown that there is always a positive measure of unbounded geometry. Additionally, some results giving control over the average expansion of vectors by the renormalization microscope are also...

  • Deradicalising Giardiasis Treatment

    Author: Sascha Krakovka
    Publication date: 2022-05-11 10:21

    Giardia intestinalis is a unicellular parasite causing the disease giardiasis. This disease is mostly prevalent in low resource settings, mandating low-cost treatment options. Treatment is mainly based on antiparasitics from the classes of 5-nitroimidazoles and benzimidazoles, which both target parasite DNA and other cellular macromolecules based on radicals generated from the drugs. In recent years resistance to those classes and cross-resistance in between them has become a problem, hence alternative antigiardials are needed.

    In this thesis we enhanced our understanding of the crucial differentiation process of encystation, which produce the environmentally stable and infective form of the parasite, the cyst (Paper I). Resistance development has been slow in G. intestinalis until now and it has been shown that resistance to the main treatment option, metronidazole, can be lost after en- and excystation so an enhanced understanding of this process can help us to identify the cause of this loss of drug tolerance. In the second study (Paper II) we followed this up by analysing two metronidazole resistant lines and one revertant on their ability to produce...

  • Oversampled radial basis function methods for solving partial differential equations

    Author: Igor Tominec
    Publication date: 2022-05-10 10:57

    Partial differential equations (PDEs) describe complex real-world phenomena such as weather dynamics,  object deformations, financial trading prices, and fluid-structure interaction.  The solutions of PDEs are commonly used to enhance the understanding of these phenomena and also as leverage to make technological improvements to consumer products. In the present thesis, we  develop numerical methods for solving PDEs using computers. The focus is on radial basis  function (RBF) methods that are appreciated for their high-order accuracy and ease of implementation in higher dimensions, but can  sometimes face numerical stability challenges.  To circumvent the stability issues, we use an oversampled approach to discretize PDEs as opposed to the more commonly used collocated approach.  Throughout the thesis, we mainly use the RBF-generated finite difference (RBF-FD) method, but we also use the RBF partition  of unity method (RBF-PUM) and Kansa's global RBF method in one part of the thesis.  The first two methods are local in the sense that the underlying discretization matrices are sparse,  while the third method is global, leading to dense discretization matrices. In Paper I we...

  • Prostate cancer incidence, treatment and mortality : Empirical longitudinal register-based studies and methods for handling missing data

    Author: Marcus Westerberg
    Publication date: 2022-05-09 13:13

    The diagnostic activity for prostate cancer has increased substantially in Sweden, primarily due to increased use of prostate-specific antigen (PSA) testing in asymptomatic men, and this has led to a large increase in diagnoses. There have also been changes in the diagnostic workup, guidelines, treatment strategies, and more effective treatments have been introduced in different phases of the disease. This thesis aims to increase the understanding of consequences of changes in diagnostic activity and treatment, with a focus on empirical studies, methodological development, and handling of missing data.

    In paper I, the survival of men with metastatic prostate cancer was investigated across calendar time periods by use of Kaplan-Meier analyses and Cox regression. The median survival from diagnosis increased with six months comparing men diagnosed 1998-2001 with men diagnosed 2010-2015, while median PSA decreased.

    In paper II, a discrete time multivariate longitudinal model was combined with a proxy for the unobserved level of diagnostic activity to produce prognoses of incidence and mortality. Simulations indicated that a higher diagnostic activity was associated with...

  • Nanoparticles based molecular electronic devices with tunable molecular functionalization shell and gas sensing measurements

    Author: Omer Sher
    Publication date: 2022-05-09 11:25

    The idea to use molecules as a basic building block in electronic circuits was developed about 50 years ago when a molecular rectifier was developed, but it has been a challenge for this field to make its way to real-world application. Now, due to the advancement in technologies, the properties of single molecules are better understandable and controllable. Some of the main motivations to build molecular electronics devices are that the conductive molecules can be as small as about 1 nm, that they are stable objects and can be tailor-made with desired electronic properties. This small size of molecule poses a challenge in their usage, one solution is to develop the hybrid devices whose properties are based on single and few molecules.

    In this study, a portable hybrid device is used and further developed called a nanoMoED device, a nano-molecular electronic device. These devices consist of gold nanoparticles (AuNPs), gold nanoelectrodes and conjugated organic molecules. The electrical resistance of the device depends on the molecules functionalizing it and, in this work, they contain phenyl rings such as 4,4’-biphenyldithiol (BPDT), p-ter-phenyl-4,4''-dithiol and oligo...

  • Symbiosis in Arbuscular Mycorrhizal Fungi : Unearthing genomic signatures

    Author: Shadi Eshghi Sahraei
    Publication date: 2022-05-09 10:34

    Arbuscular mycorrhizal (AM) fungi have been on earth since 470 million years ago. AM fungi form mutualistic symbiosis with most terrestrial plants and thus, are thought to have played important roles in land colonization together with plants. In this thesis, I studied AM fungi in their natural habitat (Paper I and II) and examined their host preference (Paper IV). Further, I identified some genomic signatures of these obligate symbionts (Paper III). In Paper I, I studied soil communities of microeukaryotes in a protected grassland. Ribosomal DNA from soil samples was amplified (1500 bp) and sequenced with Pacific Bioscience technology. Three clustering methods were used to generate operational taxonomic units (OTUs). The three methods agreed on the effects of soil moisture and presence of Fritillaria meleagris on microeukaryotic communities but disagreed about specific compositions of these communities. Soil microeukaryotic communities were dominated by protists and AM fungi had a low abundance in spite of a high spore abundance. In Paper II, the genome of ...

  • Droplet Acoustofluidics for Biochemical Applications

    Author: Zhenhua Liu
    Publication date: 2022-05-05 09:38

    Droplet microfluidics is a promising platform for biochemical applications where compartmentalized droplets serve as individual vials. Droplets are formed by using two immiscible phases, the continuous phase and the dispersed phase, making up the droplets. Droplets are interesting because they can provide fast, parallel reactions with low reagent consumption. Microscale particles, such as cells, can be encapsulated in the droplets and chemical reagents can be added via a pico-injector. However, removal of droplet background signal is hard to achieved by conventional methods, especially if you do not want to risk losing the encapsulated cells. In this thesis, I present a droplet microfluidic system that can achieve this, via droplet-internal particle manipulation using acoustophoresis.

    This droplet microfluidic system contains pico-injection and droplet split with acoustophoresis. The pico-injection is used to add fresh solution into the droplets and the droplet split with acoustophoresis is used to remove the droplet supernatant. With the combination of the pico-injector and the droplet split, the background signal of the droplets can be reduced and the cell medium in the...