DTU Orbit RSS Feed
http://orbit.dtu.dk/en/publishers-research/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08)/search/publications.html?filter=research
RSS FeedFri, 09 Oct 2015 08:24:38 GMT2015-10-09T08:24:38ZMethods and Algorithms for Economic MPC in Power Production Planning
http://orbit.dtu.dk/en/publications/methods-and-algorithms-for-economic-mpc-in-power-production-planning(5f47f4b5-882c-4a22-8516-997886e3e912).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/methods-and-algorithms-for-economic-mpc-in-power-production-planning(5f47f4b5-882c-4a22-8516-997886e3e912).html" class="link"><span>Methods and Algorithms for Economic MPC in Power Production Planning</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/leo-emil-sokoler(a562a71b-b70f-4d22-8e18-5f138d73a3d8).html" class="link person"><span>Sokoler, L. E.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/john-bagterp-joergensen(d2bc6de4-a5c6-43f7-8712-606bd02a88e9).html" class="link person"><span>Jørgensen, J. B.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/henrik-madsen(458e950b-adfd-4f2d-8d0d-ef39dee24bea).html" class="link person"><span>Madsen, H.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/niels-kjoelstad-poulsen(6c1bb106-d448-47cb-978c-3d9efee3ca97).html" class="link person"><span>Poulsen, N. K.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">247 p.</span> (DTU Compute PHD-2015; No. 377).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">This thesis concerns methods and algorithms for power production planning in contemporary and future power systems. Power production planning is a task that involves decisions across different time scales and planning horizons. Hours-ahead to days-ahead planning is handled by solving a mixed-integer linear program for unit commitment and economic dispatch of the system power generators. We focus on a minutes-ahead planning horizon, where unit commitment decisions are fixed. Economic model predictive control (EMPC) is employed to determine an optimal dispatch for a portfolio of power generators in real-time. A generator can represent a producer of electricity, a consumer of electricity, or possibly both. Examples of generators are heat pumps, electric vehicles, wind turbines, virtual power plants, solar cells, and conventional fuel-fired thermal power plants. Although this thesis is mainly concerned with EMPC for minutes-ahead production planning, we show that the proposed EMPC scheme can be extended to days-ahead planning (including unit commitment) as well. <br /><br />The power generation from renewable energy sources such as wind and solar power is inherently uncertain and variable. A portfolio with a high penetration of renewable energy is therefore a stochastic system. To accommodate the need for EMPC of stochastic systems, we generalize certainty-equivalent EMPC (CEEMPC) to mean-variance EMPC (MV-EMPC). In MV-EMPC, the objective function is a trade-off between the expected cost and the cost variance. Simulations show that MV-EMPC reduces cost and risk compared to CE-EMPC. The simulations also show that the economic performance of CE-EMPC can be much improved using a constraint back-off heuristic.
<br /><br />Efficient solution of the optimal control problems (OCPs) that arise in EMPC is important, as the OCPs are solved online. We present special-purpose algorithms for EMPC of linear systems that exploit the high degree of structure in the OCPs. A Riccati-based homogeneous and self-dual interior-point method is developed for the special case, where the OCP objective function is a linear function. We design an algorithm based on the alternating direction method of multipliers (ADMM) to solve input-constrained OCPs with convex objective functions. The OCPs that occur in EMPC of dynamically decoupled subsystems, e.g. power generators, have a block-angular structure. Subsystem decomposition algorithms based on ADMM and Dantzig-Wolfe decomposition are proposed to solve these OCPs. Subproblems that arise in the decomposition algorithms are solved using structure-exploiting algorithms. To reduce computation time of the EMPC algorithms further, warm-start and early-termination strategies are employed. Benchmarks show that the special-purpose algorithms are significantly faster than current state-of-the-art solvers.
<br /><br />As a potential application area of EMPC, we study power production planning in small isolated power systems. A critical part of power production planning in small isolated power systems is operational reserve planning. The operational reserves are activated to balance production and consumption in real-time. An EMPC scheme is presented for activation of operational reserves. Simulations based on a Faroe Islands case study show that signi_cant cost savings can be achieved using this strategy. For efficient planning of the operational reserves, we present an optimal reserve planning problem (ORPP). The ORPP is a contingency-constrained unit commitment problem that addresses low inertia challenges in small isolated power systems.
<br /><br />In summary, the main contributions of this thesis are:
<br />- A mean-variance optimization strategy for EMPC of linear stochastic systems.
<br />- Tailored algorithms for solution of the OCPs that arise in EMPC of linear stochastic systems.
<br />- Methods for power production planning in small isolated power; the ORPP for unit commitment and economic dispatch, and an EMPC scheme for activation of operational reserves.
</div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>247</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>377</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/methods-and-algorithms-for-economic-mpc-in-power-production-planning(5f47f4b5-882c-4a22-8516-997886e3e912).html2014-12-31T23:00:00ZFine-grained Information Flow for Concurrent Computation
http://orbit.dtu.dk/en/publications/finegrained-information-flow-for-concurrent-computation(a58bf713-0416-4fe3-84c0-fbe60663c3d5).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/finegrained-information-flow-for-concurrent-computation(a58bf713-0416-4fe3-84c0-fbe60663c3d5).html" class="link"><span>Fine-grained Information Flow for Concurrent Computation</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/ximeng-li(9083fd55-4fa2-4584-80be-ce977ba23719).html" class="link person"><span>Li, X.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/flemming-nielson(f63ba194-185f-465d-9b2b-1f6ff489bc3d).html" class="link person"><span>Nielson, F.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/hanne-riis-nielson(44a7d08e-b083-4419-a466-8f4392a7112f).html" class="link person"><span>Nielson, H. R.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">186 p.</span> (DTU Compute PHD-2015; No. 388).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">It is essential to protect IT systems against security threats. An example would be the control of aircraft, which uses an internal network that passengers can access. It is important to ensure that malicious code on passenger equipment cannot endanger flight safety.<br /><br />Information flow control is an important approach to the protection of systems against such threats. Notable examples include tainting analyses in languages such as Javascript, and program transformations on cryptographic algorithms to avoid information leakage through running time. A wide variety of techniques, including type systems and reference monitors, have been proposed in the context of programming languages and process calculi, to enforce such properties. The most widely used definitions of information flow security are noninterference-like properties.<br /><br />For concurrent systems where processes communicate with each other to accomplish computational tasks, fine-grained security policies can be formulated by distinguishing between whether communication can happen, and what is communicated. As the first contribution of this PhD thesis, we formulate a noninterference-like property that takes all combinations of sensitivity levels for “whether” and “what” into consideration, emphasizing the importance of the integrity case where the former is more sensitive than the latter. This case captures the effect of Message Authentication Codes (MAC) and the consequence of Denial of Service (DoS) attacks. It is also proved that the property degenerates to a classical one when the two dimensions are intentionally blurred. <br /><br />As the second contribution, we focus on the “what” dimension and further allow the flow policy to vary under different contents stored and communicated. This is the area of content-dependent (or conditional) information flow, which has recently been studied for sequential programs. We generalize the use and enforcement of content-dependent flow policies to concurrent, communicating processes. A security type system is developed, incorporating a Hoare logic component that provides approximations of the memory contents at different program points. Most proofs for the theoretical results on content-dependency are performed in the Coq proof assistant.<br /><br />The third contribution of this thesis is the obtainment of compositionality results that support modular security analyses of computer systems. <br /><br />A multiplexer pattern that separates sensitive and non-sensitive network traffic is used as a running example. Whether communications can happen is easily influenced by an attacker — attacking one of the incoming channels would suffice. In any case, the two data paths are still differentiable by the sensitivity levels of what is communicated. In case the destinations of messages are determined by their tagging, content-dependent policies are able to convey the correlation between the sensitivity level of a message and its tagging, and our Hoare-logic equipped type system allows a modular analysis of the overall system.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>186</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>388</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/finegrained-information-flow-for-concurrent-computation(a58bf713-0416-4fe3-84c0-fbe60663c3d5).html2014-12-31T23:00:00ZAn Adaptive Middleware for Improved Computational Performance
http://orbit.dtu.dk/en/publications/an-adaptive-middleware-for-improved-computational-performance(60970c81-8c64-426b-b4e5-0250f1020de0).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/an-adaptive-middleware-for-improved-computational-performance(60970c81-8c64-426b-b4e5-0250f1020de0).html" class="link"><span>An Adaptive Middleware for Improved Computational Performance</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/lars-frydendal-bonnichsen(2b1c58ed-a969-41c8-b77c-1753cf275cc6).html" class="link person"><span>Bonnichsen, L. F.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/christian-w-probst(6f0f045b-6369-4416-a241-c59d9214b6b4).html" class="link person"><span>Probst, C. W.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/sven-karlsson(0994c73f-6656-4260-b54d-3f9f426c4edb).html" class="link person"><span>Karlsson, S.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">134 p.</span> (DTU Compute PHD-2015; No. 390).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">The performance improvements in computer systems over the past 60 years have been fueled by an exponential increase in energy efficiency. In recent years, the phenomenon known as the end of Dennard’s scaling has slowed energy efficiency improvements — but improving computer energy efficiency is more important now than ever. Traditionally, most improvements in computer energy efficiency have come from improvements in lithography — the ability to produce smaller transistors — and computer architecture - the ability to apply those transistors efficiently. Since the end of scaling, we have seen diminishing returns from developments in lithography and modern computer architectures are so complicated requiring significant programming effort to exploit efficiently —software developers undertaking such a task will need all the help they can get, in order to keep the programming effort down.<br /><br />In this thesis we champion using software to improve energy efficiency — in particular we develop guidelines for reasoning and evaluating software performance on modern computers, and a middleware that has been designed for modern computers, improving computational performance both in terms of energy and execution time. Our middleware consists of a new power manager, synchronization libraries using hardware transactional memory (for locks, barriers, and task synchronization), and two concurrent map data structures, which can be deployed in computer systems with little to no effort. At a fundamental level, we are improving computational performance by exploiting modern hardware features, such as dynamic voltage-frequency scaling and transactional memory. Adapting software is an iterative process, requiring that we continually revisit it to meet new requirements or realities; a time consuming process which we hope to simplify by analyzing the realities of modern computers, and providing guidelines explaining how to get the most performance out of them.</div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>134</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>390</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/an-adaptive-middleware-for-improved-computational-performance(60970c81-8c64-426b-b4e5-0250f1020de0).html2014-12-31T23:00:00ZData Analysis of Medical Images: CT, MRI, Phase Contrast X-ray and PET
http://orbit.dtu.dk/en/publications/data-analysis-of-medical-images-ct-mri-phase-contrast-xray-and-pet(51fb84f6-eca6-40ab-9ca3-592182396128).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/data-analysis-of-medical-images-ct-mri-phase-contrast-xray-and-pet(51fb84f6-eca6-40ab-9ca3-592182396128).html" class="link"><span>Data Analysis of Medical Images: CT, MRI, Phase Contrast X-ray and PET</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/anders-nymark-christensen(ad9ef4bc-d943-459d-b52f-a11c61a9ab0d).html" class="link person"><span>Christensen, A. N.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/knut-conradsen(58fbfe83-3565-4584-8b50-3a01a72134e0).html" class="link person"><span>Conradsen, K.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/rasmus-larsen(b3ea098f-fc7a-48a0-94ae-ae033577d7f1).html" class="link person"><span>Larsen, R.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">194 p.</span> (DTU Compute PHD-2015; No. 386).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Data analysis of medical images is an important and growing area, as systems for imaging becomes still more available and complex. <br /><br />The goal of the thesis is to demonstrate solutions to data analysis problems in a cross disciplinary context. Further, to develop methods for analysis of new imaging modalities and to combine cross disciplinary knowledge from various fields to find new solutions to existing problems.<br /><br />More speciffically the thesis shows segmentation of images, classification and statistics used on a variety of quite different problems. Active Appearance models, Chan-Vese and graph-cut has been used, as well as a variety of statistical tools centred on the General Linear Model.<br /><br />The point of departure for the thesis is the NanoGuide project, in which gel based x-ray markers for use in radiotherapy has been developed. Two different types of gels has been analysed using segmentation of micro-CT images followed by a statistical analysis of homogeneity, contrast, degradation, and other qualities. By combining knowledge from the different professions in the project, a new application for one of the developed gels - in-vivo dosimetry in radiotherapy - has been studied.<br /><br />Analysis of differences between groups and of correlations between brain regions and cognitive tests in alzheimers patients is another contribution. Segmentation of fat in abdominal MRI-scans has also been studied and a robust algorithm based on graph-cut is presented.<br /><br />A relatively new modality phase-contrast x-ray and dark-field has shown promise for diagnosis of a variety of diseases in the lungs. A classification algorithm for differentiation of healthy, emphysematous and fibrotic lung tissue on pixel level is presented.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>194</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>386</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/data-analysis-of-medical-images-ct-mri-phase-contrast-xray-and-pet(51fb84f6-eca6-40ab-9ca3-592182396128).html2014-12-31T23:00:00ZAnalysis of Security Protocols in Embedded Systems
http://orbit.dtu.dk/en/publications/analysis-of-security-protocols-in-embedded-systems(d926bd84-8310-42e0-a233-05cd8283f8b5).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/analysis-of-security-protocols-in-embedded-systems(d926bd84-8310-42e0-a233-05cd8283f8b5).html" class="link"><span>Analysis of Security Protocols in Embedded Systems</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/alessandro-bruni(04d6e2ab-ca4c-408b-be86-a5ce7a5467db).html" class="link person"><span>Bruni, A.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/flemming-nielson(f63ba194-185f-465d-9b2b-1f6ff489bc3d).html" class="link person"><span>Nielson, F.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/hanne-riis-nielson(44a7d08e-b083-4419-a466-8f4392a7112f).html" class="link person"><span>Nielson, H. R.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">153 p.</span> (DTU Compute PHD-2015; No. 389).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Embedded real-time systems have been adopted in a wide range of safety-critical applications—including automotive, avionics, and train control systems—where the focus has long been on safety (i.e., protecting the external world from the potential damage caused by the system) rather than security (i.e., protecting the system from the external world). With increased connectivity of these systems to external networks the attack surface has grown, and consequently there is a need for securing the system from external attacks. Introducing security protocols in safety critical systems requires careful considerations on the available resources, especially in meeting real-time and resource constraints, as well as cost and reliability requirements. For this reason many proposed security protocols in this domain have peculiar features, not present in traditional security literature.<br /><br />In this thesis we tackle the problem of analysing security protocols in safety critical embedded systems from multiple perspectives, extending current state-of-the-art analysis techniques where the combination of safety and security hinders our efforts. Examples of protocols in automotive control systems will follow throughout the thesis. We initially take a combined perspective of the safety and security features, by giving a security analysis and a schedulability analysis of the embedded protocols, with intertwined considerations. Then we approach the problem of the expressiveness of the tools used in the analysis, extending saturation-based techniques for formal protocol verification in the symbolic model. Such techniques gain much of their efficiency by coalescing all reachable states into a single set of facts. However, distinguishing different states is a requirement for modelling the protocols that we consider. Our effort in this direction is to extend saturation-based techniques so that enough state information can be modelled and analysed. Finally, we present a methodology for proving the same security properties in the computational model, by means of typing protocol implementations.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>153</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>389</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/analysis-of-security-protocols-in-embedded-systems(d926bd84-8310-42e0-a233-05cd8283f8b5).html2014-12-31T23:00:00ZTomographic Image Reconstruction Using Training Images with Matrix and Tensor Formulations
http://orbit.dtu.dk/en/publications/tomographic-image-reconstruction-using-training-images-with-matrix-and-tensor-formulations(1a909881-23a4-4850-844b-71d0afde68c6).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/tomographic-image-reconstruction-using-training-images-with-matrix-and-tensor-formulations(1a909881-23a4-4850-844b-71d0afde68c6).html" class="link"><span>Tomographic Image Reconstruction Using Training Images with Matrix and Tensor Formulations</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/sara-soltani(d12e3a38-c482-476e-9a9e-deafc37208bf).html" class="link person"><span>Soltani, S.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/per-christian-hansen(6c6c6f32-e3b2-40d5-a407-aa8e2f051f7b).html" class="link person"><span>Hansen, P. C.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/martin-skovgaard-andersen(e065a9e6-1849-4ec2-b25f-e6394f9fe9a0).html" class="link person"><span>Andersen, M. S.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">131 p.</span> (DTU Compute PHD-2015; No. 387).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Reducing X-ray exposure while maintaining the image quality is a major challenge in computed tomography (CT); since the imperfect data produced from the few view and/or low intensity projections results in low-quality images that are suffering from severe artifacts when using conventional reconstruction methods. Incorporating a priori information about the solution is a necessity to improve the reconstruction. For example, Total Variation (TV) regularization method –assuming a piecewise constant image model – has been shown to allow reducing X-ray exposure significantly, while maintaining the image resolution compared to a classical reconstruction method such as Filtered Back Projection (FBP). <br /><br />Some priors for the tomographic reconstruction take the form of cross-section images of similar objects, providing a set of the so-called training images, that hold the key to the structural information about the solution. The training images must be reliable and application-specific. This PhD project aims at providing a mathematical and computational framework for the use of training sets as non-parametric priors for the solution in tomographic image reconstruction. Through an unsupervised machine learning technique (here, the dictionary learning), prototype elements from the training images are extracted and then incorporated in the tomographic reconstruction problem both with matrix and tensor representations of the training images.<br /><br />First, an algorithm for the tomographic image reconstruction using training images, where the training images are represented as vectors in a training matrix, is described. The dictionary learning problem is formulated as a regularized non-negative matrix factorization in order to compute a nonnegative dictionary. Then a tomographic solution with a sparse representation in the dictionary is obtained through a convex optimization formulation. Computational experiments clarify the choice and interplay of the model parameters and the regularization parameters. Furthermore, the assumptions in the tomographic problem formulation are analyzed. The sensitivity and robustness of the reconstruction to variations of the scale and rotation in the training images is investigated and algorithms to estimate the correct relative scale and orientation of the unknown image to the training images are suggested.<br /><br />Then, a third-order tensor representation for the training images images is used. The dictionary and image reconstruction problem are reformulated using the tensor representation. The dictionary learning problem is presented as a nonnegative tensor factorization problem with sparsity constraints and the reconstruction problem is formulated in a convex optimization framework by looking for a solution with a sparse representation in the tensor dictionary. Numerical results show considering a tensor formulation over a matrix formulation significantly reduces the approximation error by the dictionary as well as leads to very sparse representations of both the training images and the reconstructions. <br /><br />Further computational experiments show that in few-projection and low-dose settings our algorithm is while (not surprisingly) being superior to the classical reconstruction methods, is competitive with (or even better of) the TV regularization and tends to include more texture and sharper edges in the reconstructed images.<br /><br />The focus of the thesis is the study of mathematical and algorithmic prospectives and thus the training images and tomographic scenarios are mostly simulation based. More studies are however needed for implementing the proposed algorithm in a routine use for clinical applications and materials testing.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>131</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>387</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/tomographic-image-reconstruction-using-training-images-with-matrix-and-tensor-formulations(1a909881-23a4-4850-844b-71d0afde68c6).html2014-12-31T23:00:00ZTime-predictable Stack Caching
http://orbit.dtu.dk/en/publications/timepredictable-stack-caching(e59c1c6a-46e8-4243-8740-5667577d1c22).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/timepredictable-stack-caching(e59c1c6a-46e8-4243-8740-5667577d1c22).html" class="link"><span>Time-predictable Stack Caching</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/sahar-abbaspourseyedi(1413dab9-306e-469e-85bd-0b7e6ee9a738).html" class="link person"><span>Abbaspourseyedi, S.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/martin-schoeberl(63dcfd1c-cc7f-4e1b-8fdc-1d421f8fdb36).html" class="link person"><span>Schoeberl, M.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/jens-sparsoe(8ef14d93-7486-4521-adba-0ba0d0932be3).html" class="link person"><span>Sparsø, J.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">133 p.</span> (DTU Compute PHD-2015; No. 385).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Embedded systems are computing systems for controlling and interacting with physical environments. Embedded systems with special timing constraints where the system needs to meet deadlines are referred to as real-time systems. In hard real-time systems, missing a deadline causes the system to fail completely. Thus, in systems with hard deadlines the worst-case execution time (WCET) of the real-time software running on them needs to be bounded. <br /><br />Modern architectures use features such as pipelining and caches for improving the average performance. These features, however, make the WCET analysis more complicated and less imprecise. Time-predictable computer architectures provide solutions to this problem. As accesses to the data in caches are one source of timing unpredictability, devising methods for improving the timepredictability of caches are important. Stack data, with statically analyzable addresses, provides an opportunity to predict and tighten the WCET of accesses to data in caches.<br /><br />In this thesis, we introduce the time-predictable stack cache design and implementation within a time-predictable processor. We introduce several optimizations to our design for tightening the WCET while keeping the timepredictability of the design intact. Moreover, we provide a solution for reducing the cost of context switching in a system using the stack cache. In design of these caches, we use custom hardware and compiler support for delivering time-predictable stack data accesses. Furthermore, for systems where compiler support or hardware changes are not practical, we propose and explore two different alternatives based on only software and only hardware support.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>133</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>385</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/timepredictable-stack-caching(e59c1c6a-46e8-4243-8740-5667577d1c22).html2014-12-31T23:00:00ZAspects of the Tutte polynomial
http://orbit.dtu.dk/en/publications/aspects-of-the-tutte-polynomial(b7bbd3d4-39f7-49cb-853a-139aae217468).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/aspects-of-the-tutte-polynomial(b7bbd3d4-39f7-49cb-853a-139aae217468).html" class="link"><span>Aspects of the Tutte polynomial</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/seongmin-ok(43d463d6-b7a0-49f1-a3b7-ddf526b1934b).html" class="link person"><span>Ok, S.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/carsten-thomassen(88ab66ea-2b88-44e7-88f2-b8b883cf3ec0).html" class="link person"><span>Thomassen, C.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">110 p.</span> (DTU Compute PHD-2015; No. 384).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">This thesis studies various aspects of the Tutte polynomial, especially focusing on the Merino-Welsh conjecture.<br /><br />We write T(G;x,y) for the Tutte polynomial of a graph G with variables x and y. In 1999, Merino and Welsh conjectured that if G is a loopless 2-connected graph, then<br /><br />T(G;1,1) ≤ max{T(G;2,0), T(G;0,2)}.<br /><br />The three numbers, T(G;1,1), T(G;2,0) and T(G;0,2) are respectively the numbers of spanning trees, acyclic orientations and totally cyclic orientations of G.<br /><br />First, I extend Negami's splitting formula to the multivariate Tutte polynomial. Using the splitting formula, Thomassen and I found a lower bound for the number of spanning trees in a k-edge-connected graph. Our bound is tight for k even, but for k odd we give a slightly better lower bound which we believe is not tight. We prove that the minimum number of spanning trees in a 3-edge-connected graph with n vertices is, not surprisingly, significantly smaller than the minimum number of spanning trees in a 4-edge-connected graph. However, we conjecture that the minimum number of spanning trees of a 5-edge-connected graph is actually obtained by a 6-edge-connected graph asymptotically.<br /><br />Thomassen proved the following partial result for the Merino-Welsh conjecture. Assume the graph G is loopless, bridgeless and has n vertices and m edges.<br /><br />If m ≤ 1.066 n then T(G;1,1) ≤ T(G;2,0).<br />If m ≥ 4(n-1) then T(G;1,1) ≤ T(G;0,2).<br /> <br />I improve in this thesis Thomassen's result as follows:<br /><br />If m ≤ 1.29(n-1) then T(G;1,1) ≤ T(G;2,0).<br />If m ≥ 3.58(n-1) and G is 3-edge-connected then T(G;1,1) ≤ T(G;0,2).<br /><br />Strengthening Thomassen's idea that acyclic orientations dominate spanning trees in sparse graphs, I conjecture that the ratio T(G;2,0)/T(G;1,1) increases as G gets sparser. To support this conjecture, I prove a variant of the conjecture for series-parallel graphs.<br /><br />The Merino-Welsh conjecture has a stronger version claiming that the Tutte polynomial is convex on the line segment between (2,0) and (0,2) for loopless 2-connected graphs. Chavez-Lomeli et al. proved that this holds for coloopless paving matroids, and I provide a shorter proof of their theorem. I also prove it for minimally 2-edge-connected graphs. As a general statement for the convexity of the Tutte polynomials, I show that the Tutte polynomial of a sparse paving matroid is almost surely convex in the first quadrant. In contrast, I conjecture that the Tutte polynomial of a sparse paving matroid with fixed rank is almost never convex in the first quadrant.<br /><br />The following multiplicative version of the Merino-Welsh conjecture was considered by Noble and Royle:<br /><br />T(G;1,1)2 ≤ T(G;2,0) T(G;0,2).<br /><br />Noble and Royle proved that this multiplicative version holds for series-parallel graphs, using a computer algorithm that they designed. Using a property of the splitting formula which I found, I improve their algorithm so that it is applicable to the class of graphs with bounded treewidth (or pathwidth). As an application, I verify that the multiplicative version holds for graphs with pathwidth at most 3.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>110</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>384</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/aspects-of-the-tutte-polynomial(b7bbd3d4-39f7-49cb-853a-139aae217468).html2014-12-31T23:00:00ZDesign and Analysis of Symmetric Primitives
http://orbit.dtu.dk/en/publications/design-and-analysis-of-symmetric-primitives(7d94a26d-c657-4566-a265-461473931cd1).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/design-and-analysis-of-symmetric-primitives(7d94a26d-c657-4566-a265-461473931cd1).html" class="link"><span>Design and Analysis of Symmetric Primitives</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/martin-mehl-lauridsen(6ff3a3ba-9819-4653-82ea-69d9bc128065).html" class="link person"><span>Lauridsen, M. M.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/christian-rechberger(2b326705-e9ce-4f75-8852-a312f3ed2260).html" class="link person"><span>Rechberger, C.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/lars-ramkilde-knudsen(a62b8acd-d07f-4943-a467-85af0eb5efe0).html" class="link person"><span>Knudsen, L. R.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">213 p.</span> (DTU Compute PHD-2015; No. 382).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">The subject of this thesis is the study of symmetric cryptographic primitives. We investigate these objects from three different perspectives: cryptanalysis, design and implementation aspects. <br /><br />The first part deals with cryptanalysis of symmetric primitives, where one tries to leverage a property of the design to achieve some adversarial goal. Two of the most successful types of cryptanalysis are differential- and linear attacks. We apply variants of differential cryptanalysis to the lightweight block cipher SIMON which was proposed by researchers from the National Security Agency (NSA) in 2013. In particular, we present a search heuristic to find differentials of high probability, and we investigate the clustering of characteristics known as the differential effect. Finally, we apply impossible differential attacks using truncated differentials to a number of SIMON variants. Next, we define a theoretical model for key-less linear distinguishers, which captures the meaning of distinguishing a block cipher from an ideal permutation using linear cryptanalysis, when the key is either known or chosen by the adversary. Such models exist using differential properties but were never before defined using linear cryptanalysis. We apply this model to the standardized block cipher PRESENT. Finally, we present very generic attacks on two authenticated encryption schemes, AVALANCHE and RBS, by pointing out severe design flaws that can be leveraged to fully recover the secret key with very low complexity. <br /><br />In the second part, we delve into the matter of the various aspects of designing a symmetric cryptographic primitive. We start by considering generalizations of the widely acclaimed Advanced Encryption Standard (AES) block cipher. In particular, our focus is on a component operation in the cipher which permutes parts of the input to obtain dependency between the state bits. With this operation in focus, we give a range of theoretical results, reducing the possible choices for the operation in generalized ciphers to a particular set of classes. We then employ a computer-aided optimization technique to determine the best choices for the operation in terms of resistance towards differential- and linear cryptanalysis. Also in the vein of symmetric primitive design we present PRØST, a new and highly secure permutation. Employing existing third-party modes of operation, we present six proposals based on PRØST for the ongoing CAESAR competition for authenticated encryption with associated data. We describe the design criteria, the usage modes and give proofs of security.<br /><br />Finally, in the third part, we consider implementation aspects of symmetric cryptography, with focus on high-performance software. In more detail, we analyze and implement modes recommended by the National Institute of Standards and Technology (NIST), as well as authenticated encryption modes from the CAESAR competition, when instantiated with the AES. The data processed in our benchmarking has sizes representative to that of typical Internet traffic. Motivated by a significant improvement to special AES instructions in the most recent microarchitecture from Intel, codenamed Haswell, our implementations are tailored for this platform. Finally, we introduce the comb scheduler which is a low-overhead look-ahead strategy for processing multiple messages in parallel. We show that it significantly increases the throughput for sequential modes of operation especially, but also for parallel modes to a lesser extent.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>213</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>382</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/design-and-analysis-of-symmetric-primitives(7d94a26d-c657-4566-a265-461473931cd1).html2014-12-31T23:00:00ZDevelopment and Application of Tools for MRI Analysis - A Study on the Effects of Exercise in Patients with Alzheimer's Disease and Generative Models for Bias Field Correction in MR Brain Imaging
http://orbit.dtu.dk/en/publications/development-and-application-of-tools-for-mri-analysis--a-study-on-the-effects-of-exercise-in-patients-with-alzheimers-disease-and-generative-models-for-bias-field-correction-in-mr-brain-imaging(5998e9ca-3c2b-4734-b528-3aed0f79380d).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/development-and-application-of-tools-for-mri-analysis--a-study-on-the-effects-of-exercise-in-patients-with-alzheimers-disease-and-generative-models-for-bias-field-correction-in-mr-brain-imaging(5998e9ca-3c2b-4734-b528-3aed0f79380d).html" class="link"><span>Development and Application of Tools for MRI Analysis - A Study on the Effects of Exercise in Patients with Alzheimer's Disease and Generative Models for Bias Field Correction in MR Brain Imaging</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/christian-thode-larsen(05521ab1-11fa-4bf3-a4d1-a32d0382e79d).html" class="link person"><span>Larsen, C. T.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/koen-van-leemput(265767e4-0af5-4996-acf2-288bf3d27030).html" class="link person"><span>Van Leemput, K.</span></a> <span class="date">2015</span> Kgs. Lyngby: <a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">151 p.</span> (DTU Compute PHD-2015; No. 378).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2015</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Magnetic resonance imaging (MRI) is the de facto modality in neuroimaging studies, due to its superior image contrast in soft tissue. These studies often employ automated software pipelines that segments the image into structures and tissue. This reduces the time needed for analysis as well as statistical bias that may arise due to disagreements in delineations made by human experts. One such pipeline is Freesurfer.<br /><br />This thesis presents results from the intervention study \Preserving cognition, quality of life, physical health and functional ability in Alzheimer’s disease: the effect of physical exercise” (ADEX), where longitudinal Freesurfer analysis was used to obtain segmentations of the hippocampal subfields and cortical regions in a subgroup of participants before and after a four-month exercise period. The participants performed moderate-to-high aerobic exercise for one hour, three times per week. The study hypothesized that the intervention would lead to reduced loss of tissue in the hippocampus and cortical regions, and that volumetric changes over time would correlate with cognitive performance measures. It was not possible to measure any effects in the hippocampus or cortical regions due to the intervention. However, it was found that exercise load (attendance and training intensity) correlated with changes in the hippocampus and in frontal and cingulate cortical thickness. Furthermore, changes in frontal and cingulate cortical thickness were found to correlate with changes in several cognitive performance measures, including mental speed, attention and verbal uency.<br /><br />MRI suffers from an image artifact often referred to as the "bias field”. This effect complicates automatized analysis of the images. For this reason, bias field correction is typical an early preprocessing step in many pipelines. Freesurfer currently employs the popular N3 bias field correction algorithm early in the pipeline, to solve this problem.<br /><br />In this thesis, the reader is introduced to generative models for bias field correction. It is further shown how N3, which has traditionally been described as a "histogram sharpening” method, actually employs an underlying generative model, and that the bias field is estimated using an algorithm that is identical to generalized expectation maximization, but relies on heuristic parameter updates.<br /><br />The thesis progresses to present a new generative model for longitudinal correction of the bias field, as well as a model that does not require brain masking or probabilistic, anatomical atlases in order to perform well. Finally, the thesis presents the realization of these models in the software package "Intensity Inhomogeneity Correction”, which will be made publicly available.<br /></div></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of Publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html" class="link"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>151</td></tr><tr class="status"><th>State</th><td>Submitted - <span class="date">2015</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2015</td></tr><tr><th>Number</th><td>378</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/development-and-application-of-tools-for-mri-analysis--a-study-on-the-effects-of-exercise-in-patients-with-alzheimers-disease-and-generative-models-for-bias-field-correction-in-mr-brain-imaging(5998e9ca-3c2b-4734-b528-3aed0f79380d).html2014-12-31T23:00:00Z