DTU Orbit RSS Feed
http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08)/search/publications.html?page=2
RSS FeedWed, 26 Oct 2016 08:59:23 GMT2016-10-26T08:59:23ZModelling Digital Media Objects
http://orbit.dtu.dk/en/publications/modelling-digital-media-objects(71fd187c-773c-4db4-986f-f5564397db2d).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/modelling-digital-media-objects(71fd187c-773c-4db4-986f-f5564397db2d).html" class="link"><span>Modelling Digital Media Objects</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/rasmus-troelsgaard(db5a2a52-a5ec-46ae-88b5-53085f142cf4).html" class="link person"><span>Troelsgaard, R.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/lars-kai-hansen(3bdd1c4b-6c83-462b-a4cd-1ffbdcbff128).html" class="link person"><span>Hansen, L. K.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/jan-larsen(172b79ba-ccb1-4bea-8cd3-30dfd62f4624).html" class="link person"><span>Larsen, J.</span></a> <span class="date">2016</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">83 p.</span> (DTU Compute PHD-2016; No. 439).<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: 2016</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 goal of this thesis is to investigate two relevant issues regarding computational representation and classification of digital multi-media objects. With a special focus on music, a model for representation of objects comprising multiple heterogeneous data types is investigated. Necessary to this work are considerations regarding integration of multiple diverse data modalities and evaluation of the resulting concept representation.<br /><br />Regarding modelling of data exhibiting certain sequential structure, a number of theoretical and empirical results are presented. These are results related to model parameter estimation and the use of sequence models in a classification scenario. The latter being of importance in various digital multimedia navigation and retrieval tasks.<br /><br />In the fields of topic modelling and multi-modal integration, we formulate a model to describe entities composed of multiple aspects. The particular aspects considered in the publications are sound, song lyrics, and user-provided metadata. This model integrates the diverse data types comprising the objects and defines concrete unified representations in a joint “semantic” space. Within the context of this model, general measures of similarity between such multi-modal objects are investigated.<br /><br />In the fields of method of moments and sequence modelling, we increase practical applicability of a certain moment based parameter estimation method for Hidden Markov models by showing how to use full-length sequences in the estimation process. Consequently, this impacts the quality of the estimated model parameters.<br /><br />Subsequently, we show how to perform time series classification using a composite likelihood formulated from third order moments defined by the Hidden Markov model. Compared to the conventional likelihood based method, our contribution is less computationally expensive, while retaining the level of classification performance.</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>83</td></tr><tr class="status"><th>State</th><td><span class="prefix">Submitted - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2016</td></tr><tr><th>Number</th><td>439</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/modelling-digital-media-objects(71fd187c-773c-4db4-986f-f5564397db2d).html2015-12-31T23:00:00ZRoots of the Chromatic Polynomial
http://orbit.dtu.dk/en/publications/roots-of-the-chromatic-polynomial(31df590e-f634-4bf0-9879-768c1548e968).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/roots-of-the-chromatic-polynomial(31df590e-f634-4bf0-9879-768c1548e968).html" class="link"><span>Roots of the Chromatic Polynomial</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/thomas-perrett(85f59c7c-4146-4bd4-82df-36b251109766).html" class="link person"><span>Perrett, T.</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">2016</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">115 p.</span> (DTU Compute PHD-2016; No. 438).<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: 2016</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 chromatic polynomial of a graph G is a univariate polynomial whose evaluation at any positive integer q enumerates the proper q-colourings of G. It was introduced in connection with the famous four colour theorem but has recently found other applications in the field of statistical physics. In this thesis we study the real roots of the chromatic polynomial, termed chromatic roots, and focus on how certain properties of a graph affect the location of its chromatic roots. <br /><br />Firstly, we investigate how the presence of a certain spanning tree in a graph affects its chromatic roots. In particular we prove a tight lower bound on the smallest non-trivial chromatic root of a graph admitting a spanning tree with at most three leaves. Here, non-trivial means different from 0 or 1. This extends a theorem of Thomassen on graphs with Hamiltonian paths. We also prove similar lower bounds on the chromatic roots of certain minor-closed families of graphs.<br /><br />Later, we study the Tutte polynomial of a graph, which contains the chromatic polynomial as a specialisation. We discuss a technique of Thomassen using which it is possible to deduce that the roots of the chromatic polynomial are dense in certain intervals. We extend Thomassen’s technique to the Tutte polynomial and as a consequence, deduce a density result for roots of the Tutte polynomial. This partially answers a conjecture of Jackson and Sokal. <br /><br />Finally, we refocus our attention on the chromatic polynomial and investigate the density of chromatic roots of several graph families. In particular, we show that the chromatic roots of planar graphs are dense in the interval (3; 4), except for a small interval around _ + 2 _ 3:618, where _ denotes the golden ratio. We also investigate the chromatic roots of related minor-closed classes of graphs and bipartite graphs.<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>115</td></tr><tr class="status"><th>State</th><td><span class="prefix">Submitted - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2016</td></tr><tr><th>Number</th><td>438</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/roots-of-the-chromatic-polynomial(31df590e-f634-4bf0-9879-768c1548e968).html2015-12-31T23:00:00ZOn Practical Sampling of Bidirectional Reflectance
http://orbit.dtu.dk/en/publications/on-practical-sampling-of-bidirectional-reflectance(8765732b-59a5-4826-9206-5d7b19c85962).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/on-practical-sampling-of-bidirectional-reflectance(8765732b-59a5-4826-9206-5d7b19c85962).html" class="link"><span>On Practical Sampling of Bidirectional Reflectance</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/jannik-boll-nielsen(adb04f8e-34ea-4bbd-9325-afc7857afd05).html" class="link person"><span>Nielsen, J. B.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/henrik-aanaes(82129afb-048a-424a-ba22-419bd0bab56f).html" class="link person"><span>Aanæs, H.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/jeppe-revall-frisvad(e7a06e6e-86d5-454f-a78b-c998215c6169).html" class="link person"><span>Frisvad, J. R.</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> <span class="date">2016</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">142 p.</span> (DTU Compute PHD-2016; No. 435).<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: 2016</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Accurate material models are a key part in producing convincing, photo-realistic, images in computer graphics. Elaborate analytical models exist, allowing graphics designers to manually design material appearance. However, given the complex nature and wide variability of material appearance, measuring this from the real world is an impractical and time-consuming process. Having a practical way of measuring material appearance will not only be of great value to the graphics community, but also open up for a wide range of new application areas, including industrial production quality control, digital prototyping and manufacturing, and interactive real-time product visualization. <br /><br />In this thesis, the challenge of making material appearance measurements practical is addressed. Specifically, the Bidirectional Reflectance Distribution Function (BRDF), which is the quantity describing material appearance, is thoroughly analysed using both optimisation tools and multivariate statistics, in search of making BRDFs more accessible.<br /><br />The work demonstrated includes an insight into the challenges of fitting analytical models to measured data and on the compromises one is bound to make when simplifying the real world with a parametric BRDF model. Specifically we identify what error measures work well for obtaining perceputally good results and how a simple BRDF model may be modified to better match real world data. With an offset in this, a linear, data-driven, BRDF model is proposed and a framework for reconstructing full and accurate BRDFs from only a few measurements is presented. It is here demonstrated that with as little as 20 point-samples, a BRDF can accurately be reconstructed. Furthermore utilising the field of view of a camera, this may be reduced to as little as two images. With this, the thesis demonstrates how BRDF measurements can be made practical, and it exemplifies this with a range of datasets intended for various purposes, each including high quality measured BRDFs.<br /><br />Where the classical approach to BRDF capture may take weeks in measurement time, we here successfully demonstrate that is can in fact be reduced to no more than minutes or even seconds using our framework.</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>142</td></tr><tr class="status"><th>State</th><td><span class="prefix">Submitted - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2016</td></tr><tr><th>Number</th><td>435</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/on-practical-sampling-of-bidirectional-reflectance(8765732b-59a5-4826-9206-5d7b19c85962).html2015-12-31T23:00:00ZFPGA Acceleration by Dynamically-Loaded Hardware Libraries
http://orbit.dtu.dk/en/publications/fpga-acceleration-by-dynamicallyloaded-hardware-libraries(9b55d27b-8123-4852-96b0-808be8961b60).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/fpga-acceleration-by-dynamicallyloaded-hardware-libraries(9b55d27b-8123-4852-96b0-808be8961b60).html" class="link"><span>FPGA Acceleration by Dynamically-Loaded Hardware Libraries</span></a></h2>Lomuscio, A., <a rel="Person" href="http://orbit.dtu.dk/en/persons/alberto-nannarelli(e1a0a7a2-7f9d-4a0b-bae5-b24f2fc93d2b).html" class="link person"><span>Nannarelli, A.</span></a> & Re, M. <span class="date">2016</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">19 p.</span> (DTU Compute-Technical Report-2016; No. 3).<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">Report – Annual report year: 2016</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Hardware acceleration is a viable solution to obtain energy efficiency in data intensive computation.<br /><br />In this work, we present a hardware framework to dynamically load hardware libraries, HLL, on reconfigurable platforms (FPGAs). Provided a library of application-specific processors, we load on-the-y the speciffic processor in the FPGA, and we transfer the execution from the CPU to the FPGA-based accelerator. <br /><br />Results show that significant speed-up and energy efficiency can be obtained by HLL acceleration on system-on-chips where reconfigurable fabric is placed next to the CPUs.</div></div><div class="scientificreport"><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>19</td></tr><tr class="status"><th>State</th><td><span class="prefix">Published - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute-Technical Report-2016</td></tr><tr><th>Number</th><td>3</td></tr><tr><th>ISSN (print)</th><td>1601-2321</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/fpga-acceleration-by-dynamicallyloaded-hardware-libraries(9b55d27b-8123-4852-96b0-808be8961b60).html2015-12-31T23:00:00ZGabor frames on locally compact abelian groups and related topics
http://orbit.dtu.dk/en/publications/gabor-frames-on-locally-compact-abelian-groups-and-related-topics(39151a48-c470-40a3-bec8-abe7d10419a2).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/gabor-frames-on-locally-compact-abelian-groups-and-related-topics(39151a48-c470-40a3-bec8-abe7d10419a2).html" class="link"><span>Gabor frames on locally compact abelian groups and related topics</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/mads-sielemann-jakobsen(a8285507-1522-444f-a01d-a8f379fbca28).html" class="link person"><span>Jakobsen, M. S.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/ole-christensen(5d530d48-deb4-4634-b4c0-a52fd0a529f8).html" class="link person"><span>Christensen, O.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/jakob-lemvig(ea2e1ec1-1c79-46b8-aeba-2bbdafeafe99).html" class="link person"><span>Lemvig, J.</span></a> <span class="date">2016</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">26 p.</span> (DTU Compute PHD-2016; No. 436).<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: 2016</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 consists of four papers. The first one introduces generalized translation invariant systems and considers their frame properties, the second and third paper give new results on the theory of Gabor frames, and the fourth is a review paper with proofs and new results on the Feichtinger algebra.<br /><br />The generalized translation invariant (GTI) systems provide, for the first time, a framework which can describe frame properties of both discrete and continuous systems. The results yield the well-known characterizations of dual frame pairs and Parseval frames of Gabor-, wavelet-, curvelet- and shearlet-type and for (generalized) shift-invariant systems and their continuous formulations.<br /><br />This thesis advances the theory of both separable and non-separable, discrete, semicontinuous and continuous Gabor systems. In particular, the well established structure theory for separable lattice Gabor frames is extended and generalized significantly to Gabor systems with time-frequency shifts along closed subgroups in the time-frequency plane. This includes density results, the Walnut representation, the Wexler-Raz biorthogonality relations, the Bessel duality and the duality principle between Gabor frames and Gabor Riesz bases.<br /><br />The theory of GTI systems and Gabor frames in this thesis is developed and presented in the setting of locally compact abelian groups, however, even in the euclidean setting the results given here improve the existing theory.<br /><br />Finally, the thesis contains a review paper with proofs of all the major results on the Banach space of functions known as the Feichtinger algebra. This includes many of its different characterizations and treatment of its many equivalent norms, its minimality among all time-frequency shift invariant Banach spaces and aspects of its dual space, operators on the space and the kernel theorem for the Feichtinger algebra. The work also includes new findings such as a characterization among all Banach spaces, a forgotten theorem by Reiter on Banach space isomorphisms of the Feichtinger algebra, and new useful inequalities.</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>26</td></tr><tr class="status"><th>State</th><td><span class="prefix">Submitted - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2016</td></tr><tr><th>Number</th><td>436</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/gabor-frames-on-locally-compact-abelian-groups-and-related-topics(39151a48-c470-40a3-bec8-abe7d10419a2).html2015-12-31T23:00:00ZHigh Performance with Prescriptive Optimization and Debugging
http://orbit.dtu.dk/en/publications/high-performance-with-prescriptive-optimization-and-debugging(f2ea9bb8-d850-4d88-938c-5d63319964d9).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/high-performance-with-prescriptive-optimization-and-debugging(f2ea9bb8-d850-4d88-938c-5d63319964d9).html" class="link"><span>High Performance with Prescriptive Optimization and Debugging</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/nicklas-bo-jensen(1539cf5e-c1e6-4a98-98bb-1fc8e7d407b2).html" class="link person"><span>Jensen, N. B.</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">2016</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-2016; No. 437).<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: 2016</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Parallel programming is the dominant approach to achieve high performance in computing today. Correctly writing efficient and fast parallel programs is a big challenge mostly carried out by experts. We investigate optimization and debugging of parallel programs.<br /><br />We argue that automatic parallelization and automatic vectorization is attractive as it transparently optimizes programs. The thesis contributes an improved dependence analysis for explicitly parallel programs. These improvements lead to more loops being vectorized, on average we achieve a speedup of 1.46 over the existing dependence analysis and vectorizer in GCC. <br /><br />Automatic optimizations often fail for theoretical and practical reasons. When they fail we argue that a hybrid approach can be effective. Using compiler feedback, we propose to use the programmer’s intuition and insight to achieve high performance. Compiler feedback enlightens the programmer why a given optimization was not applied, and suggest how to change the source code to make it more amenable to optimizations. We show how this can yield significant speedups and achieve 2.4 faster execution on a real industrial use case. <br /><br />To aid in parallel debugging we propose the prescriptive debugging model, which is a user-guided model that allows the programmer to use his intuition to diagnose bugs in parallel programs. The model is scalable, yet capable enough, to be general-purpose. In our evaluation we demonstrate low run time overhead and logarithmic scalability. This enable the model to be used on extremely large parallel systems.</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><span class="prefix">Submitted - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2016</td></tr><tr><th>Number</th><td>437</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/high-performance-with-prescriptive-optimization-and-debugging(f2ea9bb8-d850-4d88-938c-5d63319964d9).html2015-12-31T23:00:00ZModel Stickiness in Spray Drying
http://orbit.dtu.dk/en/publications/model-stickiness-in-spray-drying(fd944869-f19b-400c-ba5a-6e0c45d78c79).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/model-stickiness-in-spray-drying(fd944869-f19b-400c-ba5a-6e0c45d78c79).html" class="link"><span>Model Stickiness in Spray Drying</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/thomas-petersen(a0603943-26c7-42ee-9ccd-bba181353f85).html" class="link person"><span>Petersen, T.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/ole-hassager(b3b2f7bb-1e38-4937-ba11-7f4d63f0012d).html" class="link person"><span>Hassager, O.</span></a>, <a rel="Person" href="http://orbit.dtu.dk/en/persons/jakob-sloth(d5a03b4c-2dc1-4fad-b633-58128246f419).html" class="link person"><span>Sloth, J.</span></a> & <a rel="Person" href="http://orbit.dtu.dk/en/persons/thorvald-uhrskov-ullum(c0ddc42e-d7e3-49a8-b7af-f6731018443b).html" class="link person"><span>Ullum, T. U.</span></a> <span class="date">2015</span> Kgs. Kyngby: <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">103 p.</span><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: 2016</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 work presented in this thesis concerns wall deposits encountered in spray drying caused by products that exhibit so-called 'stickiness'. The thesis delves into the understanding of the phenomenon of sticky wall deposits in spray drying and proposes a simple criterion for use in industrial design of spray dryers.<br />The experimental work centers around a new technique for measuring when, during drying, a particle becomes non-sticky based on a single droplet drying technique used to study drying kinetics. An acoustic levitator is used to dry a levitated droplet in conditions similar to those a droplet would encounter in a spray dryer. The droplet is recorded using a CCD-camera during drying and the subsequent stickiness test. After a user-specifed drying time a piston strikes the partially dried particle at a user-specifed velocity. After the impact the piston surface is inspected and if the particle was sticky it is seen adhering to the surface, while a clean piston signifes a non-sticky particle. The setup was designed specically to test the stickiness of a particle produced by drying a droplet of the desired feed - something unlike methods of literature where dry particles have been humidied before tests. The setup allowed for parameter variations in the temperature and humidity of the drying air, impact velocity, piston surface and more.<br />Results of measuring the stickiness of skim milk is reported for varying impact velocity, drying temperature and relative humidity. It is found that normalizing the critical drying time to get non-sticky particles with the initial diameter squared leads to a single value for a given set of parameters, if the initial diameter is in a limited range. The normalized critical drying time was found to increase linearly when increasing the relative humidity ofthe drying air. Furthermore, the dependency was the same independently of temperature.The drying time appeared to decrease linearly with increasing temperature, although with a smaller dependence. Measurements with increasing impact velocity showed that the requireddrying time increased linearly. This nding is opposite of what is typically reportedin literature and it is an important part of the hypothesis for stickiness developed here .Finally, measurements with maltodextrin (with dextrose equivalent of 18) are reported for varying relative humidity and impact velocity and the same trends were shown as those found for skim milk. <br />Replacing the CCD-camera with a high-speed camera allows a user to produce videos of the impact of a sticky particle with a wall in very high temporal resolution. This was done for skim milk powder for varying impact velocity, humidity and piston surface material.The results were mostly qualitative, with a little quantitative analysis where possible.The first observation was that there was very little dierence between a particle that just adhered (sticky) and one that just bounced (non-sticky). Both were forced to make some contact with the piston upon impact, deformed only slightly and then typically moved slightly away before either stopping with contact (sticky) or bouncing with no contact (non-sticky). Sticky particles had a large apparent contact angle, similar to what would be expected for a liquid with poor wetting properties. The velocity did not seem to change this much, although slight deformation was seen when the impact velocity was at the highest used values. The phenomenon did not appear to change noticeable when the droplet was dried in a high relative humidity environment. The qualitative difference observable between Teon and stainless steel was very limited. On Stainless steel the droplet seemed to wet slightly more after the initial impact while the contact area was constant for Teon. Modelling work was carried out to help understand the phenomenon, but also to investigate how the impact scaled for particle size. This was done using the Level Set Method implemented in the Finite Element based COMSOL Microuidics software. Using two level functions allowed for the denition of density and viscosity functions which were different throughout the particle and different from the surrounding air. This was used to model an in homogeneous droplet which consisted of a skin with high viscosity and a core with lower viscosity. The surrounding air had an even lower viscosity and a lower density.The droplet was modelled without elastic properties. The simulations were initiated withthe condition that the droplet was moving towards the wall with a predened velocity.The simulation was run while individually varying initial droplet velocity, viscosity of skin, core and air, density of droplet and air, surface tension, droplet radius and the radius ofthe core relative to the droplet. A parameter analogous to the radius of the contact areawas dened and the dependency of this parameter upon the ones listed above was mapped.The radius of spreading, normalized with the droplet radius, correlated with the squareroot of the Reynolds' number (based on material properties of the skin) multiplied withthe volume of the droplet divided by the volume of the skin. A simple analytical model was used to show that this dependence on the Reynolds' number could be a result ofviscous dissipation of kinetic energy in a zone near the movement of the triple line.<br />The observations made in the experimental and modelling sections were combined to propose a simple criterion for use in conjunction with other tools for design or possible control of spray dryers. A single droplet drying technique is to be used to obtain a characteristic drying curve for the product of interest. The linear dependencies found is then used to reduce the necessary number of stickiness measurements to as little as three (although more measurements increase accuracy). These data are then used to produce a simple criterion which gives a critical residual moisture content under which a particle must be for it to be non-sticky as depending on the relative humidity in which it is dried, the initial radius of the droplet and the velocity with which it impacts the wall.<br />Finally, the hypothesis for stickiness that this work leads to is summarized as follows. Asa droplet is dried it forms a region near the surface in which the solvent concentrationdecreases and the viscosity increases. As it impacts a wall in the spray dryer it does sowith some kinetic energy. The kinetic energy forces the wet particle to deform and wetthe surface, while energy is dissipated in the viscous ow. If the viscosity in the surface region is high enough and the region is large enough the energy is dissipated with verylittle movement which means the contact area is small. A small contact area means a low energy of adhesion and the particle is easily removed by other eects. If the energy is not dissipated rapidly enough however a larger contact area is established and the particle will adhere strongly. This hypothesis suggests that the surface properties of the sticky particleare not the only effects that matter and therefore that measuring rehumidied particles is not the same as dried droplets. Furthermore it has the important dierence from the established hypothesis that the contact is forced and not spontaneous.</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. Kyngby</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>103</td></tr><tr class="status"><th>State</th><td><span class="prefix">Published - </span><span class="date">2015</span></td></tr></tbody></table></div></div></div>Wed, 31 Dec 2014 23:00:00 GMThttp://orbit.dtu.dk/en/publications/model-stickiness-in-spray-drying(fd944869-f19b-400c-ba5a-6e0c45d78c79).html2014-12-31T23:00:00ZThe Dynamics of a Railway Vehicle on a Disturbed Track: Modelling of lateral irregularities
http://orbit.dtu.dk/en/publications/the-dynamics-of-a-railway-vehicle-on-a-disturbed-track(6d24daac-617f-41f6-aaba-16dd18bfb16e).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/the-dynamics-of-a-railway-vehicle-on-a-disturbed-track(6d24daac-617f-41f6-aaba-16dd18bfb16e).html" class="link"><span>The Dynamics of a Railway Vehicle on a Disturbed Track: Modelling of lateral irregularities</span></a></h2><a rel="Person" href="http://orbit.dtu.dk/en/persons/lasse-engbo-christiansen(9d379f06-f620-404a-8397-3dfdaf9a3804).html" class="link person"><span>Christiansen, L. E.</span></a> <span class="date">2001</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">122 p.</span><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">Report – Annual report year: 2001</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 work studies a mathematical model of a bogie, the aim being to investigate the correlation between the lateral motion of the bogie and lateral irregularities of the track that the bogie is running along. When modelling the wheel-rail forces real wheel and rail profiles are used. One of the main differences between this work and previous ones is that the vertical degrees of freedom are included making the model very realistic as this allows a dynamic calculation of the penetration depth in the contact area.<br /><br />In chapter two the model of the bogie is described, this is followed by a chapter describing how the wheel-rail interaction is modelled. The fourth chapter gives a presentation of the numerical algorithms used in this work.<br /><br />The fifth chapter presents the results obtained during this work, it begins with an brief investigation of the dynamics when running on a straight track. The second section introduces some statistical measures and treats centerline irregularities when running with a fixed speed of 30 m/s. The following section investigates the behaviour of the model when the track possesses gauge irregularities. The fourth and fifth sections describe the changes observed when changing the velocity of the railway vehicle and the phase between the sinusoidal irregularities of the rails. In the final section the eigenfrequencies of the bogie are found and the most important modes are described. Chapter six is a discussion of the obtained results, and these are related to the possibility of measuring track irregularities using accelerometers mounted on the bogie. Finally, some appendixes are included bringing some details of the derivations related to the wheel-rail interaction and details about the used pro files.</div></div><div class="scientificreport"><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>122</td></tr><tr class="status"><th>State</th><td><span class="prefix">Published - </span><span class="date">2001</span></td></tr></tbody></table></div><h3 class="subheader">Bibliographical note</h3><p>This is a Master of Science Thesis</p></div></div>Sun, 31 Dec 2000 23:00:00 GMThttp://orbit.dtu.dk/en/publications/the-dynamics-of-a-railway-vehicle-on-a-disturbed-track(6d24daac-617f-41f6-aaba-16dd18bfb16e).html2000-12-31T23:00:00ZCauchy Noise Removal by Nonconvex ADMM with Convergence Gaurantees
http://orbit.dtu.dk/en/publications/cauchy-noise-removal-by-nonconvex-admm-with-convergence-gaurantees(9353ee8b-835f-4063-9e72-c31893f7c4d0).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/cauchy-noise-removal-by-nonconvex-admm-with-convergence-gaurantees(9353ee8b-835f-4063-9e72-c31893f7c4d0).html" class="link"><span>Cauchy Noise Removal by Nonconvex ADMM with Convergence Gaurantees</span></a></h2>Mei, J-J., <a rel="Person" href="http://orbit.dtu.dk/en/persons/yiqiu-dong(fbd1dafa-2d1c-4149-85b8-8998ea308e5f).html" class="link person"><span>Dong, Y.</span></a>, Huang, T-Z. & Yin, W. <span class="date">2016</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">24 p.</span> (DTU Compute-Technical Report-2016; No. 10).<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">Report – Annual report year: 2016</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">Image restoration is one of the most important and essential issues in image processing. Cauchy noise in engineering application has the non-Gaussian and impulsive property. In order to preserve edges and details of images, the total variation (TV) based variational model has been studied for restoring images degraded by blur and Cauchy noise. Due to the nonconvexity and nonsmoothness, there exist computational and theoretical challenges. In this paper, adapting recent results, we develop an alternating direction method of multiplier (ADMM) in spite of the challenges. The convergence to a stationary point is guaranteed theoretically under certain conditions. Experimental results demonstrate that the proposed method is competitive with other methods in terms of visual and quantitative measures. Especially, by comparing to the PSNR values, our method can improve about 0.5dB on average.</div></div><div class="scientificreport"><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>24</td></tr><tr class="status"><th>State</th><td><span class="prefix">Published - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute-Technical Report-2016</td></tr><tr><th>Number</th><td>10</td></tr><tr><th>ISSN (print)</th><td>1601-2321</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/cauchy-noise-removal-by-nonconvex-admm-with-convergence-gaurantees(9353ee8b-835f-4063-9e72-c31893f7c4d0).html2015-12-31T23:00:00ZContrast Invariant SNR
http://orbit.dtu.dk/en/publications/contrast-invariant-snr(597f7317-fb88-4f7b-b736-9e29c8c71dae).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/contrast-invariant-snr(597f7317-fb88-4f7b-b736-9e29c8c71dae).html" class="link"><span>Contrast Invariant SNR</span></a></h2>Weiss, P., Escande, P. & <a rel="Person" href="http://orbit.dtu.dk/en/persons/yiqiu-dong(fbd1dafa-2d1c-4149-85b8-8998ea308e5f).html" class="link person"><span>Dong, Y.</span></a> <span class="date">2016</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">8 p.</span> (DTU Compute-Technical Report-2016; No. 9).<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">Report – Annual report year: 2016</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="abstract"><div class="textblock">We design an image quality measure independent of local contrast changes, which constitute simple models of illumination changes. Given two images, the algorithm provides the image closest to the first one with the component tree of the second. This problem can be cast as a specific convex program called isotonic regression. We provide a few analytic properties of the solutions to this problem. We also design a tailored first order optimization procedure together with a full complexity analysis. The proposed method turns out to be practically more efficient and reliable than the best existing algorithms based on interior point methods. The algorithm has potential applications in change detection, color image processing or image fusion. A Matlab implementation is available at http://www.math.univ-toulouse.fr/_weiss/PageCodes.html.</div></div><div class="scientificreport"><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>8</td></tr><tr class="status"><th>State</th><td><span class="prefix">Published - </span><span class="date">2016</span></td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute-Technical Report-2016</td></tr><tr><th>Number</th><td>9</td></tr><tr><th>ISSN (print)</th><td>1601-2321</td></tr></tbody></table></div></div>Thu, 31 Dec 2015 23:00:00 GMThttp://orbit.dtu.dk/en/publications/contrast-invariant-snr(597f7317-fb88-4f7b-b736-9e29c8c71dae).html2015-12-31T23:00:00Z