Pun-Mo Ho. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (leaves 90-95). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Background --- p.1 / Chapter 1.2 --- Existing Approaches --- p.2 / Chapter 1.3 --- Our Approach --- p.3 / Chapter 1.4 --- Structure of the Thesis --- p.4 / Chapter 2 --- Related Work --- p.5 / Chapter 2.1 --- Compression for Navigation --- p.5 / Chapter 2.1.1 --- Light Field/Lumigraph --- p.5 / Chapter 2.1.2 --- Surface Light Field --- p.6 / Chapter 2.1.3 --- Concentric Mosaics --- p.6 / Chapter 2.1.4 --- On the Compression --- p.7 / Chapter 2.2 --- Compression for Relighting --- p.7 / Chapter 2.2.1 --- Previous Approaches --- p.7 / Chapter 2.2.2 --- Our Approach --- p.8 / Chapter 3 --- Image-Based Relighting --- p.9 / Chapter 3.1 --- Plenoptic Illumination Function --- p.9 / Chapter 3.2 --- Sampling and Relighting --- p.11 / Chapter 3.3 --- Overview --- p.13 / Chapter 3.3.1 --- Codec Overview --- p.13 / Chapter 3.3.2 --- Image Acquisition --- p.15 / Chapter 3.3.3 --- Experiment Data Sets --- p.16 / Chapter 4 --- Data Preparation --- p.18 / Chapter 4.1 --- Block Division --- p.18 / Chapter 4.2 --- Color Model --- p.23 / Chapter 4.3 --- Mean Extraction --- p.24 / Chapter 5 --- Principal Component Analysis --- p.29 / Chapter 5.1 --- Overview --- p.29 / Chapter 5.2 --- Singular Value Decomposition --- p.30 / Chapter 5.3 --- Dimensionality Reduction --- p.34 / Chapter 5.4 --- Evaluation --- p.37 / Chapter 6 --- Eigenimage Coding --- p.39 / Chapter 6.1 --- Transform Coding --- p.39 / Chapter 6.1.1 --- Discrete Cosine Transform --- p.40 / Chapter 6.1.2 --- Discrete Wavelet Transform --- p.47 / Chapter 6.2 --- Evaluation --- p.49 / Chapter 6.2.1 --- Statistical Evaluation --- p.49 / Chapter 6.2.2 --- Visual Evaluation --- p.52 / Chapter 7 --- Relighting Coefficient Coding --- p.57 / Chapter 7.1 --- Quantization and Bit Allocation --- p.57 / Chapter 7.2 --- Evaluation --- p.62 / Chapter 7.2.1 --- Statistical Evaluation --- p.62 / Chapter 7.2.2 --- Visual Evaluation --- p.62 / Chapter 8 --- Relighting --- p.65 / Chapter 8.1 --- Overview --- p.66 / Chapter 8.2 --- First-Phase Decoding --- p.66 / Chapter 8.3 --- Second-Phase Decoding --- p.68 / Chapter 8.3.1 --- Software Relighting --- p.68 / Chapter 8.3.2 --- Hardware-Assisted Relighting --- p.71 / Chapter 9 --- Overall Evaluation --- p.81 / Chapter 9.1 --- Compression of IAIs --- p.81 / Chapter 9.1.1 --- Statistical Evaluation --- p.81 / Chapter 9.1.2 --- Visual Evaluation --- p.86 / Chapter 9.2 --- Hardware-Assisted Relighting --- p.86 / Chapter 10 --- Conclusion --- p.89 / Bibliography --- p.90

Cette thèse présente l’étude analytique et expérimentale d’un système de compression axial constitue d’un compresseur, d’un plenum et d’une vanne de régulation du débit. Fondée sur le modèle de Moore-Greitzer, une approche analytique est utilisée par la suite pour interpréter les données expérimentales. Les analyses linéaires et non linéaires sont présentées, ainsi que les résultats obtenus par intégration numérique des équations du modèle non linéaire complet. Un modèle théorique simple pour les modes acoustiques observes est aussi présente. Dans les expériences, les mesures de pression ont été réalisées pour différents taux de rotation et différents vannages. Ceci a permis la détermination de l’augmentation de pression au travers du compresseur et du débit en utilisant un Venturi, ceux qui fournissent la courbe de fonctionnement du compresseur. Les fluctuations de pression sont mesurées par quatre microphones places en amont du compresseur. Nous n’avons pas observé de pompage, mais un décrochage tournant s’est produit quand le système était suffisamment vanne. Avant ceci, des fluctuations correspondant aux modes acoustiques de la conduite ont été trouvées. Au début du décrochage, une cellule de décrochage tournante d’amplitude croissante a été observée. Cependant, la cellule a rapidement éclaté, remplacée par des fluctuations aléatoires. Ainsi, pour le système de compression considéré, la situation de décrochage est mieux décrite comme étant aléatoire plutôt que rotative. Le traitement du signal (spectres en fréquence ainsi que les autocorrélations et corrélations croisées) a permis d’analyser les fluctuations de pression des modes acoustiques et du décrochage développé. Les modes acoustiques montrent des pics spectraux proches des valeurs prédites par la théorie. En présence de décrochage, le spectre est à large bande, et contient un pic de fréquence basse (∼20 Hz) suivi d’une queue s’étendant jusqu’à 1 kHz, bien au-dessus de la fréquence de rotation du rotor (∼150 Hz). Il apparait une gamme fréquentielle située entre le pic à basse fréquence et la chute à haute fréquence, dans laquelle le spectre évolue en loi de puissance. Les autocorrélations et corrélations croisées entre les différents microphones montrent des oscillations à ∼20 Hz. En appliquant un filtre passe-bas aux données, les fonctions de corrélations croisées des signaux filtres des différents microphones suggèrent l’existence d’une cellule rotative qui effectue une rotation complète en ∼0.05 s. Ceci pourrait expliquer le pic spectral a ∼20 Hz et les oscillations des fonctions de corrélation. De ce fait, les fluctuations de pression apparaissent comme contenant une forte composante aléatoire à haute fréquence, et une cellule rotative. La décorrelation du signal filtre quand la séparation en temps augmente indique que la cellule elle-même a un caractère aléatoire, plutôt qu’une forme et une vitesse rotative fixes comme dans une cellule de décrochage classique. / This thesis presents an analytical and experimental study of an axial compression system consisting of a compressor, plenum and throttle. The analysis is based on the Moore-Greitzer model, the results being later employed to interpret the experimental ones. Linear and weakly nonlinear analyses are presented, as are some results obtained by numerical integration of the fully nonlinear model equations. A simple theoretical model of the experimentally observed acoustic modes is also presented. In the experiments, pressure measurements were carried out for different rotation rates and throttle settings. This allowed the determination of the pressure rise across the compressor and the flow rate using a Venturi, yielding the compressor characteristic function. Pressure fluctuations were measured using four microphones placed upstream of the compressor. We did not observe surge, but rotating stall occurred when the system was sufficiently throttled. Prior to stall, fluctuations corresponding to acoustic duct modes were found. At stall onset, a rotating stall cell of growing amplitude was observed. However, the cell rapidly broke down and gave way to random fluctuations. Thus, for the given compression system, developed stall is perhaps better described as random, rather than rotating. Signal processing (frequency spectra, as well as auto- and cross-correlations) was used to analyse the pressure fluctuations of the acoustic modes and developed stall. The acoustic modes give spectral peaks located close to the expected theoretical values. In the presence of stall, the spectrum is broadband, having a low frequency (∼20 Hz) peak followed by a tail which extends up to ∼1 kHz, well above the rotation frequency (∼150 Hz) of the rotor. There appears to be a frequency range between the low-frequency peak and the high-frequency fall-off in which the spectrum approximates a power law. The autocorrelations and cross-correlations between different microphones show ∼20 Hz oscillations. Low pass filtering the data, the cross-correlation functions of the filtered signals of different microphones suggest a rotating cell which takes ∼0.05 s for a complete rotation. This could explain the ∼20 Hz spectral peak and oscillations of the correlation functions. Thus, the pressure fluctuations appear to consist of a strong high-frequency, random component and a rotating cell. Decorrelation of the filtered signal as separation time increases indicates that, rather than maintaining form and rotational velocity like a classical stall cell, the cell itself exhibits randomness.

Cette thèse, commencée fin 1989, est consacrée à l'étude de l'influence pratique de nouvelles propriétés mathématiques apportées par la théorie des {\em ondelettes\/} dans le domaine du codage en sous-bandes, en vue d'application à la compression d'images fixes. On montre, par une analyse théorique de la décomposition multi-résolution de signaux à temps discret, que l'apport pratique essentiel de la théorie des ondelettes, dans ce cadre, est la propriété de {\em régularité\/} des bancs de filtres itérés en octaves utilisés pour le codage en sous-bandes. On développe ensuite un certain nombre d'outils, qui vont rendre possible une étude exhaustive du rôle de la régularité pour des applications en compression d'images: Une étude mathématique détaillée de la notion de régularité est menée; elle permet d'obtenir des algorithmes d'estimation optimale de régularité. Ensuite, on développe plusieurs méthodes de calcul de bancs de filtres, permettant de réaliser des bons compromis entre la régularité et les autres propriétés des filtres généralement considérées comme utiles pour le codage d'images (orthogonalité, phase linéaire, sélectivité en fréquence, etc.). On propose également des algorithmes rapides permettant, à peu de frais, de réduire notablement la charge de calcul nécessaire à la réalisation d'un banc de filtres itéré en octaves. Finalement, on mène une étude expérimentale du rôle des propriétés des filtres calculés, pour un schéma simple de compression d'images fixes, où la transformée en ondelettes est séparable. Dans le cadre restreint choisi, on met en avant l'intérêt potentiel du critère de régularité par rapport à ceux de sélectivité en fréquence et de phase linéaire.

Afin d'améliorer la qualité de service offerte aux abonnés, il est nécessaire de faire évoluer les réseaux d'accès GPRS/EDGE. Cette thèse se propose d'étudier différentes problématiques qui visent à améliorer les débits offerts aux utilisateurs et à mettre en oeuvre un mécanisme de handover qui réponde à leurs besoins. La première partie de cette thèse étudie différents mécanismes d'allocation dynamique de ressources sur l'interface Abis. Ces mécanismes visent en particulier à permettre le déploiement de la technologie E-GPRS tout en préservant la structure de l'interface Abis existante, qui s'appuie sur une structure de trame MIC. La seconde partie de cette thèse analyse différentes approches de handover qui peuvent être mises en oeuvre pour assurer la mobilité dans les réseaux E-GPRS. Nous y formulons quelques propositions pour améliorer les performances du basculement et mettre en place un véritable handover. Nous présentons les évolutions récentes de la normalisation, puis exposons les résultats de nos simulations qui permettent de comparer les performances des différents mécanismes. Notre étude montre les gains de performances obtenus par la préservation des états de transmission au niveau RLC. La troisième partie de cette thèse considère le problème des handover inter-systèmes. Nous y avons analysé le basculement entre une station de base E-GPRS et un point d'accès WIFI intégré au réseau d'accès mobile. Nous avons analysé l'impact du passage d'une pile protocolaire à une autre, ainsi que de la rupture de débit. Nos résultats montrent que l'introduction d'une couche de convergence au niveau liaison de données permet de limiter l'impact du handover au niveau réseau d'accès. Nous analysons, par ailleurs, les mécanismes de handover à mettre en place pour le transfert de données en mode Streaming. Nous démontrons l'intérêt de l'activation de la couche de liaison de données dans le cas de transferts avec le protocole UDP. La quatrième partie étudie les mécanismes de compression qui peuvent être mis en oeuvre pour réduire la taille des messages de signalisation SIP. Cette étude vise à réduire le temps de transmission de la signalisation - à travers des bearer bas débit - et à économiser l'utilisation des ressources radio.

<p>Safe winter navigation is a hot topic. Not only is the traffic density increasing but theenvironmental considerations are also getting bigger. An oil leakage from a big oiltanker can be of catastrophic proportions in the wrong area and more trafficincreases the risk of an accident. A project that aims for safer winter navigation isSafeWIN. The aim of this project is to develop a forecasting system for compressiveice and thus make winter navigation safer.This thesis is part of above mentioned project and aims to investigate what influenceice compression and ice class has on winter navigation. Vessels are exclusivelyAFRAMAX size tankers sailing on Primorsk in the Gulf of Finland during 2006. Transitdata comes from AIS tracks recorded by the Swedish Maritime Administration. Adatabase with tanker transits has been created and this information is the source forthe studies in this thesis. Included in the database are wind data, ice particulars andtransit information such as speed, and time at different activities during the transit.Average values for a transit has been investigated for comparison and to get a pictureof an average transit.Velocity, waiting time and time with assisting icebreaker are parameters that arebelieved to show how a tanker performs in winter navigation. These parameters arecompared with ice compression and ice class separately to see if there is acorrelation. Ice compression has also been investigated for correlation towards windforce to see if stronger wind generates stronger compression.Using the velocity in different ice compressions an estimate of ice resistance that stemfrom ice compression has been extracted by means of Lindqvist’s formula.</p>

A coronary angiogram (CATH) is the diagnostic tool used to visualize the coronary arteries of a persons heart. These arteries can be accessed through the radial artery. Various methods of applying compression to the radial puncture site have been used, but no research has been done to show what best practice is. The purpose of this pilot study was to compare two devices and three methods for achieving hemostasis after a transradial angiogram while assessing vascular complications and time endpoints. A mechanical device (Terumo TR wristband) and a hydrophilic wound dressing (Clo-Sur P.A.D.) were used. The Terumo band was studied twice, using the current method and a fast-release method. Taking into account the small sample size of this pilot study, statistically significant differences are seen in time to discharge in the fast-release Terumo and Clo-Sur P.A.D. groups, as compared with the control Terumo group, without increasing vascular complications.

The ability to produce potable water economically is the primary purpose of seawater desalination research. Reverse osmosis (RO) and multi-stage flash (MSF) cost more than potable water produced from fresh water resources. As an alternative to RO and MSF, this research investigates a high-efficiency mechanical vapor-compression distillation system that employs an improved water flow arrangement. The incoming salt concentration was 0.15% salt for brackish water and 3.5% salt for seawater, whereas the outgoing salt concentration was 1.5% and 7%, respectively. Distillation was performed at 439 K (331oF) and 722 kPa (105 psia) for both brackish water feed and seawater feed. Water costs of the various conditions were calculated for brackish water and seawater feeds using optimum conditions considered as 25 and 20 stages, respectively. For brackish water at a temperature difference of 0.96 K (1.73oF), the energy requirement is 2.0 kWh/m3 (7.53 kWh/kgal). At this condition, the estimated water cost is $0.39/m3 ($1.48/kgal) achieved with 10,000,000 gal/day distillate, 30-year bond, 5% interest rate, and $0.05/kWh electricity. For seawater at a temperature difference of 0.44 K (0.80oF), the energy requirement is 3.97 kWh/m3 (15.0 kWh/kgal) and the estimated water cost is $0.61/m3 ($2.31/kgal). Greater efficiency of the vapor compression system is achieved by connecting multiple evaporators in series, rather than the traditional parallel arrangement. The efficiency results from the gradual increase of salinity in each stage of the series arrangement in comparison to parallel. Calculations using various temperature differences between boiling brine and condensing steam show the series arrangement has the greatest improvement at lower temperature differences. The following table shows the improvement of a series flow arrangement compared to parallel: ?T (K) Improvement (%)*1.111 2.222 3.333 15.21 10.80 8.37 * Incoming salt concentration: 3.5% Outgoing salt concentration: 7% Temperature: 450 K (350oF) Pressure: 928 kPa (120 psig) Stages: 4

Safe winter navigation is a hot topic. Not only is the traffic density increasing but theenvironmental considerations are also getting bigger. An oil leakage from a big oiltanker can be of catastrophic proportions in the wrong area and more trafficincreases the risk of an accident. A project that aims for safer winter navigation isSafeWIN. The aim of this project is to develop a forecasting system for compressiveice and thus make winter navigation safer.This thesis is part of above mentioned project and aims to investigate what influenceice compression and ice class has on winter navigation. Vessels are exclusivelyAFRAMAX size tankers sailing on Primorsk in the Gulf of Finland during 2006. Transitdata comes from AIS tracks recorded by the Swedish Maritime Administration. Adatabase with tanker transits has been created and this information is the source forthe studies in this thesis. Included in the database are wind data, ice particulars andtransit information such as speed, and time at different activities during the transit.Average values for a transit has been investigated for comparison and to get a pictureof an average transit.Velocity, waiting time and time with assisting icebreaker are parameters that arebelieved to show how a tanker performs in winter navigation. These parameters arecompared with ice compression and ice class separately to see if there is acorrelation. Ice compression has also been investigated for correlation towards windforce to see if stronger wind generates stronger compression.Using the velocity in different ice compressions an estimate of ice resistance that stemfrom ice compression has been extracted by means of Lindqvist’s formula.

Scientists often need to know how data was derived in addition to what it is. The detailed tracking of data transformation or provenance allows result reproducibility, knowledge reuse and data analysis. Scientific workflows are increasingly being used to represent provenance as they are capable of recording complicated processes at various levels of detail. In context of knowledge reuse and sharing; search technology is of paramount importance specially considering the huge and ever increasing amount of scientific data. It is computationally hard to produce a single exact answer to the user's query due to sheer volume and complicated structure of provenance.  One solution to this difficult problem is to produce a list of candidate matches and let user select the most relevant result. Here search result presentation becomes very important as the user is required to make the final decision by looking at the workflows in the result list. Presentation of these candidate matches needs to be brief, precise, clear and revealing. This is a challenging task in case of workflows as they contain textual content as well as graphical structure. Current workflow search engines such as Yahoo Pipes! or myExperiment ignore the actual workflow specification and use metadata to create summaries. Workflows which lack metadata do not make good summaries even if they are useful and relevant as search criteria. This work investigates the possibility of creating meaningful and usable summaries or snippets based on structure and specification of workflows. We shall  (1) present relevant published work done regarding snippet building techniques (2) explain how we mapped current techniques to our work (3) describe how we identified techniques from interface design theory in order to make usable graphical interface (4) present implementation of two new algorithms for workflow graph compression and their complexity analysis (5) identify future work in our implementation and outline open research problems in snippet building field.

The ability to produce potable water economically is the primary purpose of seawater desalination research. Reverse osmosis (RO) and multi-stage flash (MSF) cost more than potable water produced from fresh water resources. As an alternative to RO and MSF, this research investigates a high-efficiency mechanical vapor-compression distillation system that employs an improved water flow arrangement. The incoming salt concentration was 0.15% salt for brackish water and 3.5% salt for seawater, whereas the outgoing salt concentration was 1.5% and 7%, respectively. Distillation was performed at 439 K (331oF) and 722 kPa (105 psia) for both brackish water feed and seawater feed. Water costs of the various conditions were calculated for brackish water and seawater feeds using optimum conditions considered as 25 and 20 stages, respectively. For brackish water at a temperature difference of 0.96 K (1.73oF), the energy requirement is 2.0 kWh/m3 (7.53 kWh/kgal). At this condition, the estimated water cost is $0.39/m3 ($1.48/kgal) achieved with 10,000,000 gal/day distillate, 30-year bond, 5% interest rate, and $0.05/kWh electricity. For seawater at a temperature difference of 0.44 K (0.80oF), the energy requirement is 3.97 kWh/m3 (15.0 kWh/kgal) and the estimated water cost is $0.61/m3 ($2.31/kgal). Greater efficiency of the vapor compression system is achieved by connecting multiple evaporators in series, rather than the traditional parallel arrangement. The efficiency results from the gradual increase of salinity in each stage of the series arrangement in comparison to parallel. Calculations using various temperature differences between boiling brine and condensing steam show the series arrangement has the greatest improvement at lower temperature differences. The following table shows the improvement of a series flow arrangement compared to parallel: ?T (K) Improvement (%)*1.111 2.222 3.333 15.21 10.80 8.37 * Incoming salt concentration: 3.5% Outgoing salt concentration: 7% Temperature: 450 K (350oF) Pressure: 928 kPa (120 psig) Stages: 4