Les ombres sont un élément fondamental pour garantir le réalisme des images de synthèses. Elles donnent des informations nécessaires pour discerner correctement le positionnement relatif des objets. Le calcul d'ombre dures exactes en temps-réel reste toutefois un problème difficile en Informatique Graphique. Les Shadow Maps sont privilégiées pour leurs excellentes performances, mais elles ne peuvent garantir un résultat exact par pixel. Les volumes d'ombre sont plus lents que les Shadow Maps mais produisent des ombres exactes en toute circonstances. Bien qu'ils aient récemment atteint un haut niveau de performances, ils restent sensibles à la complexité géométrique et sont limités à des scènes de taille modeste. Dans cette thèse, nous proposons un nouvel algorithme de rendu d'ombres dures en temps réel basé sur les volumes d'ombre qui reste rapide sur des scènes de très grande tailles. De plus, nous introduisons une nouvelle stratégie de partitionnement jusque-là inexplorée en Informatique Graphique : les arbres métriques. / Shadows are a fundamental visual effect for computer generated images. They provide essential spatial hints allowing us to correctly perceive objects positions in the scene. Despite its importance, computing pixel accurate shadows in real-time is a challenging problem in Computer Graphics. Shadow Maps are still the favored solution due to their high efficiency, but they can't always provide pixel accurate results. Shadow Volumes are slower than Shadow Maps but always garantee pixel accurate shadows. While recent works significantly improved their performance, they remain sensitive to scalability, which limits their application scope. In this thesis, we propose a new algorithm based on Shadow Volumes that remains fast even on very large scenes. Furthermore, we introduce a new partitioning strategy that is yet unexplored in Computer Graphics : the metric tree.

Die Modellierung und Simulation dynamischer Vorgänge ist ein etabliertes Werkzeug bei der Entwicklung und Untersuchung technisch-physikalischer Systeme. Gleichungsbasierte und objektorientierte Modellbeschreibungssprachen wie Modelica ermöglichen eine effiziente Modellierung bei gleichzeitig guter Wiederverwendbarkeit und Übersichtlichkeit. Der Modelica-Compiler übersetzt das Modell in eine mathematisch lösbare Form und erzeugt ein ausführbares Simulationsprogramm. Mithilfe symbolischer Verfahren kann der Berechnungsaufwand und die Zeit zur Auswertung der Modellgleichungen positiv beeinflusst werden. Die vorliegende Arbeit identifiziert Potenziale zur automatisierten Laufzeitverbesserung von Modelica-Modellen und untersucht geeignete Verfahren, welche in den open source Compiler OpenModelica implementiert werden. Die vorgeschlagenen Methoden basieren auf einer graphentheoretischen Repräsentation des Gleichungssystems. Graphenalgorithmen werden zur symbolischen Manipulation des Gleichungssystem benutzt, um die Ausführungszeit zu reduzieren und ein nicht-deterministisches Laufzeitverhalten zu vermeiden. Es werden Methoden aufgezeigt, welche automatisiert und generalisiert ohne manuelles Eingreifen in den Übersetzungsprozess zu einer beschleunigten Modellauswertung führen. Die vorliegende Arbeit leistet somit einen Beitrag für die Echtzeitfähigkeit von Modellen mit steigender Komplexität, effizienter Modellerstellung und leistungsfähigen, automatisierten Simulationsläufen, wie bspw. Parameterstudien oder virtuellen Testläufen. / Modelling and simulation of dynamic processes is a well-established tool for the development of technical and physical systems. Equation-based and object-oriented modelling languages such as Modelica enable an efficient way of modelling as well as good reusability and structuring of the model. The Modelica-Compiler translates the model to a solvable representation and compiles an executable simulation program. With the help of symbolic compiler methods, the calculation costs and executation time can be reduced. The present paper identifies potentials for the automated improvement of execution times of Modelica-based simulation models and investigates suitable methods which have been implemented in the OpenModelica Compiler. The proposed compiler methods are based on a graph representation of the underlying equation system. Graph algorithms are used for the symbolic manipulation of the system of equations to reduce execuation time and non-deterministic runtime behaviour. All presented methods can be applied automatically without addtional manual modifications during the compilation process. Hence, the present paper is a contribution to achieve faster simulation times of models with increasing complexity. This is an eminent requirement for a productive and highly automated virtual development process.

The topic of the master's thesis is a concept and implementation of the prototype application TOPZ demostrating the data warehouse. All theoretical facts about the data storage space are discussed at first. It is also specified the areas of possible improvements of the data warehouse. The specification of requirements and concept of the demonstration application are described in the following part. Testing the performace of the data warehouse is discussed in last chapter.

Using transparency is an important technique for modelling partial coverage in computer graphics and is used for a variety of effects such as hair, fur, glass, smoke and particle effects. Replacing the per object sorting as the dominant algorithm used in real time 3D engines with anew algorithm with less rendering artifacts could let artists use more transparency and effects in their scenes. This thesis evaluates two algorithms as a possible replacement to the per object sorting algorithm, the per pixel linked list and the weighted blended order-independent transparency (WBOIT) in a visual comparison and a performance evaluation. The results shows that neither of the algorithms are suitable as replacement as a unified and general transparency algorithm but there are cases where the algorithms can be used as a replacement. The per pixel linked list has good visual quality but lacks in performance and can be used in limited scope such as hair renderings. The WBOIT algorithm has good performance but the approximation used can give noticeable errors in color when there is a large difference in depth between transparent objects. / I datorgrafik är transparens en viktig teknik för att efterlikna partiell blockering och används i ett flertal områden, så som rendering av hår, päls, glas, rök och partikeleffekter. Om det går att ersätta per objekt-sortering som den mest använda algoritmen i realtidsmotorer för 3D med en ny algoritm med färre renderingsfel blir det möjligt för 3D-designersatt använda mer genomskinlighet och effekter i sina scener. I den här rapporten utvärderas och jämförs två algoritmer, per pixel linked lists och weighted blended order-independent transparency mot per objektsortering i en visuell- och prestandautvärdering. Implementeringen är gjort i spelmotorn Stingray. Resultaten visar att ingen av algoritmerna är en värdig kandidat föratt användas som en universell och generell transparensalgoritm men det finns tillfällen där deras styrkor kan användas. Per pixel linked lists håller en hög visuell kvalité men har inte tillräckligt bra prestanda och kan därför endast användas för mindre områden, så som hårrendering. Weighted blended order-independent transparency har istället bra prestanda men kan skapa påtagliga visuella fel när det är stora skillnader i djup mellan objekt.

Culling out non-visible portions of 3D scenes is important for rendering large complex worlds at an interactive frame rate. Past 3D engines used static prebaked visibility data which was generated using complex algorithms. Hardware Occlusion Queries are a modern feature that allows engines to determine if objects are invisible on the fly. This allows for fully dynamic destructible and editable environments as opposed to static prebaked environments of the past. This paper presents an algorithm that uses Hardware Occlusion Queries to cull fully dynamic scenes in real-time. This algorithm is relatively simple in comparison to other real-time occlusion culling techniques, making it possible for the average developer to render large detailed scenes. It also requires very little work from the artists who design the scenes since no portals, occluders, or other special objects need to be used.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第20071号 / 医博第4164号 / 新制||医||1018(附属図書館) / 33187 / 京都大学大学院医学研究科医学専攻 / (主査)教授 黒田 知宏, 教授 松田 文彦, 教授 楠見 明弘 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

The need for accurate 3D reconstructions of complex and large environments or structures has risen dramatically in recent years. In this context, devices known as portable mobile mapping systems have lately emerged as fast and accurate reconstruction solutions. While most of the research and commercial works have relied so far on laser scanners, solutions solely based on cameras and photogrammetry are attracting an increasing interest for the minor costs, size and power consumption of cameras. This thesis presents a novel handheld mobile mapping system based on stereo vision and image-based 3D reconstruction techniques. The main novelty of the system is that it leverages Visual Simultaneous Localization And Mapping (V-SLAM) technology to support and control the acquisition of the images. The real-time estimates of the system trajectory and 3D structure of the scene are used not only to enable a live feedback of the mapped area, but also to optimize the saving of the images, provide geometric and radiometric quality measures of the imagery, and robustly control the acquisition parameters of the cameras. To the best of authors’ knowledge, the proposed system is the first handheld mobile mapping system to offer these features during the acquisition of the images, and the results support its advantages in enabling accurate and controlled visual mapping experiences even in complex and challenging scenarios.

No description available.

Transfigurations is a symphonic work in one movement for orchestra and live computer processing utilizing the graphical audio programming language Pure Data. The score and patch for this piece are accompanied by an essay describing the audio processing techniques and the compositional processes employed in this work. Programming methods discussed include strategies for data capture, patch structure, user interface, and processor management. All audio processing in the work is realized in realtime. These sounds are derived directly from the orchestra in performance, except for the last. The processes involved in Transfigurations include pitch and amplitude tracking, pitch-shifting, filtering, frequency and amplitude modulation, granular synthesis, delay, and convolution. The final sounds from the computer employ stochastic processes for synthesis which are derived from the germinal materials of the piece. The essay also discusses the aesthetic philosophy and formal structure of the work, principle themes and motives, and formative pitch materials, as well as the compositional processes in each section. The final discourse of the essay considers microphone and loudspeaker setups, patch preparation and leveling, and strategies for rehearsal and performance. / Music Composition / Accompanied by one .pdf file, Transfigurations for orchestra and live computer processing: Full score, and one .pd file.

This report explores the integration of 5G cellular networks with Time-Sensitive Networks (TSN), focusing on network latencies. While mobile 5G networks offer high bandwidth and low latency, TSN is an IEEE standard-based wired solution used in real-time applications for deterministic data transmission with low latency and high reliability. Integrating TSN with 5G could significantly enhance scalability, particularly in industrial automation, by providing flexibility, efficiency, and responsiveness. When combining 5G capabilities with TSN, seamless communication across wired and wireless domains becomes achievable. 5G supports several Quality of Service (QoS) flows with different priorities that must map to TSN QoS to ensure smooth integration. To meet TSN's requirements for latency and jitter, effective traffic translation and forwarding between the networks are crucial. This thesis aims to address key questions regarding traffic translation, QoS implementation, and latency in both TSN and 5G networks. Through experiments and evaluations, we assess latency and network capabilities. Understanding these metrics is essential for devising effective integration strategies. Our findings indicate that integrating 5G with TSN is feasible for achieving low latencies for packet sizes below 128-bytes. However, as packet sizes increase, latencies and jitter rise significantly. This result indicates that Firecell 5G technology may have difficulties efficiently handling larger packet sizes without latency degradation. Additionally, leveraging QoS functionalities in the current version of OpenAirInterface (OAI), which forms the foundation of the utilized Firecell 5G, is currently unattainable. This thesis highlights the importance of implementing QoS functionalities to maintain low latencies and jitter in targeted network applications. / Denna rapport undersöker den potentiella integrationen av 5G-mobilnät med tidskänsliga nätverk (TSN), med fokus på nätverksfördröjningar. Medan 5G-nät erbjuder hög bandbredd och låg fördröjning, är TSN en IEEE-standardiserad trådbunden teknik som används i realtidsapplikationer för deterministisk dataöverföring med låg fördröjning och hög tillförlitlighet. Att integrera TSN med 5G kan avsevärt förbättra skalbarheten, särskilt inom industriell automation, genom att erbjuda flexibilitet, effektivitet och låg fördröjning. Genom att kombinera 5G med TSN blir tidskänslig kommunikation över trådbundna och trådlösa nätverk möjligt. 5G stödjer flera kvalitetsnivåer för tjänster (QoS) med olika prioriteringar som måste anpassas till TSN:s QoS för att säkerställa en smidig integration. För att möta TSN:s krav på fördröjning och jitter är effektiv datahantering mellan nätverken avgörande. Denna avhandling behandlar centrala frågor kring datahantering, QoS-implementering och fördröjning i både TSN och 5G-nätverk. Genom experiment och utvärderingar får vi fram information om fördröjning och nätverkskapacitet i TSN-5G, vilket är avgörande för att utveckla effektiva integrationsstrategier. Våra resultat visar att integrationen av TSN-5G uppnår låg fördröjning för paketstorlekar under 128 byte. När paketstorleken ökar, ökar också fördröjning och jitter avsevärt. Resultaten visar att Firecell 5G-teknologin har svårigheter att effektivt hantera större paketstorlekar utan ökad fördröjning. Utöver detta är det för närvarande inte möjligt att använda QoS-funktioner i den nuvarande versionen av OpenAirInterface (OAI), vilket är mjukvaran Firecell 5G bygger på. Denna avhandling understryker vikten av att kunna implementera QoS-funktioner för att upprätthålla låg fördröjning och jitter i specifika nätverksapplikationer.