004 Datenverarbeitung; Informatik
Refine
Year of publication
Document Type
- Part of Periodical (73)
- Doctoral Thesis (33)
- Master's Thesis (29)
- Diploma Thesis (27)
- Bachelor Thesis (25)
- Study Thesis (11)
- Conference Proceedings (4)
- Habilitation (1)
Language
- English (203) (remove)
Keywords
- Software Engineering (6)
- Bluetooth (4)
- ontology (4)
- API (3)
- E-KRHyper (3)
- Enterprise 2.0 (3)
- Knowledge Compilation (3)
- OWL (3)
- OWL <Informatik> (3)
- Ontologie <Wissensverarbeitung> (3)
This paper introduces Vocville, a causal online game for learning vocabularies. I am creating this application for my master thesis of my career as a "Computervisualist" (computer visions) for the University of Koblenz - Landau. The application is an online browser game based on the idea of the really successful Facebook game FarmVille. The application is seperated in two parts; a Grails application manages a database which holds the game objects like vocabulary, a Flex/Flash application generates the actual game by using these data. The user can create his own home with everything in it. For creating things, the user has to give the correct translation of the object he wants to create several times. After every query he has to wait a certain amount of time to be queried again. When the correct answer is given sufficient times, the object is builded. After building one object the user is allowed to build others. After building enough objects in one area (i.e. a room, a street etc.) the user can activate other areas by translating all the vocabularies of the previous area. Users can also interact with other users by adding them as neighbors and then visiting their homes or sending them gifts, for which they have to fill in the correct word in a given sentence.
Only little information is available about the diffusion of cloud computing in German higher educational institutions. A better understanding of the state of the art in this field would support the modernization of the higher educational institutions in Germany and allow the development of more adequate cloud products and more appropriate business models for this niche. For this purpose, a literature research on Cloud Computing and IT-diffusion will be run and an empirical investigation with an online questionnaire addressed to higher educational institutions in Germany will be performed to illustrate the state of the art of Cloud Computing in German higher educational institutions as well as the threats and opportunities perceived by employees of higher educational institutions data centers connected to the usage of the cloud.
In addition to that, different experts from universities and businesses will be interviewed to complete the knowledge and information collected through the online questionnaire and during the research phase. The expected results will serve to create a recommendation for higher educational institutions in Germany about either they should migration to the cloud or not and introduce a list of guiding questions of critical issues to consider before using cloud-computing technologies.
Computed tomography (CT) and magnetic resonance imaging (MRI) in the medical area deliver huge amounts of data, which doctors have to handle in a short time. These data can be visualised efficiently with direct volume rendering. Consequently most direct volume rendering applications on the market are specialised on medical tasks or integrated in medical visualisa- tion environments. Highly evolved applications for tasks like diagnosis or surgery simulation are available in this area. In the last years, however, another area is making increasing use of com- puted tomography. Companies like phoenix |x-ray, founded in 1999 pro- duce CT-scanners especially dedicated to industrial applications like non destructive material testing (NDT). Of course an application like NDT has different demands on the visualisation than a typical medical application. For example a typical task for non destructive testing would be to high- light air inclusions (pores) in a casting. These inclusions usually cover a very small area and are very hard to classify only based on their density value as this would also highlight the air around the casting. This thesis presents multiple approaches to improve the rendering of in- dustrial CT data, most of them based on higher dimensional transfer func- tions. Therefore the existing volume renderer application of VRVis was extended with a user interface to create such transfer functions and exist- ing render modes were adapted to profit from the new transfer functions. These approaches are especially suited to improve the visualisation of sur- faces and material boundaries as well as pores. The resulting renderings make it very easy to identify these features while preserving interactive framerates.
This thesis addresses the problem of terrain classification in unstructured outdoor environments. Terrain classification includes the detection of obstacles and passable areas as well as the analysis of ground surfaces. A 3D laser range finder is used as primary sensor for perceiving the surroundings of the robot. First of all, a grid structure is introduced for data reduction. The chosen data representation allows for multi-sensor integration, e.g., cameras for color and texture information or further laser range finders for improved data density. Subsequently, features are computed for each terrain cell within the grid. Classification is performedrnwith a Markov random field for context-sensitivity and to compensate for sensor noise and varying data density within the grid. A Gibbs sampler is used for optimization and is parallelized on the CPU and GPU in order to achieve real-time performance. Dynamic obstacles are detected and tracked using different state-of-the-art approaches. The resulting information - where other traffic participants move and are going to move to - is used to perform inference in regions where the terrain surface is partially or completely invisible for the sensors. Algorithms are tested and validated on different autonomous robot platforms and the evaluation is carried out with human-annotated ground truth maps of millions of measurements. The terrain classification approach of this thesis proved reliable in all real-time scenarios and domains and yielded new insights. Furthermore, if combined with a path planning algorithm, it enables full autonomy for all kinds of wheeled outdoor robots in natural outdoor environments.
The following thesis analyses the functionality and programming capabilitiesrnof compute shaders. For this purpose, chapter 2 gives an introductionrnto compute shaders by showing how they work and how they can be programmed. In addition, the interaction of compute shaders and OpenGL 4.3 is shown through two introductory examples. Chapter 3 describes an NBodyrnsimulation that has been implemented in order to show the computational power of compute shaders and the use of shared memory. Then it is shown in chapter 4 how compute shaders can be used for physical simulationsrnand where problems may arise. In chapter 5 a specially conceived and implemented algorithm for detecting lines in images is described and then compared with the Hough transform. Lastly, a final conclusion is drawn in chapter 6.
Interactive video retrieval
(2006)
The goal of this thesis is to develop a video retrieval system that supports relevance feedback. One research approach of the thesis is to find out if a combination of implicit and explicit relevance feedback returns better retrieval results than a system using explicit feedback only. Another approach is to identify a model to weight existing feature categories. For this purpose, a state-of-the-art analysis is presented and two systems implemented, which run under the conditions of the international TRECVID workshop. It will be a basis system for further research approaches in the field of interactive video retrieval. Amongst others, it shall participate in the 2006 search task of the mentioned workshop.
In this thesis the feasibility of a GPGPU (general-purpose computing on graphics processing units) approach to natural feature description on mobile phone GPUs is assessed. To this end, the SURF descriptor [4] has been implemented with OpenGL ES 2.0/GLSL ES 1.0 and evaluated across different mobile devices. The implementation is multiple times faster than a comparable CPU variant on the same device. The results proof the feasibility of modern mobile graphics accelerators for GPGPU tasks especially for the detection phase in natural feature tracking used in augmented reality applications. Extensive analysis and benchmarking of this approach in comparison to state of the art methods have been undertaken. Insights into the modifications necessary to adapt and modify the SURF algorithm to the limitations of a mobile GPU are presented. Further, an outlook for a GPGPU-based tracking pipeline on a mobile device is provided.
The goal of this Bachelor thesis is to implement and evaluate the "Simulating of Collective Misbelief"-model into the NetLogo programming language. Therefore, the model requirements have to be specified and implemented into the NetLogo environment. Further tool-related re-quirements have to be specified to enable the model to work in NetLogo. After implementation several simulations will be conducted to answer the research question stated above.
This paper documents the development of an abstract physics layer (APL) for Simspark. After short introductions to physics engines and Simspark, reasons why an APL was developed are explained. The biggest part of this paper describes the new design and why certain design choices were made based on requirements that arose during developement. It concludes by explaining how the new design was eventually implemented and what future possibilities the new design holds.
Augmented Reality bedeutet eine reale Umgebung mit, meistens grafischen, virtuellen Inhalten zu erweitern. Oft sind dabei die virtuellen Inhalte der Szene jedoch nur ein Overlay und interagieren nicht mit den realen Bestandteilen der Szene. Daraus ergibt sich ein Authentizitätsproblem für Augmented Reatliy Anwendungen. Diese Arbeit betrachtet Augmented Reality in einer speziellen Umgebung, mit deren Hilfe eine authentischere Darstellung möglich ist. Ziel dieserArbeitwar die Erstellung eines Systems, das Zeichnungen durch Techniken der Augmented Reality mit virtuellen Inhalten erweitert. Durch das Anlegen einer Repräsentation soll es der Anwendung dabei möglich sein die virtuellen Szeneelementemit der Zeichnung interagieren zu lassen. Dazu wurden verschiedene Methoden aus den Bereichen des Pose Tracking und der Sketch Recognition disktutiert und für die Implementierung in einem prototypischen System ausgewählt. Als Zielhardware fungiert ein Android Smartphone. Kontext der Zeichnungen ist eine Dungeon Karte, wie sie in Rollenspielen vorkommt. Die virtuellen Inhalte nehmen dabei die Form von Bewohnern des Dungeons an, welche von einer Agentensimulation verwaltet werden. Die Agentensimulation ist Gegenstand einer eigenen Diplomarbeit [18]. Für das Pose Tracking wurde ARToolkitPlus eingesetzt, ein optisches Tracking System, das auf Basis von Markern arbeitet. Die Sketch Recognition ist dafür zuständig die Inhalte der Zeichnung zu erkennen und zu interpretieren. Dafür wurde ein eigener Ansatz implementiert der Techniken aus verschiedenen Sketch Recognition Systemen kombiniert. Die Evaluation konzentriert sich auf die technischen Aspekte des Systems, die für eine authentische Erweiterung der Zeichnung mit virtuellen Inhalten wichtig sind.
For software engineers, conceptually understanding the tools they are using in the context of their projects is a daily challenge and a prerequisite for complex tasks. Textual explanations and code examples serve as knowledge resources for understanding software languages and software technologies. This thesis describes research on integrating and interconnecting
existing knowledge resources, which can then be used to assist with understanding and comparing software languages and software technologies on a conceptual level. We consider the following broad research questions that we later refine: What knowledge resources can be systematically reused for recovering structured knowledge and how? What vocabulary already exists in literature that is used to express conceptual knowledge? How can we reuse the
online encyclopedia Wikipedia? How can we detect and report on instances of technology usage? How can we assure reproducibility as the central quality factor of any construction process for knowledge artifacts? As qualitative research, we describe methodologies to recover knowledge resources by i.) systematically studying literature, ii.) mining Wikipedia, iii.) mining available textual explanations and code examples of technology usage. The theoretical findings are backed by case studies. As research contributions, we have recovered i.) a reference semantics of vocabulary for describing software technology usage with an emphasis on software languages, ii.) an annotated corpus of Wikipedia articles on software languages, iii.) insights into technology usage on GitHub with regard to a catalog of pattern and iv.) megamodels of technology usage that are interconnected with existing textual explanations and code examples.
The aim of this paper is to identify and understand the risks and issues companies are experiencing from the business use of social media and to develop a framework for describing and categorising those social media risks. The goal is to contribute to the evolving theorisation of social media risk and to provide a foundation for the further development of social media risk management strategies and processes. The study findings identify thirty risk types organised into five categories (technical, human, content, compliance and reputational). A risk-chain is used to illustrate the complex interrelated, multi-stakeholder nature of these risks and directions for future work are identified.
The World Wide Web (WWW) has become a very important communication channel. Its usage has steadily grown within the past. Interest by website owners in identifying user behaviour has been around since Tim Berners-Lee developed the first web browser in 1990. But as the influence of the online channel today eclipses all other media the interest in monitoring website usage and user activities has intensified as well. Gathering and analysing data about the usage of websites can help to understand customer behaviour, improve services and potentially increase profit.
It is further essential for ensuring effective website design and management, efficient mass customization and effective marketing. Web Analytics (WA) is the area addressing these considerations. However, changing technologies and evolving Web Analytic methods and processes present a challenge to organisations starting with Web Analytic programmes. Because of lacking resources in different areas and other types of websites especially small and medium-sized enterprises (SME) as well as non-profit organisations struggle to operate WA in an effective manner.
This research project aims to identify the existing gap between theory, tool possibilities and business needs for undertaking Web Analytic programmes. Therefore the topic was looked at from three different ways: the academic literature, Web Analytic tools and an interpretative case study. The researcher utilized an action research approach to investigate Web Analytics presenting an holistic overview and to identify the gaps that exists. The outcome of this research project is an overall framework, which provides guidance for SMEs who operate information websites on how to proceed in a Web Analytic programme.
101worker is the modular knowledge engineering component of the 101companies project. It has developed maintainability and performance problems due to growing organically, rather than following best software design practices. This thesis lays out these problems, drafts a set of requirements for refactoring the system and then describes and analyzes the resulting implementation. The solution involves collation of scattered and redundant information, setup of unit and functional test suites and incrementalization of the bus architecture of 101worker.
The publication of freely available and machine-readable information has increased significantly in the last years. Especially the Linked Data initiative has been receiving a lot of attention. Linked Data is based on the Resource Description Framework (RDF) and anybody can simply publish their data in RDF and link it to other datasets. The structure is similar to the World Wide Web where individual HTML documents are connected with links. Linked Data entities are identified by URIs which are dereferenceable to retrieve information describing the entity. Additionally, so called SPARQL endpoints can be used to access the data with an algebraic query language (SPARQL) similar to SQL. By integrating multiple SPARQL endpoints it is possible to create a federation of distributed RDF data sources which acts like one big data store.
In contrast to the federation of classical relational database systems there are some differences for federated RDF data. RDF stores are accessed either via SPARQL endpoints or by resolving URIs. There is no coordination between RDF data sources and machine-readable meta data about a source- data is commonly limited or not available at all. Moreover, there is no common directory which can be used to discover RDF data sources or ask for sources which offer specific data. The federation of distributed and linked RDF data sources has to deal with various challenges. In order to distribute queries automatically, suitable data sources have to be selected based on query details and information that is available about the data sources. Furthermore, the minimization of query execution time requires optimization techniques that take into account the execution cost for query operators and the network communication overhead for contacting individual data sources. In this thesis, solutions for these problems are discussed. Moreover, SPLENDID is presented, a new federation infrastructure for distributed RDF data sources which uses optimization techniques based on statistical information.
Ontologies play an important role in knowledge representation for sharing information and collaboratively developing knowledge bases. They are changed, adapted and reused in different applications and domains resulting in multiple versions of an ontology. The comparison of different versions and the analysis of changes at a higher level of abstraction may be insightful to understand the changes that were applied to an ontology. While there is existing work on detecting (syntactical) differences and changes in ontologies, there is still a need in analyzing ontology changes at a higher level of abstraction like ontology evolution or refactoring pattern. In our approach we start from a classification of model refactoring patterns found in software engineering for identifying such refactoring patterns in OWL ontologies using DL reasoning to recognize these patterns.
SOA-Security
(2007)
This paper is a part of the ASG project (Adaptive Services Grid) and addresses some IT security issues of service oriented architectures. It defines a service-oriented security concept, it explores the SOA security challenge, it describes the existing WS-Security standard, and it undertakes a first step into a survey on best practice examples. In particular, the ASG middleware platform technology (JBossWS) is analyzed with respect to its ability to handle security functions.
This paper describes the development of security requirements for non-political Internet voting. The practical background is our experience with the Internet voting within the Gesellschaft für Informatik (GI - Informatics Society) 2004 and 2005. The theoretical background is the international state-of-the-art of requirements about electronic voting, especially in the US and in Europe. A focus of this paper is on the user community driven standardization of security requirements by means of a Protection Profile of the international Common Criteria standard.
Virtual Goods + ODRL 2012
(2012)
This is the 10th international workshop for technical, economic, and legal aspects of business models for virtual goods incorporating the 8th ODRL community group meeting. This year we did not call for completed research results, but we invited PhD students to present and discuss their ongoing research work. In the traditional international group of virtual goods and ODRL researchers we discussed PhD research from Belgium, Brazil, and Germany. The topics focused on research questions about rights management in the Internet and e-business stimulation. In the center of rights management stands the conception of a formal policy expression that can be used for human readable policy transparency, as well as for machine readable support of policy conformant systems behavior up to automatic policy enforcement. ODRL has proven to be an ideal basis for policy expressions, not only for digital copy rights, but also for the more general "Policy Awareness in the World of Virtual Goods". In this sense, policies support the communication of virtual goods, and they are a virtualization of rules-governed behavior themselves.
Various best practices and principles guide an ontology engineer when modeling Linked Data. The choice of appropriate vocabularies is one essential aspect in the guidelines, as it leads to better interpretation, querying, and consumption of the data by Linked Data applications and users.
In this paper, we present the various types of support features for an ontology engineer to model a Linked Data dataset, discuss existing tools and services with respect to these support features, and propose LOVER: a novel approach to support the ontology engineer in modeling a Linked Data dataset. We demonstrate that none of the existing tools and services incorporate all types of supporting features and illustrate the concept of LOVER, which supports the engineer by recommending appropriate classes and properties from existing and actively used vocabularies. Hereby, the recommendations are made on the basis of an iterative multimodal search. LOVER uses different, orthogonal information sources for finding terms, e.g. based on a best string match or schema information on other datasets published in the Linked Open Data cloud. We describe LOVER's recommendation mechanism in general and illustrate it alongrna real-life example from the social sciences domain.
This bachelor thesis deals with the comparison related to the similarity of recorded WiFi patterns during the tracing of a path through the streets of a large city. Both MAC address only comparison has been investigated as well as the incorporation of RSSI values, whereby the localization accuracy has been evaluated. Methods for the detection of different types and combinations of loops in the path are demonstrated likewise the attempt to estimate the degree of urban development in the environment of the user by assessing the received signal strength and signal-to-noise ratio of GPS satellites and GSM cell towers.
In order to observe a user- proximity to a certain spot on a large public square the absorption of WiFi signals by the human body has been taken into account. Finally, the results of a comparison of the computing performance of a modern smartphone versus the alternative of remote calculation on a server including data transmission via cellular data network are presented.
In this thesis the possibilities for real-time visualization of OpenVDB
files are investigated. The basics of OpenVDB, its possibilities, as well
as NanoVDB and its GPU port, were studied. A system was developed
using PNanoVDB, the graphics API port of OpenVDB. Techniques were
explored to improve and accelerate a single ray approach of ray tracing.
To prove real-time capability, two single scattering approaches were
also implemented. One of these was selected, further investigated and
optimized to achieve interactive real-time rendering.
It is important to give artists immediate feedback on their adjustments, as
well as the possibility to change all parameters to ensure a user friendly
creation process.
In addition to the optical rendering, corresponding benchmarks were
collected to compare different improvement approaches and to prove
their relevance. Attention was paid to the rendering times and memory
consumption on the GPU to ensure optimal use. A special focus, when
rendering OpenVDB files, was put on the integrability and extensibility of
the program to allow easy integration into an existing real-time renderer
like U-Render.
This paper presents a method for the evolution of SHI ABoxes which is based on a compilation technique of the knowledge base. For this the ABox is regarded as an interpretation of the TBox which is close to a model. It is shown, that the ABox can be used for a semantically guided transformation resulting in an equisatisfiable knowledge base. We use the result of this transformation to effciently delete assertions from the ABox. Furthermore, insertion of assertions as well as repair of inconsistent ABoxes is addressed. For the computation of the necessary actions for deletion, insertion and repair, the E-KRHyper theorem prover is used.
Knowledge compilation is a common technique for propositional logic knowledge bases. The idea is to transform a given knowledge base into a special normal form ([MR03],[DH05]), for which queries can be answered efficiently. This precompilation step is very expensive but it only has to be performed once. We propose to apply this technique to knowledge bases defined in Description Logics. For this, we introduce a normal form, called linkless concept descriptions, for ALC concepts. Further we present an algorithm, based on path dissolution, which can be used to transform a given concept description into an equivalent linkless concept description. Finally we discuss a linear satisfiability test as well as a subsumption test for linkless concept descriptions.
This paper offers an informal overview and discussion on first order predicate logic reasoning systems together with a description of applications which are carried out in the Artificial Intelligence Research Group of the University in Koblenz. Furthermore the technique of knowledge compilation is shortly introduced.
This paper shows how multiagent systems can be modeled by a combination of UML statecharts and hybrid automata. This allows formal system specification on different levels of abstraction on the one hand, and expressing real-time system behavior with continuous variables on the other hand. It is not only shown how multi-robot systems can be modeled by a combination of hybrid automata and hierarchical state machines, but also how model checking techniques for hybrid automata can be applied. An enhanced synchronization concept is introduced that allows synchronization taking time and avoids state explosion to a certain extent.
In this paper we describe a series of projects on location based and personalised information systems. We start wit a basic research project and we show how we came with the help of two other more application oriented project to a product. This is developed by a consortium of enterprises and it already is in use in the city of Koblenz.
In this paper we describe a network for distributing personalized information within a pervasive university. We discuss the system architecture of our Bluetooth-based CampusNews-system, both, from the administrator and the user viewpoint. We furthermore present first statistical data about the usage of the partial installation at the Koblenz campus together with an outlook to future work.
In this paper we describe a network for distributing personalized Information in a metropolitan area. We discuss the system architecture of our Bluetooth-based information system as well as the reasoning process that fits users" needs with potential messages. We furthermore present our findings on parallelizing Bluetooth connection setup and performance.
Identifying reusable legacy code able to implement SOA services is still an open research issue. This master thesis presents an approach to identify legacy code for service implementation based on dynamic analysis and the application of data mining techniques. rnrnAs part of the SOAMIG project, code execution traces were mapped to business processes. Due to the high amount of traces generated by dynamic analyses, the traces must be post-processed in order to provide useful information. rnrnFor this master thesis, two data mining techniques - cluster analysis and link analysis - were applied to the traces. First tests on a Java/Swing legacy system provided good results, compared to an expert- allocation of legacy code.
With the ongoing process of building business networks in today- economy, business to-business integration (B2B Integration) has become a strategic tool for utilizing and optimizing information exchange between business partners. Industry and academia have made remarkable progress in implementing and conceptualizing different kinds of electronic inter-company relationships in the last years. Nevertheless, academic findings generally focus exclusively on certain aspects of the research object, e.g. document standards, process integration or other descriptive criteria. Without arncommon framework these results stay unrelated and their mutual impact on each other remains largely unexplained. In this paper we explore motivational factors of B2B integration in practice. In a research project using a uniform taxonomy (eXperience methodology) we classified real-world B2B integration projects from a pool of over 400 case studies using a pre-developed framework for integration scenarios. The result of our partly exploratory research shows the influence of the role of a company in the supply chain and its motive to invest in a B2B solution.
Semantic desktop environments aim at improving the effectiveness and efficiency of users carrying out daily tasks within their personal information management infrastructure (PIM). They support the user by transferring and exploiting the explicit semantics of data items across different PIM applications. Whether such an approach does indeed reach its aim of facilitating users" life and—if so—to which extent, however, remains an open question that we address in this paper with the first summative evaluation of a semantic desktop approach. We approach the research question exploiting our own semantic desktop infrastructure, X-COSIM. As data corpus, we have used over 100 emails and 50 documents extracted from the organizers of a conference-like event at our university. The evaluation has been carried out with 18 subjects. We have developed a test environment to evaluate COSIMail and COSIFile, two semantic PIM applications based on X-COSIM. As result, we have found a significant improvement for typical PIM tasks compared to a standard desktop environment.
Multi-agent systems are a mature approach to model complex software systems by means of Agent-Oriented Software Engineering (AOSE). However, their application is not widely accepted in mainstream software engineering. Parallel to this the interdisciplinary field of Agent-based Social Simulation (ABSS) finds increasing recognition beyond the purely academic realm which starts to draw attention from the mainstream of agent researchers. This work analyzes factors to improve the uptake of AOSE as well as characteristics which separate the two fields AOSE and ABSS to understand their gap. Based on the efficiency-oriented micro-agent concept of the Otago Agent Platform (OPAL) we have constructed a new modern and self-contained micro-agent platform called µ². The design takes technological trends into account and integrates representative technologies, such as the functionally-inspired JVM language Clojure (with its Transactional Memory), asynchronous message passing frameworks and the mobile application platform Android. The mobile version of the platform shows an innovative approach to allow direct interaction between Android application components and micro-agents by mapping their related internal communication mechanisms. This empowers micro-agents to exploit virtually any capability of mobile devices for intelligent agent-based applications, robotics or simply act as a distributed middleware. Additionally, relevant platform components for the support of social simulations are identified and partially implemented. To show the usability of the platform for simulation purposes an interaction-centric scenario representing group shaping processes in a multi-cultural context is provided. The scenario is based on Hofstede's concept of 'Cultural Dimensions'. It does not only confirm the applicability of the platform for simulations but also reveals interesting patterns for culturally augmented in- and out-group agents. This explorative research advocates the potential of micro-agents as a powerful general system modelling mechanism while bridging the convergence between mobile and desktop systems. The results stimulate future work on the micro-agent concept itself, the suggested platform and the deeper exploration of mechanisms for seemless interaction of micro-agents with mobile environments. Last but not least the further elaboration of the simulation model as well as its use to augment intelligent agents with cultural aspects offer promising perspectives for future research.
Computers and especially computer networks have become an important part of our everyday life. Almost every device we use is equipped with a computer or microcontroller. Recent technology has even boosted this development by miniaturization of the size of microcontrollers. These are used to either process or collect data. Miniature senors may sense and collect huge amounts of information coming from nature, either from environment or from our own bodies. To process and distribute the data of these sensors, wireless sensor networks (WSN) have been developed in the last couple of years. Several microcontrollers are connected over a wireless connection and are able to collect, transmit and process data for various applications. Today, there are several WSN applications available, such as environment monitoring, rescue operations, habitat monitoring and smart home applications. The research group of Prof. Elaine Lawrence at the University of Technology, Sydney (UTS) is focusing on mobile health care with WSN. Small sensors are used to collect vital data. This data is sent over the network to be processed at a central device such as computer, laptop or handheld device. The research group has developed several prototypes of mobile health care. This thesis will deal with enhancing and improving the latest prototype based on CodeBlue, a hardware and software framework for medical care.
This thesis introduces fnnlib, a C++ library for recurrent neural network simulations that I developed between October 2009 and March 2010 at Osaka University's Graduate School of Engineering. After covering the theory behind recurrent neural networks, backpropagation through time, recurrent neural networks with parametric bias, continuous-time recurrent neural networks, and echo state networks, the design of the library is explained. All of the classes as well as their interrelationships are presented along with reasons as to why certain design decisions were made. Towards the end of the thesis, a small practical example is shown. Also, fnnlib is compared to other neural network libraries.
The STOR project aims at the development of a scientific component system employing models and knowledge for object recognition in images. This interim report elaborates on the requirements for such a component system, structures the application area by identifying a large set of basic operations, and shows how a set of appropriate data structures and components can be derived. A small case studies exemplifies the approach.
The mitral valve is one of four human heart valves. It is located in the left heart and acts as a unidirectional passageway for blood between the left atrium and the left ventricle. A correctly functioning mitral valve prevents a backflow of blood into the pulmonary circulation (lungs) and thus constitutes a vital part of the cardiac cycle. Pathologies of the mitral valve can manifest in a variety of symptoms with severity ranging from chest pain and fatigue to pulmonary edema (fluid accumulation in the tissue and air space of lungs), which may ultimately cause respiratory failure.
Malfunctioning mitral valves can be restored through complex surgical interventions, which greatly benefit from intensive planning and pre-operative analysis. Visualization techniques provide a possibility to enhance such preparation processes and can also facilitate post-operative evaluation. The work at hand extends current research in this field, building upon patient-specific mitral valve segmentations developed at the German Cancer Research Center, which result in triangulated 3D models of the valve surface. The core of this work will be the construction of a 2D-view of these models through global parameterization, a method that can be used to establish a bijective mapping between a planar parameter domain and a surface embedded in higher dimensions.
A flat representation of the mitral valve provides physicians with a view of the whole surface at once, similar to a map. This allows assessment of the valve's area and shape without the need for different viewing angles. Parts of the valve that are occluded by geometry in 3D become visible in 2D.
An additional contribution of this work will be the exploration of different visualizations of the 3D and 2D mitral valve representations. Features of the valve can be highlighted by associating them with specified colors, which can for instance directly convey pathology indicators.
Quality and effectiveness of the proposed methods were evaluated through a survey conducted at the Heidelberg University Hospital.
Social networks are ubiquitous structures that we generate and enrich every-day while connecting with people through social media platforms, emails, and any other type of interaction. While these structures are intangible to us, they carry important information. For instance, the political leaning of our friends can be a proxy to identify our own political preferences. Similarly, the credit score of our friends can be decisive in the approval or rejection of our own loans. This explanatory power is being leveraged in public policy, business decision-making and scientific research because it helps machine learning techniques to make accurate predictions. However, these generalizations often benefit the majority of people who shape the general structure of the network, and put in disadvantage under-represented groups by limiting their resources and opportunities. Therefore it is crucial to first understand how social networks form to then verify to what extent their mechanisms of edge formation contribute to reinforce social inequalities in machine learning algorithms.
To this end, in the first part of this thesis, I propose HopRank and Janus two methods to characterize the mechanisms of edge formation in real-world undirected social networks. HopRank is a model of information foraging on networks. Its key component is a biased random walker based on transition probabilities between k-hop neighborhoods. Janus is a Bayesian framework that allows to identify and rank plausible hypotheses of edge formation in cases where nodes possess additional information. In the second part of this thesis, I investigate the implications of these mechanisms - that explain edge formation in social networks - on machine learning. Specifically, I study the influence of homophily, preferential attachment, edge density, fraction of inorities, and the directionality of links on both performance and bias of collective classification, and on the visibility of minorities in top-k ranks. My findings demonstrate a strong correlation between network structure and machine learning outcomes. This suggests that systematic discrimination against certain people can be: (i) anticipated by the type of network, and (ii) mitigated by connecting strategically in the network.
Colonoscopy is the gold standard for the detection of colorectal polyps that can progress into cancer. In such an examination, physicians search for polyps in endoscopic images. Thereby polyps can be removed. To support experts with a computer-aided diagnosis system, the University of Koblenz-Landau currently makes some efforts in research different methods for automatic detection. Comparable to traditional pattern recognition systems, features are initially extracted and a classifier is trained on such data. Afterwards, unknown endoscopic images can be classified with the previously trained classifier. This thesis concentrates on the extension of the feature extraction module in the existing system. New detection methods are compared to existing techniques. Several features are implemented, incorporating Graylevel Co-occurrence Matrices, Local Binary Patterns and Discrte Wavelet Transform. Different modifications on those features are applied and evaaluated.
Magnetic resonance (MR) tomography is an imaging method, that is used to expose the structure and function of tissues and organs in the human body for medical diagnosis. Diffusion weighted (DW) imaging is a specific MR imaging technique, which enables us to gain insight into the connectivity of white matter pathways noninvasively and in vivo. It allows for making predictions about the structure and integrity of those connections. In clinical routine this modality finds application in the planning phase of neurosurgical operations, such as in tumor resections. This is especially helpful if the lesion is deeply seated in a functionally important area, where the risk of damage is given. This work reviews the concepts of MR imaging and DW imaging. Generally, at the current resolution of diffusion weighted data, single white matter axons cannot be resolved. The captured signal rather describes whole fiber bundles. Beside this, it often appears that different complex fiber configurations occur in a single voxel, such as crossings, splittings and fannings. For this reason, the main goal is to assist tractography algorithms who are often confound in such complex regions. Tractography is a method which uses local information to reconstruct global connectivities, i.e. fiber tracts. In the course of this thesis, existing reconstruction methods such as diffusion tensor imaging (DTI) and q-ball imaging (QBI) are evaluated on synthetic generated data and real human brain data, whereas the amount of valuable information provided by the individual reconstruction mehods and their corresponding limitations are investigated. The output of QBI is the orientation distribution function (ODF), where the local maxima coincides with the underlying fiber architecture. We determine those local maxima. Furthermore, we propose a new voxel-based classification scheme conducted on diffusion tensor metrics. The main contribution of this work is the combination of voxel-based classification, local maxima from the ODF and global information from a voxel- neighborhood, which leads to the development of a global classifier. This classifier validates the detected ODF maxima and enhances them with neighborhood information. Hence, specific asymmetric fibrous architectures can be determined. The outcome of the global classifier are potential tracking directions. Subsequently, a fiber tractography algorithm is designed that integrates along the potential tracking directions and is able to reproduce splitting fiber tracts.
Expert-driven business process management is an established means for improving efficiency of organizational knowledge work. Implicit procedural knowledge in the organization is made explicit by defining processes. This approach is not applicable to individual knowledge work due to its high complexity and variability. However, without explicitly described processes there is no analysis and efficient communication of best practices of individual knowledge work within the organization. In addition, the activities of the individual knowledge work cannot be synchronized with the activities in the organizational knowledge work.rnrnSolution to this problem is the semantic integration of individual knowledgernwork and organizational knowledge work by means of the patternbased core ontology strukt. The ontology allows for defining and managing the dynamic tasks of individual knowledge work in a formal way and to synchronize them with organizational business processes. Using the strukt ontology, we have implemented a prototype application for knowledge workers and have evaluated it at the use case of an architectural fifirm conducting construction projects.
With the Multimedia Metadata Ontology (M3O), we have developed a sophisticated model for representing among others the annotation, decomposition, and provenance of multimedia metadata. The goal of the M3O is to integrate the existing metadata standards and metadata formats rather than replacing them. To this end, the M3O provides a scaffold needed to represent multimedia metadata. Being an abstract model for multimedia metadata, it is not straightforward how to use and specialize the M3O for concrete application requirements and existing metadata formats and metadata standards. In this paper, we present a step-by-step alignment method describing how to integrate and leverage existing multimedia metadata standards and metadata formats in the M3O in order to use them in a concrete application. We demonstrate our approach by integrating three existing metadata models: the Core Ontology on Multimedia (COMM), which is a formalization of the multimedia metadata standard MPEG-7, the Ontology for Media Resource of the W3C, and the widely known industry standard EXIF for image metadata
In dieser Ausarbeitung beschreibe ich die Ergebnisse meiner Untersuchungen zur Erweiterung des LogAnswer-Systemsmit nutzerspezifischen Profilinformationen. LogAnswer ist ein natürlichsprachliches open-domain Frage-Antwort-System. Das heißt: es beantwortet Fragen zu beliebigen Themen und liefert dabei konkrete (möglichst knappe und korrekte) Antworten zurück. Das System wird im Rahmen eines Gemeinschaftsprojekts der Arbeitsgruppe für künstliche Intelligenz von Professor Ulrich Furbach an der Universität Koblenz-Landau und der Arbeitsgruppe Intelligent Information and Communication Systems (IICS) von Professor Hermann Helbig an der Fernuniversität Hagen entwickelt. Die Motivation meiner Arbeit war die Idee, dass der Prozess der Antwortfindung optimiert werden kann, wenn das Themengebiet, auf das die Frage abzielt, im Vorhinein bestimmt werden kann. Dazu versuchte ich im Rahmen meiner Arbeit die Interessensgebiete von Nutzern basierend auf Profilinformationen zu bestimmen. Das Semantic Desktop System NEPOMUK wurde verwendet um diese Profilinformationen zu erhalten. NEPOMUK wird verwendet um alle Daten, Dokumente und Informationen, die ein Nutzer auf seinem Rechner hat zu strukturieren. Dazu nutzt das System ein sogenanntes Personal Information Model (PIMO) in Form einer Ontologie. Diese Ontologie enthält unter anderem eine Klasse "Topic", welche die wichtigste Grundlage für das Erstellen der in meiner Arbeit verwendeten Nutzerprofile bildete. Konkret wurde die RDF-Anfragesprache SPARQL verwendet, um eine Liste aller für den Nutzer relevanten Themen aus der Ontologie zu filtern. Die zentrale Idee meiner Arbeit war es nun diese Profilinformationen zur Optimierung des Ranking von Antwortkandidaten einzusetzen. In LogAnswer werden zu jeder gestellten Frage bis zu 200 potentiell relevante Textstellen aus der deutschen Wikipedia extrahiert. Diese Textstellen werden auf Basis von Eigenschaften (wie z.B. lexikalische Übereinstimmungen zwischen Frage und Textstelle) geordnet, da innerhalb des zur Verfügung stehenden Zeitlimits nicht alle Kandidaten bearbeitet werden können.
Mein Ansatz verfolgte das Ziel, diesen Algorithmus durch Nutzerprofile so zu erweitern, dass Antwortkandidaten, welche für den Benutzer relevante Informationen enthalten, höher in der Rangfolge eingeordnet werden. Zur Umsetzung dieser Idee musste eine Methode gefunden werden, um zu bestimmen ob ein Antwortkandidat mit dem Profil übereinstimmt. Da sich die in einer Textstelle enthaltenen Informationen in den meisten Fällen auf das übergeordnete Thema des Artikels beziehen, ohne den Namen des Artikels explizit zu erwähnen, wurde in meiner Implementierung der Artikelname betrachtet, um zu ermitteln, zu welchem Themengebiet die Textstelle Informationen liefert. Als zusätzliches Hilfsmittel wurde außerdem die DBpedia-Ontologie eingesetzt, welche die Informationen der Wikipedia strukturiert im RDF Format enthält. Mit Hilfe dieser Ontologie war es möglich, jeden Artikel in Kategorien einzuordnen, die dann mit den im Profil enthaltenen Stichworten verglichen wurden. Zur Untersuchung der Auswirkungen des Ansatzes auf das Ranking-Verfahren wurden mehrere Testläufe mit je 200 Testfragen durchgeführt. Die erste Testmenge bestand aus zufällig ausgewählten Fragen, die mit meinem eigenen Nutzerprofil getestet wurden. Dieser Testlauf lieferte kaum nutzbare Ergebnisse, da nur bei 29 der getesteten Fragen überhaupt ein Antwortkandidat mit dem Profil in Verbindung gebracht werden konnte. Außerdem konnte eine potentielle Verbesserung der Ergebnisse nur bei einer dieser 29 Fragen festgestellt werden, was zu der Schlussfolgerung führte, dass der Einsatz von Profildaten nicht für Anwendungsfälle geeignet ist, in denen die Fragen keine Korrelation mit dem genutzten Profil aufweisen.
Da die Grundannahme meiner Arbeit war, dass Nutzer in erster Linie Fragen zu den Interessensgebieten stellen, welche sich aus ihrem Profil ableiten lassen, sollten die weiteren Testläufe genau diesen Fall beleuchten. Dazu wurden 200 Testfragen aus dem Bereich Sport ausgewählt und mit einem Profil getestet, welches Stichworte zu unterschiedlichen Sportarten enthielt. Die Tests mit den Sportfragen waren wesentlich aussagekräftiger. Auch hier deuteten die Ergebnisse darauf hin, dass der Ansatz kein großes Potential zur Verbesserung des Rankings hat. Eine genauere Betrachtung einiger ausgewählter Beispiele zeigte allerdings, dass die Integration von Profildaten für bestimmte Anwendungsfälle, wie z.B. offene Fragen für die es mehr als eine korrekte Antwort gibt, durchaus zu einer Verbesserung der Ergebnisse führen kann. Außerdem wurde festgestellt, dass viele der schlechten Ergebnisse auf Inkosistenzen in der DBpedia-Ontologie und grundsätzliche Probleme im Umgang mit Wissensbasen in natürlicher Sprache beruhen.
Die Schlussfolgerung meiner Arbeit ist, dass der in dieser Arbeit vorgestellte Ansatz zur Integration von Profilinformationen für den aktuellen Anwendungsfall von LogAnswer nicht geeignet ist, da vor allem Faktenwissen aus sehr unterschiedlichen Domänen abgefragt wird und offene Fragen nur einen geringen Anteil ausmachen.
The term "Augmented Reality (AR)" denotes the superposition of additional virtual objects and supplementary information over real images. The joint project Enhanced Reality (ER)1 aims at a generic AR-system. The ER-project is a cooperation of six different research groups of the Department of Computer Science at the University of Koblenz-Landau. According to Ronald Azuma an AR-system combines real and virtual environments, where the real and virtual objects are registered in 3-D, and it provides interactivity in real time [Azu97]. Enhanced Reality extends Augmented Reality by requiring the virtual objects to be seamlessly embedded into the real world as photo-realistic objects according to the exact lighting conditions. Furthermore, additional information supplying value-added services may be displayed and interaction of the user may even be immersive. The short-term goal of the ER-project is the exploration of ER-fundamentals using some specific research scenarios; the long-term goal is the development of a component-based ER-framework for the creation of ER-applications for arbitrary application areas. ER-applications are developed as single-user applications for users who are moving in a real environment and are wearing some kind of visual output device like see-through glasses and some mobile end device. By these devices the user is able to see reality as it is, but he can also see the virtual objects and the additional information about some value-added service. Furthermore he might have additional devices whereby he can interact with the available virtual objects. The development of a generic framework for ER-applications requires the definition of generic components which are customizable and composable to build concrete applications and it requires a homogeneous data model which supports all components equally well. The workgroup "Software Technology"2 is responsible for this subproject. This report gives some preliminary results concerning the derivation of a component-based view of ER. There are several augmented reality frameworks like ARVIKA, AMIRE, DWARF, MORGAN, Studierstube and others which offer some support for the development of AR-applications. All of them ease the use of existing subsystems like AR-Toolkit, OpenGL and others and leverage the generation process for realistic systems by making efficient use of those subsystems. Consequently, they highly rely on them.
Tagging systems are intriguing dynamic systems, in which users collaboratively index resources with the so-called tags. In order to leverage the full potential of tagging systems, it is important to understand the relationship between the micro-level behavior of the individual users and the macro-level properties of the whole tagging system. In this thesis, we present the Epistemic Dynamic Model, which tries to bridge this gap between the micro-level behavior and the macro-level properties by developing a theory of tagging systems. The model is based on the assumption that the combined influence of the shared background knowledge of the users and the imitation of tag recommendations are sufficient for explaining the emergence of the tag frequency distribution and the vocabulary growth in tagging systems. Both macro-level properties of tagging systems are closely related to the emergence of the shared community vocabulary. rnrnWith the help of the Epistemic Dynamic Model, we show that the general shape of the tag frequency distribution and of the vocabulary growth have their origin in the shared background knowledge of the users. Tag recommendations can then be used for selectively influencing this general shape. In this thesis, we especially concentrate on studying the influence of recommending a set of popular tags. Recommending popular tags adds a feedback mechanism between the vocabularies of individual users that increases the inter-indexer consistency of the tag assignments. How does this influence the indexing quality in a tagging system? For this purpose, we investigate a methodology for measuring the inter-resource consistency of tag assignments. The inter-resource consistency is an indicator of the indexing quality, which positively correlates with the precision and recall of query results. It measures the degree to which the tag vectors of indexed resources reflect how the users perceive the similarity between resources. We argue with our model, and show it with a user experiment, that recommending popular tags decreases the inter-resource consistency in a tagging system. Furthermore, we show that recommending the user his/her previously used tags helps to increase the inter-resource consistency. Our measure of the inter-resource consistency complements existing measures for the evaluation and comparison of tag recommendation algorithms, moving the focus to evaluating their influence on the indexing quality.