Kaiserslautern - Fachbereich Informatik
Refine
Year of publication
- 2012 (14) (remove)
Document Type
- Doctoral Thesis (7)
- Master's Thesis (3)
- Report (3)
- Conference Proceeding (1)
Has Fulltext
- yes (14)
Keywords
- Bildverarbeitung (1)
- Bioinformatik (1)
- Carbon footprint (1)
- Fault Tree Analysis (1)
- Geoinformationssystem (1)
- Geovisualization (1)
- Information Extraction (1)
- Information Visualization (1)
- Layout (1)
- Molekulare Bioinformatik (1)
Faculty / Organisational entity
The development of autonomous vehicle systems demands the increased usage of software based control mechanisms. Generally, this leads to very complex systems, whose proper functioning has to be ensured. In our work we aim at investigating and assessing the potential effects of software issues on the safety, reliability and availability of complex embedded autonomous systems. One of the key aspects of the research concerns the mapping of functional descriptions in form of integrated behavior-based control networks to State-Event Fault Tree models.
Hadoop ist ein beliebtes Framework für verteilte Berechnungen über große
Datenmengen (Big Data) mittels MapReduce. Hadoop zu verwenden ist einfach: Der
Entwickler definiert die Eingabedatenquelle und implementiert die beiden
Methoden Map und Reduce. Über die verteilte Berechnung und Fehlerbehandlung muss
er sich dabei keine Gedanken machen, das erledigt das Hadoop-Framework.
Allerdings kann die Analyse von Big Data sehr lange dauern und da sich die
Eingabedaten jede Sekunde ändern, ist es vielleicht nicht immer die beste
Idee, die vollständige Berechnung jedes Mal aufs Neue auf die kompletten
Eingabedaten anzuwenden. Es wäre geschickter, sich das Ergebnis der
vorherigen Berechnung zu betrachten und nur die Deltas zu analysieren, also
Daten, die seit der letzten Berechnung hinzugefügt oder gelöscht wurden. In dem Gebiet der
selbstwartbaren materialisierten Sichten in relationalen Datenbanksystemen gibt
es bereits mehrere Ansätze, die sich mit der Lösung dieses Problems
beschäftigen. Eine Strategie liest nur die Deltas und inkrementiert oder
dekrementiert die Ergebnisse der vorherigen Berechnung. Allerdings ist diese
Inkrement-Operation sehr teuer, deshalb ist es manchmal besser, das komplette
alte Ergebnis zu lesen und es mit den Deltas der Eingabedaten zu kombinieren.
In dieser Masterarbeit wird ein neues Framework namens Marimba vorgestellt,
welches sich genau um diese Probleme der inkrementellen Berechnung kümmert. Einen
Map\-Re\-duce-Job in Marimba zu schreiben ist genau so einfach wie einen Hadoop-Job.
Allerdings werden hier keine Mapper- und Reducer-Klasse implementiert, sondern
eine Translator- und Serializer-Klasse. Der Translator ähnelt dem Mapper: Er
bestimmt, wie die Eingabedaten gelesen und daraus Zwischenwerte berechnet
werden. Der Serializer erzeugt die Ausgabe des Jobs. Wie diese Ausgabe berechnet
wird, gibt der Benutzer durch Implementierung einiger Methoden an, um Werte zu
aggregieren und invertieren.
Vier MapReduce-Jobs, darunter auch das Paradebeispiel für MapReduce WordCount,
wurden im Marimba-Framework implementiert. Das Entwickeln von inkrementellen
Map-Reduce-Jobs ist mit dem Framework extrem einfach geworden. Außerdem konnte
mit Performanztests gezeigt werden, dass die inkrementelle Berechnung deutlich
schneller ist als eine vollständige Neuberechnung.
Ein weiterer unter den vier implementierten Jobs berechnet
Wortauftrittswahrscheinlichkeiten in geschriebenen Sätzen. Dies kann
beispielsweise für Spracherkennung verwendet werden. Wenn ein Wort in einer
gesprochenen SMS nicht richtig verstanden wurde, hilft der Algorithmus zu raten,
welches Wort am wahrscheinlichsten an einer bestimmten Stelle stehen könnte,
abhängig von den vorherigen Wörtern im Satz. Damit dieser Algorithmus auch
brauchbare Ergebnisse liefert, ist die Menge und die Qualität der Eingabedaten
wichtig. Durchaus brauchbare Ergebnisse wurden durch die Verarbeitung von
Millionen von Twitter-Feeds, die deutsche Twitter-Nutzer in den letzten Monaten
geschrieben haben, erreicht.
Recently, a new Quicksort variant due to Yaroslavskiy was chosen as standard sorting
method for Oracle's Java 7 runtime library. The decision for the change was based on
empirical studies showing that on average, the new algorithm is faster than the formerly
used classic Quicksort. Surprisingly, the improvement was achieved by using a dual pivot
approach — an idea that was considered not promising by several theoretical studies in the
past. In this thesis, I try to find the reason for this unexpected success.
My focus is on the precise and detailed average case analysis, aiming at the flavor of
Knuth's series “The Art of Computer Programming”. In particular, I go beyond abstract
measures like counting key comparisons, and try to understand the efficiency of the
algorithms at different levels of abstraction. Whenever possible, precise expected values are
preferred to asymptotic approximations. This rigor ensures that (a) the sorting methods
discussed here are actually usable in practice and (b) that the analysis results contribute to
a sound comparison of the Quicksort variants.
Generic layout analysis--process of decomposing document image into homogeneous regions for a collection of diverse document images--has many important applications in document image analysis and understanding such as preprocessing of degraded warped, camera-captured document images, high performance layout analysis of document images containing complex cursive scripts, and word spotting in historical document images at page level. Many areas in this field like generic text line extraction method are considered as elusive goals so far, still beyond the reach of the state-of-the-art methods [NJ07, LSZT07, KB06]. This thesis addresses this problem in such a way that it presents generic, domain-independent, text line extraction and text and non-text segmentation methods, and then describes some important applications, that were developed based on these methods. An overview of the key contributions of this thesis is as follows.
The first part of this thesis presents a generic text line extraction method using a combination of matched filtering and ridge detection techniques, which are commonly used in computer vision. Unlike the state-of-the-art text line extraction methods in the literature, the generic text line extraction method can be equally and robustly applied to a large variety of document image classes including scanned and camera-captured documents, binary and grayscale documents, typed-text and handwritten documents, historical and contemporary documents, and documents containing different scripts. Different standard datasets are selected for performance evaluation that belong to different categories of document images such as the UW-III [GHHP97] dataset of scanned documents, the ICDAR 2007 [GAS07] and the UMD [LZDJ08] datasets of handwritten documents, the DFKI-I [SB07] dataset of camera-captured documents, Arabic/Urdu script documents dataset, and German calligraphic (Fraktur) script historical documents dataset. The generic text line extraction method achieves 86% (n = 23,763 text lines in 650 documents) text line detection accuracy which is better than the aggregate accuracy of 73% of the best performing domain-specific state-of-the-art methods. To the best of the author's knowledge, it is the first general-purpose text line extraction method that can be equally used for a diverse collection of documents.
This thesis also presents an active contour (snake) based curled text line extraction method for warped, camera-captured document images. The presented approach is applied to DFKI-I [SB07] dataset of camera-captured, Latin script document images for curled text line extraction. It achieves above 95% (n = 3,091 text lines in 102 documents) text line detection accuracy, which is significantly better than the competing state-of-the-art curled text line extraction methods. The presented text line extraction method can also be applied to document images containing different scripts like Chinese, Devanagari, and Arabic after small modifications.
The second part of this thesis presents an improved version of the state-of-the-art multiresolution morphology (Leptonica) based text and non-text segmentation method [Blo91], which is a domain-independent page segmentation approach and can be equally applied to a diverse collection of binarized document images. It is demonstrated that the presented improvements result in an increase in segmentation accuracy from 93% to 99% (n = 113 documents).
This thesis also introduces a discriminative learning based approach for page segmentation, where a self-tunable multi-layer perceptron (MLP) classifier [BS10] is trained for distinguishing between text and non-text connected components. Unlike other classification based page segmentation approaches in the literature, the connected components based discriminative learning based approach is faster than pixel based classification methods and does not require a block segmentation method beforehand. A segmentation accuracy of $96\%$ ($n = 113$ documents) is achieved in comparison to the state-of-the-art multiresolution morphology (Leptonica) based page segmentation method [Blo91] that achieves a segmentation accuracy of 93%. In addition to text and non-text segmentation of Latin script documents, the presented approach can also be adapted for document images containing other scripts as well as for other specialized layout analysis tasks such as digit and non-digit segmentation [HBSB12], orientation detection [RBSB09], and body-text and side-note segmentation [BAESB12].
Finally, this thesis presents important applications of the two generic layout analysis techniques, ridge-based text line extraction method and the multi-resolution morphology based text and non-text segmentation method, discussed above. First, a complete preprocessing pipeline is described for removing different types of degradations from grayscale warped, camera-captured document images that includes removal of grayscale degradations such as non-uniform shadows and blurring through binarization, noise cleanup applying page frame detection, and document rectification using monocular dewarping. Each of these preprocessing steps shows significant improvement in comparison to the analyzed state-of-the-art methods in the literature. Second, a high performance layout analysis method is described for complex Arabic script document images written in different languages such as Arabic, Urdu, and Persian and different styles for example Naskh and Nastaliq. The presented layout analysis system is robust against different types of document image degradations and shows better performance for text and non-text segmentation, text line extraction, and reading order determination on a variety of Arabic and Urdu document images as compared to the state-of-the-art methods. It can be used for large scale Arabic and Urdu documents' digitization processes. These applications demonstrate that the layout analysis methods, ridge-based text line extraction and the multi-resolution morphology based text and non-text segmentation, are generic and can be applied easily to a large collection of diverse document images.
This research for this thesis was conducted to develop a framework which supports the automatic configuration of project-specific software development processes by selecting and combining different technologies: the Process Configuration Framework. The research draws attention to the problem that while the research community develops new technologies, the industrial companies continue only using their well-known ones. Because of this, technology transfer takes decades. In addition, there is the fact that there is no solution which solves all problems in a software development project. This leads to a number of technologies which need to be combined for one project.
The framework developed and explained in this research mainly addresses those problems by building a bridge between research and industry as well as by supporting software companies during the selection of the most appropriate technologies combined in a software process. The technology transformation gap is filled by a repository of (new) technologies which are used as a foundation of the Process Configuration Framework. The process is configured by providing SPEM process pattern for each technology, so that the companies can build their process by plugging into each other.
The technologies of the repository were specified in a schema including a technology model, context model, and an impact model. With context and impact it is possible to provide information about a technology, for example, its benefits to quality, cost or schedule. The offering of the process pattern as output of the Process Configuration Framework is performed in several stages:
I Technology Ranking:
1 Ranking based on Application Domain, Project & Impact
2 Ranking based on Environment
3 Ranking based on Static Context
II Technology Combination:
4 Creation of all possible Technology Chains
5 Restriction of the Technology Chains
6 Ranking based on Static and Dynamic Context
7 Extension of the Chains by Quality Assurance
III Process Configuration:
8 Process Component Diagram
9 Extension of the Process Component Diagram
10 Instantiation of the Components by Technologies of the Technology Chain
11 Providing process patterns
12 Creation of the process based on Patterns
The effectiveness and quality of the Process Configuration Framework have additionally been evaluated in a case study. Here, the Technology Chains manually created by experts were compared to the chains automatically created by the framework after it was configured by those experts. This comparison depicted that the framework results are similar and therefore can be used as a recommendation.
We conclude from our research that support during the configuration of a process for software projects is important especially for non-experts. This support is provided by the Process Configuration Framework developed in this research. In addition our research has shown that this framework offers a possibility to speed up the technology transformation gap between the research community and industrial companies.
The safety of embedded systems is becoming more and more important nowadays. Fault Tree Analysis (FTA) is a widely used technique for analyzing the safety of embedded systems. A standardized tree-like structure called a Fault Tree (FT) models the failures of the systems. The Component Fault Tree (CFT) provides an advanced modeling concept for adapting the traditional FTs to the hierarchical architecture model in system design. Minimal Cut Set (MCS) analysis is a method that works for qualitative analysis based on the FTs. Each MCS represents a minimal combination of component failures of a system called basic events, which may together cause the top-level system failure. The ordinary representations of MCSs consist of plain text and data tables with little additional supporting visual and interactive information. Importance analysis based on FTs or CFTs estimates the contribution of each potential basic event to a top-level system failure. The resulting importance values of basic events are typically represented in summary views, e.g., data tables and histograms. There is little visual integration between these forms and the FT (or CFT) structure. The safety of a system can be improved using an iterative process, called the safety improvement process, based on FTs taking relevant constraints into account, e.g., cost. Typically, relevant data regarding the safety improvement process are presented across multiple views with few interactive associations. In short, the ordinary representation concepts cannot effectively facilitate these analyses.
We propose a set of visualization approaches for addressing the issues above mentioned in order to facilitate those analyses in terms of the representations.
Contribution:
1. To support the MCS analysis, we propose a matrix-based visualization that allows detailed data of the MCSs of interest to be viewed while maintaining a satisfactory overview of a large number of MCSs for effective navigation and pattern analysis. Engineers can also intuitively analyze the influence of MCSs of a CFT.
2. To facilitate the importance analysis based on the CFT, we propose a hybrid visualization approach that combines the icicle-layout-style architectural views with the CFT structure. This approach facilitates to identify the vulnerable components taking the hierarchies of system architecture into account and investigate the logical failure propagation of the important basic events.
3. We propose a visual safety improvement process that integrates an enhanced decision tree with a scatter plot. This approach allows one to visually investigate the detailed data related to individual steps of the process while maintaining the overview of the process. The approach facilitates to construct and analyze improvement solutions of the safety of a system.
Using our visualization approaches, the MCS analysis, the importance analysis, and the safety improvement process based on the CFT can be facilitated.
Predicting secondary structures of RNA molecules is one of the fundamental problems of and thus a challenging task in computational structural biology. Existing prediction methods basically use the dynamic programming principle and are either based on a general thermodynamic model or on a specific probabilistic model, traditionally realized by a stochastic context-free grammar. To date, the applied grammars were rather simple and small and despite the fact that statistical approaches have become increasingly appreciated over the past years, a corresponding sampling algorithm based on a stochastic RNA structure model has not yet been devised. In addition, basically all popular state-of-the-art tools for computational structure prediction have the same worst-case time and space requirements of O(n^3) and O(n^2) for sequence length n, limiting their applicability for practical purposes due to the often quite large sizes of native RNA molecules. Accordingly, the prime demand imposed by biologists on computational prediction procedures is to reach a reduced waiting time for results that are not significantly less accurate.
We here deal with all of these issues, by describing algorithms and performing comprehensive studies that are based on sophisticated stochastic context-free grammars of similar complexity as those underlying thermodynamic prediction approaches, where all of our methods indeed make use of the concept of sampling. We also employ the approximation technique known from theoretical computer science in order to reach a heuristic worst-case speedup for RNA folding.
Particularly, we start by describing a way for deriving a sequence-independent random sampler for an arbitrary class of RNAs by means of (weighted) unranking. The resulting algorithm may generate any secondary structure of a given fixed size n in only O(n·log(n)) time, where the results are observed to be accurate, validating its practical applicability.
With respect to RNA folding, we present a novel probabilistic sampling algorithm that generates statistically representative and reproducible samples of the entire ensemble of feasible structures for a particular input sequence. This method actually samples the possible foldings from a distribution implied by a suitable (traditional or length-dependent) grammar. Notably, we also propose several (new) ways for obtaining predictions from generated samples. Both variants have the same worst-case time and space complexities of O(n^3) and O(n^2) for sequence length n. Nevertheless, evaluations of our sampling methods show that they are actually capable of producing accurate (prediction) results.
In an attempt to resolve the long-standing problem of reducing the time complexity of RNA folding algorithms without sacrificing much of the accuracy of the results, we invented an innovative heuristic statistical sampling method that can be implemented to require only O(n^2) time for generating a fixed-size sample of candidate structures for a given sequence of length n. Since a reasonable prediction can still efficiently be obtained from the generated sample set, this approach finally reduces the worst-case time complexity by a liner factor compared to all existing precise methods. Notably, we also propose a novel (heuristic) sampling strategy as opposed to the common one typically applied for statistical sampling, which may produce more accurate results for particular settings. A validation of our heuristic sampling approach by comparison to several leading RNA secondary structure prediction tools indicates that it is capable of producing competitive predictions, but may require the consideration of large sample sizes.
Dealing with information in modern times involves users to cope with hundreds of thousands of documents, such as articles, emails, Web pages, or News feeds.
Above all information sources, the World Wide Web presents information seekers with great challenges.
It offers more text in natural language than one is capable to read.
The key idea for this research intends to provide users with adaptable filtering techniques, supporting them in filtering out the specific information items they need.
Its realization focuses on developing an Information Extraction system,
which adapts to a domain of concern, by interpreting the contained formalized knowledge.
Utilizing the Resource Description Framework (RDF), which is the Semantic Web's formal language for exchanging information,
allows extending information extractors to incorporate the given domain knowledge.
Because of this, formal information items from the RDF source can be recognized in the text.
The application of RDF allows a further investigation of operations on recognized information items, such as disambiguating and rating the relevance of these.
Switching between different RDF sources allows changing the application scope of the Information Extraction system from one domain of concern to another.
An RDF-based Information Extraction system can be triggered to extract specific kinds of information entities by providing it with formal RDF queries in terms of the SPARQL query language.
Representing extracted information in RDF extends the coverage of the Semantic Web's information degree and provides a formal view on a text from the perspective of the RDF source.
In detail, this work presents the extension of existing Information Extraction approaches by incorporating the graph-based nature of RDF.
Hereby, the pre-processing of RDF sources allows extracting statistical information models dedicated to support specific information extractors.
These information extractors refine standard extraction tasks, such as the Named Entity Recognition, by using the information provided by the pre-processed models.
The post-processing of extracted information items enables representing these results in RDF format or lists, which can now be ranked or filtered by relevance.
Post-processing also comprises the enrichment of originating natural language text sources with extracted information items by using annotations in RDFa format.
The results of this research extend the state-of-the-art of the Semantic Web.
This work contributes approaches for computing customizable and adaptable RDF views on the natural language content of Web pages.
Finally, due to the formal nature of RDF, machines can interpret these views allowing developers to process the contained information in a variety of applications.
One of the fundamental problems in computational structural biology is the prediction of RNA secondary structures from a single sequence. To solve this problem, mainly two different approaches have been used over the past decades: the free energy minimization (MFE) approach which is still considered the most popular and successful method and the competing stochastic context-free grammar (SCFG) approach. While the accuracy of the MFE based algorithms is limited by the quality of underlying thermodynamic models, the SCFG method abstracts from free energies and instead tries to learn about the structural behavior of the molecules by training the grammars on known real RNA structures, making it highly dependent on the availability of a rich high quality training set. However, due to the respective problems associated with both methods, new statistics based approaches towards RNA structure prediction have become increasingly appreciated. For instance, over the last years, several statistical sampling methods and clustering techniques have been invented that are based on the computation of partition functions (PFs) and base pair probabilities according to thermodynamic models. A corresponding SCFG based statistical sampling algorithm for RNA secondary structures has been studied just recently. Notably, this probabilistic method is capable of producing accurate (prediction) results, where its worst-case time and space requirements are equal to those of common RNA folding algorithms for single sequences.
The aim of this work is to present a comprehensive study on how enriching the underlying SCFG by additional information on the lengths of generated substructures (i.e. by incorporating length-dependencies into the SCFG based sampling algorithm, which is actually possible without significant losses in performance) affects the reliability of the induced RNA model and the accuracy of sampled secondary structures. As we will see, significant differences with respect to the overall quality of generated sample sets and the resulting predictive accuracy are typically implied. In principle, when considering the more specialized length-dependent SCFG model as basis for statistical sampling, a higher accuracy of predicted foldings can be reached at the price of a lower diversity of generated candidate structures (compared to the more general traditional SCFG variant or sampling based on PFs that rely on free energies).