This work describes a novel software tool for visualizing anatomical segmentations of medical images. It was developed as part of a bachelor's thesis project, with a view to supporting research into automatic anatomical brain image segmentation. The tool builds on a widely-used visualization approach for 3D image volumes, where sections in orthogonal directions are rendered on screen as 2D images. It implements novel display modes that solve common problems with conventional viewer programs. In particular, it features a double-contour display mode to aid the user's spatial orientation in the image, as well as modes for comparing two competing segmentation labels pertaining to one and the same anatomical region. The tool was developed as an extension to an existing open-source software suite for medical image processing. The visualization modes are, however, suitable for implementation in the context of other viewer programs that follow a similar rendering approach. The modified code can be found here: soundray.org/mm-segmentation-visualization.tar.gz.

Eine Untersuchung verschiedener Algorithmen aus dem Themengebiet des maschinellen Lernens anhand des Problems, desReizens beim Skat.

Our work finds the fine grained edits in context of neighbouring tokens in Wikipedia articles. We cluster those edits according to similar neighbouring context. We encode neighbouring context into vector space using word vectors. We evaluate clusters returned by our algorithm on extrinsic and intrinsic metric and compare it with previous work. We analyse the relation between extrinsic and intrinsic measurements of fine grained edit tokens.

Over the past few decades society’s dependence on software systems has grown significantly. These systems are utilized in nearly every matter of life today and often handle sensitive, private data. This situation has turned software security analysis into an essential and widely researched topic in the field of computer science. Researchers in this field tend to make the assumption that the quality of the software systems' code directly affects the possibility for security gaps to arise in it. Because this assumption is based on properties of the code, proving it true would mean that security assessments can be performed on software, even before a certain version of it is released. A study based on this implication has already attempted to mathematically assess the existence of such a correlation, studying it based on quality and security metric calculations. The present study builds upon that study in finding an automatic method for choosing well-fitted software projects as a sample for this correlation analysis and extends the variety of projects considered for the it. In this thesis, the automatic generation of graphical representations both for the correlations between the metrics as well as for their evolution is also introduced. With these improvements, this thesis verifies the results of the previous study with a different and broader project input. It also focuses on analyzing the correlations between the quality and security metrics to real-world vulnerability data metrics. The data is extracted and evaluated from dedicated software vulnerability information sources and serves to represent the existence of proven security weaknesses in the studied software. The study discusses some of the difficulties that arise when trying to gather such information and link it to the difference in the information contained in the repositories of the studied projects. This thesis confirms the significant influence that quality metrics have on each other. It also shows that it is important to view them together as a whole and suppose that their correlation could influence the appearance of unwanted vulnerabilities as well. One of the important conclusions I can draw from this thesis is that the visualization of metric evolution graphs, helps the understanding of the values as well as their connection to each other in a more meaningful way. It allows for better grasp of their influence on each other as opposed to only studying their correlation values. This study confirms that studying metric correlations and evolution trends can help developers improve their projects and prevent them from becoming difficult to extend and maintain, increasing the potential for good quality as well as more secure software code.

Zum Inhalt: Die Dissertation ist in einem empirisch-qualitativen Forschungssetting eingebettet, bei der das Team-Teaching als Unterrichtsmethode im Forschungsfokus steht. Die Erhebung des empirischen Datenmaterials erfolgt einerseits mittels teilnehmender Beobachtung im Forschungsfeld und andererseits durch ein Gruppeninterview, das mit Lehrkräften geführt wird, die über einen gewissen Beobachtungszeitraum Erfahrungen mit der Unterrichtsmethode des Team-Teaching gesammelt haben. Für die hier zu Grunde gelegte Feldforschung wird das „theoretical sampling“ zum Einsatz kommen, das 1967 von Glaser und Strauss im Rahmen einer empirischen Untersuchung begründet worden ist. Zum Aufbau: In der Einleitung wird das Forschungsinteresse, der Forschungsstand und die zentrale Forschungsfrage der Dissertation beschrieben. Im zweiten Kapitel folgt der theoretische Teil mit einer differenzierten Definition des Team-Teachings als Unterrichtsmethode. Es folgt eine allgemeine Definition zum Unterricht, um die spezifischen Merkmale des Team-Teachings differenziert herausarbeiten zu können. Die empirischen Daten werden im Rahmen einer teilnehmenden Beobachtung im Unterricht der Oberstufe einer Förderschule mit dem Förderschwerpunkt Lernen und sozial-emotionale Entwicklung erhoben. Deshalb sind im Theorieteil der Förderschwerpunkt und die curricularen Aspekte zum Förderschwerpunkt dargestellt. Die Bedeutung der Unterrichtsmethode des Team-Teachings für das Unterrichtssetting in einer Förderschule mit dem sozial-emotionalem Förderschwerpunkt ist ebenfalls im Theorieteil verortet. Im dritten Kapitel werden die qualitativen und quantitativen Erhebungsinstrumente beschrieben und Hypothesen formuliert. Die Darstellung der Methoden zur Aufbereitung des Datenmaterials ist im vierten Kapitel verortet. Die Interpretation der Forschungsergebnisse erfolgt im fünften Kapitel. Die qualitative und quantitative Datenauswertung des empirischen Datenmaterials ist im sechsten Kapitel dargestellt. Die Dissertation schließt im siebten Kapitel mit einen Fazit und Ausblick.