Contributions in Merlin

The quantitative comparison of the form of the endocast, which is the internal surface of the braincase, is a central issue in paleoanthropology (i.e., the study of human evolution based on fossil evidence). The major difficulty is that there are only few anatomical locations (so-called landmarks) defining biological correspondence between individual endocasts. Such point homologies, however, are the basis of the most powerful methods of morphometric analysis, the geometric morphometric toolbox. As a consequence, methods of geometric morphometrics are only of limited use in the analysis of endocranial form, and the morphometric analysis of such three-dimensional surfaces, which are poor in landmarks, is an open and lively discussed problem. In a quite different field, computer graphics, the parameterization of a surface embedded in the space $\mathbb{R}^3$ is a well-known problem. For that purpose, the surface is mapped to a two-dimensional parameter domain such as the plane or the sphere. The parameterization of a triangulated surface is a basic step in many applications in the area of geometry processing such as texture mapping, morphing, remeshing and data compression. Conformal, i.e., angle-preserving parameterizations, represents a special case. This thesis uses these techniques and methods to propose a new approach to the problem of quantitative comparison of endocranial surfaces with only few point homologies. For this purpose, concepts from geometric morphometrics are fused with concepts from computer graphics. Triangulated endocranial surfaces are conformally mapped to the unit sphere. The resulting spherical parameterizations are calibrated according to user-defined biological constraints. The result is a consistent coordinate system: a position on the unit sphere corresponds to a biologically homologous position on each surface in the sample. The sphere is a well-known, two-dimensional domain, and spherical functions can be expanded into a series of spherical harmonic functions. Thereby, a three-dimensional Fourier descriptor is produced, which permits shape analysis in frequency space. The integration of the new surface representations into the software MorphoTools permits the various statistical analyses, scientific visualization of the results and comparison with classic geometric morphometric methods. To test the new method, neurocrania of humans and great apes (gorillas, chimpanzees and bonobos) are analyzed. Der quantitative Vergleich der Form des Neuroendocraniums (Innenfläche des Hirnschädels; “Endocast”; Endocranium) ist ein zentrales Problem in der Paläoanthropologie (d.h. der Erforschung der menschlichen Evolution anhand fossiler Funde). Das Hauptproblem dabei ist, dass es nur wenige anatomische Fixpunkte (sog. Landmarken) auf dem Neurocranium gibt, die biologische Übereinstimmung zwischen individuellen Neurocrania definieren. Die leistungsstärksten Methoden der Geometrischen Morphometrie basieren allerdings auf solchen Punkthomologien und können deshalb nicht, oder nur bedingt, angewendet werden. Die biologische Analyse solcher dreidimensionalen Oberflächen, die arm an Landmarken sind, ist ein Problem, für das es bisher noch keine befriedigende Lösung gibt und das zurzeit rege diskutiert wird. In einem ganz anderen Forschungsgebiet, der Computergrafik, ist die Parametrisierung einer Fläche, die im Raum $\mathbb{R}^3$ eingebettet ist, ein wohlbekanntes Problem. Die Fläche wird dazu auf einen geeigneten zweidimensionalen Parameterraum, wie z.B. die Ebene oder die Sphäre, abgebildet. Die Parametrisierung einer triangulierten Oberfläche ist grundlegend für viele Anwendungen im Bereich Geometry-Processing, wie Texture Mapping, Morphing, Remeshing und Datenkompression. Ein Spezialfall sind konformale (d.h. winkeltreue) Abbildungen. Die vorliegende Arbeit befasst sich mit der Anwendung dieser Techniken und Methoden auf das Problem, Oberflächen mit nur wenigen Punkthomologien quantitativ und biologisch relevant zu vergleichen. Dazu werden Konzepte aus der Geometrischen Morphometrie und aus der Computergrafik verschmolzen. Triangulierte endocraniale Oberflächen werden konformal auf die Einheitssphäre abgebildet. Die sphärischen Parametrisierungen werden unter Einbezug von biologischen Bedingungen kalibriert. Das Resultat ist ein einheitliches Koordinatensystem: Eine bestimmte Position auf der Einheitskugel entspricht biologisch derselben Position auf jeder der zu vergleichenden Flächen. Die Sphäre ist eine wohlbekannte zweidimensionale Domäne, und sphärische Funktionen lassen sich in eine Reihe von sphärischen harmonischen Funktionen entwickeln. Dadurch wird ein dreidimensionaler Fourierdeskriptor erzeugt, der die Gestaltanalyse im Frequenzraum erlaubt. Die Einbindung der neuen Oberflächenrepräsentationen in die Software Morpho-Tools ermöglicht die Anwendung statistischer Analysen, wissenschaftliche Visualisierungen der Resultate sowie den Vergleich mit rein landmarkbasierten Methoden. Als Test für die neue Methode werden Neurocrania von Menschen und Menschenaffen (Gorillas, Schimpansen und Zwergschimpansen) analysiert.

This paper presents a model in which final goods producers outsource intermediate input production. Intermediate inputs are differentiated and their production can be located at home or abroad. The model is used to examine competitive location policy in a (two-country) free trade area (FTA). It is shown that national public infrastructure investment has a positive effect on both the number of intermediate input producers and the return to the immobile factor in the home country. International outsourcing from home declines. Opposite effects are triggered in the partner country. In a welfare analysis we characterize national infrastructure policies that aim to maximize national income (net of tax costs) and compare the non-cooperative FTA-equilibrium with optimal policies from an integrated point of view. We show whether or not there is a need for policy coordination. Firm subsidies are discussed as an alternative to public infrastructure investment.

We propose a test for the presence of money illusion based on subjective survey information on individual satisfaction with income. Using data from the German Socio-Economic Panel for the period 1993–2003, we cannot reject the null hypothesis of no money illusion.

null

In a laboratory experiment we show that minimum wages have significant and lasting effects on subjects’ reservation wages. The temporary introduction of a minimum wage leads to a rise in subjects’ reservation wages which persists even after the minimum wage has been removed. Firms are therefore forced to pay higher wages after the removal of the minimum wage than before its introduction. As a consequence, the employment effects of removing the minimum wage are significantly smaller than are the effects of its introduction. The impact of minimum wages on reservation wages may also explain the anomalously low utilization of subminimum wages if employers are given the opportunity of paying less than a minimum wage previously introduced. It may further explain why employers often increase workers' wages after an increase in the minimum wage by an amount exceeding that necessary for compliance with the higher minimum. At a more general level, our results suggest that economic policy may affect people’s behavior by shaping the perception of what is a fair transaction and by creating entitlement effects.

null

The fundamental drawback of current stereo and multi-view visualization is the necessity to perform multi pass rendering (one pass for each view) and subsequent image composition + masking for generating multiple stereo views. Thus the rendering time increases in general linearly with the number of views. In this paper we introduce a new method for multi-view splatting based on deferred blending. Our method exploits the programmability of modern graphic processing units (GPUs) for rendering multiple stereo views in a single rendering pass. The views are calculated directly on the GPU including sub-pixel wavelength selective views. We describe our algorithm precisely and provide details about its implementation. Experimental results demonstrate the performance advantage of our multi-view point splatting algorithm compared to the standard multi-pass approach.

null

Mobile learning has a lot of potential for supporting learning in situations such as in a museum, at a tourist sight or when exploring biological phenomena at a riverside. There learners can interact with their environment and still make use of the advantages of computational power. However, we have found many of such projects hindered by placing the technology too much in the focus of the learner. Instead of interacting with the environment, we found the learners interacting with the device, heads down and ignoring the environment. We found the issue of focus to be a massive problem, one which needs a completely new metaphor for the design of an educational and technical setting. Until now, the mobile devices have been interpreted as small desktops, always in the foreground of the learners' focus. Instead, we propose a different approach, deduced from the usage of mobile phones. Mobile applications need to be designed explicitly to free the learners' focus and push the application to the background. The good news is that the actual changes to be made in existing systems are not as fundamental as one may think.