Interaktive Erstellung von wissenschaftlichen Illustrationen

  • In dieser Arbeit werden Verfahren vorgestellt, mit dem sich hochaufgelöste wissenschaftliche Illustrationen in einem interaktiven Vorgang erstellen lassen. Die Basis dafür bildet die neu eingeführte GPU-basierte Illustrations-Pipeline, in der auf Grundlage eines 3D-Modells Bildebenen frei angelegt und miteinander kombiniert werden können. In einer Ebene wird ein bestimmter Aspekt der Illustration mit einer auswählbaren Technik gezeigt. Die Parameter der Technik sind interaktiv editierbar. Um Effizienz zu gewährleisten ist das gesamte Verfahren so konzipiert, dass es soweit wie möglich die Berechnungen auf der GPU durchführt. So ist es möglich, dass die Illustrationen mit interaktiven Frameraten gerendert werden.
  • Scientific Illustrations have been used since the sixteenth century to depict new scientific theories and findings. Today they are playing an ever more important role in the publication of new research articles, not only in natural sciences like biology or medicine, but also in general sciences like archaeology or palaeontology. They are often created via a tedious and time consuming process by specially trained drafters. An illustration is based on a concrete specimen, but usually shows an abstracted or idealised object to express a theory. In the field of non-photorealistic-rendering many drawing techniques that are used in the creation of scientific illustrations have been researched for centuries. Current algorithmic methods generate aesthetically pleasing illustrations that are suitable for publication in a scientific context. Unfortunately these methods are predominantly created for an off-line and non-interactive scenario as the calculation of the final images takes a long time. The creation of a desired custom illustration using these algorithms is hence a time-consuming task. In this work a framework is presented with which high-dpi scientific illustrations and schematics can be created interactively by a non-illustrating researcher. It is based on an extensible GPU-based illustration-pipeline to interactively render characteristics of a 3d-model into image-layers that are combined in an image-editing fashion. The user can choose the techniques used to render each layer and manipulate its key aspects and parameters. The pipeline is constructed to utilise the GPU as much as possible to ensure efficiency. An important new aspect is the space of abstraction- and correction-possibilities that the pipeline allows, realised using drawing functions. The first allows the user to deform and sculpt the 3d-model and create the abstraction of a concrete specimen. The user can furthermore draw both on the model and into the layers as well as adjust the results of the algorithms to his liking. The drawing functions are realised completely on the GPU and allow points, lines and parametric curves to be drawn. As well as being able to load 3d-models the framework also supports the processing of fixed-focus image stacks as they are created by motor-focus microscopes. A state-of-the-art algorithm has been extended to reconstruct the 3d-surface on the GPU to create better results more efficiently. The main hypotheses of this work have been validated by a qualitative evaluation of a prototype implementation and a quantitative evaluation through an online survey. The main use case of such a framework is the creation of scientific illustrations by non-illustrating researchers. They are able to experiment with different viewing and illumination scenarios and parameters interactively and thereby create sketches that can be used to discuss a work order of a scientific illustration drawn by trained drafters. The online survey has also shown that the scientific illustrations created with the prototype are publishable in a scientific context. Non-illustrating researchers that don't have access to trained drafters are therefore able to create scientific illustrations that are accepted by the community. The prototype was successfully employed by users in the evaluation, and it also showed that there is interest and potential for further research and extension. Some users asked for the prototype to be extended to also process volumetric data, while others were interested in adding more drawing techniques to extend the scope of the prototype outside just the palaeontological domain at which it was targeted.

Download full text files

Export metadata

Author:Sebastian Schäfer
Place of publication:Frankfurt am Main
Referee:Detlef Krömker, Matthias KaschubeORCiDGND
Advisor:Detlef Krömker
Document Type:Doctoral Thesis
Date of Publication (online):2016/01/22
Year of first Publication:2015
Publishing Institution:Universitätsbibliothek Johann Christian Senckenberg
Granting Institution:Johann Wolfgang Goethe-Universität
Date of final exam:2016/01/19
Release Date:2016/01/22
Page Number:190
Institutes:Informatik und Mathematik / Informatik
Dewey Decimal Classification:0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 000 Informatik, Informationswissenschaft, allgemeine Werke
Licence (German):License LogoDeutsches Urheberrecht