Scientific visualization is concerned with facilitating discovery, exploration, and analysis of scientific datasets that arise from physical experiment and computer simulations. Its application areas cover all disciplines of computational science and include areas such as mechanical engineering, astrophysics, fluid dynamics, medicine and many more. While many physical techniques can be replicated using computational algorithms, scientific visualization techniques can go beyond typical experimental setups.
The lecture focuses on fundamental aspects of computational data representation and covers foundational visualization techniques for scalar fields, vector fields, and tensor fields. Here, both direct representations of data sets are covered, as well as indirect analysis approaches such as e.g. topological analysis that provide abstract summaries of data sets. Participants will be equipped with solid methodological background and a comprehensive overview of the most important contemporary techniques. Furthermore, problems and approaches in visualizing very large datasets – which exceed the size of main memory – are discussed. Last but not least, the lecture will cover the application of scientific visualization in practice, and available tools such as e.g. ParaView and the VTK library will be discussed.
As a successful application of scientific visualiztion requires a high level of practical proficiency, the exercises focus on practical matters and focus on the implementation of algorithms and an examination of their characteristics. Consequently, participants will conduct biweekly programming projects in small teams.