New visual analytics approaches for volume data time series in materials science

Supervisors: Prof. Martin Müller (HZG), Prof. Stephan Olbrich (UHH)

In materials science, experiments to study dynamical processes within materials are conducted within tomographic X-ray detectors (examples include material changes within medical implants and solidification processes in magnesium). These detectors generate large, time-varying volumetric datasets within very short timespans. For successful experiments, detailed analysis of dynamical (i.e., time-dependent) 3D structures in the acquired data is crucial, both for immediate analysis to refine experiment setups, as well as for later in-depth analysis.

The objective of this PhD project is to investigate how novel visual analytics approaches can be used to study dynamical 3D structures in time-dependent volumetric data from tomographic scans. You will perform research into methods for structure detection, tracking, and interactive visual analysis, that is, you will deal with techniques from computer science fields including visualization, computer vision, artificial intelligence, data mining, and computer graphics. The goal is to develop novel approaches to capture structures in the data that are of crucial relevance for the success of the experiments. Focus will be put on interactive near-real-time analysis for use during experiment runtime. We will evaluate and deploy the developed methods together with domain scientists.

The position will be integrated into the Visual Data Analysis Group at the Regional Computing Centre of Universität Hamburg (RRZ-VDA, http://uhh.de/rrz-research), in close collaboration with the German Engineering Materials Science Center (GEMS, https://www.hzg.de/institutes_platforms/gems/) operated by HZG. The successful candidate is expected to show ample interest for frequent exchange with our collaboration partners in material science, and to integrate into the RRZ-VDA group specializing in interactive visual analysis, feature detection, uncertainty analysis, and large data.