ESR 07

Generative prediction of weathering and aged appearance

-Open Position-


Topic Description

The visual appearance of real-world materials does not remain static over time, as the physical objects are subject to environmental effects. Faithfully predicting such changes has a wide range of applications. In industrial designs, certain surfaces attract fingerprints right away, others scratch easily, colours fade through light exposure, and surfaces attract dirt or dust. In architectural applications, the influence of weather and seasons changes materials over time, even to the extent of erosion and build-up of matter. In computer graphics, realistic wear of surfaces is often what separates a “photo-real” look from an unrealistically clean “computer graphics” appearance.

Based on the state of the art in both machine learning and physical modelling, the ESR 07 PhD candidate within PRIME will be developing phenomenological models of material ageing, to plausibly age objects across a wide range of materials and environments, including previously unobserved materials, achieving plausible, photo-realistic renditions of their weathered and worn surface characteristics. These models will be driven by data from physical simulations and from photographs under controlled conditions and time series. The work will be supported through secondments and collaboration with industry leaders in surface appearance authoring.

About the research group

The ESR will be hosted by the Digital Reality group (led by Prof. Tim Weyrich), in the Department of Computer Science at FAU Erlangen-Nürnberg. The group’s research is dedicated to bridging reality and the digital domain by drawing from computer graphics, computer vision, machine learning and trans-disciplinary domains, combining well-grounded engineering with a scientific approach to phenomena in the real world. The ESR will benefit from the thriving ecosystem of FAU, one of Germany‘s oldest and most distinguished universities, and yet, ranked as “most innovative university in Germany” (Reuter Ranking), with close ties to world-leading high-tech industry nearby.

Employment details

Pay is standardised by the EU, allows for comfortable living as a student by local standards, and adjusted for local cost of living per country.
Employment duration within PRIME is 36 months. We prefer candidates to be available as soon as possible.


Formal Requirements

  • Master’s degree in Computer Science, Applied Mathematics, Physics or Engineering. For exceptional candidates, Bachelor’s degree in the same disciplines will also be considered.
  • No Ph.D. yet.
  • Less than 3 years of professional employment after obtaining your master’s degree.
  • Less than 12 months of residency in Germany in the last 3 years.


Technical Requirements

  • Some experience with deep learning using TensorFlow or PyTorch.
  • Preference will be given to candidates with experience with training large neural networks, preferably GANs, and being familiar with the current state of the art in the area.
  • Preference will be given to candidates with prior experience in several of the following skills: C++/Python, pathtracing/rendering, OpenGL, OpenCL/CUDA, Matlab, OpenCV, DSLR/machine-vision cameras.
  • Not strictly required, but a definite plus: publications at top tier conferences (e.g., CVPR, SIGGRAPH)

Soft Skills Requirements

  • Fluent spoken and written English, minimum level B2, C1 definitely preferred.
  • Self-reliant working style.
  • Clear communication.
  • Curious and ambitious personality.
  • Willing to take up the challenge of building large and robust systems.

How to apply!

Qualified applicants are invited to submit the following information to tim.weyrich@fau.de:

  • A single Page CV
  • A scan of one’s Master’s diploma, along with a transcript of records for the master studies
  • A motivation letter why one wants to join PRIME, along with a justification why one thinks one fits the research topic.
  • Letters of recommendation from two previous employers and/or teachers at the university the applicant graduated from (ideally, this includes the supervisor of the master thesis).

Please collate all this information into a single PDF file, and provide it to us either as mail attachment at the above address, or via a file sharing service such as Google Drive.