Rupert Polley

Aus Aifbportal
Wechseln zu:Navigation, Suche
2021-08-18 FO Polley-Rupert 003 Outlook.jpg
  • M.Sc. Rupert Polley

  • FZI-Mitarbeiter
  • Tel.:+49 721 9654-390
  • Email: polley∂fzi de

Hello, I am Rupert and I studied computer science at the KIT until early 2021 with a focus on machine learning. Since June 2021, I am a research scientist at the FZI Research Center for Information Technology in the Technical Cognitive Systems (TKS) department and as a PhD student at the partner department Applied Technical Cognitive Systems (ATKS) at the AIFB.

For my PhD theses I am researching the creation of HD-Maps from aerial imagery. HD-Maps in comparison with commonly known maps are much more precise and contain information needed for autonomous vehicles. Typically they are created manually and I am looking into automating this process with deep neural networks. Depending on the resolution, aerial images contain helpful information like road surface markings, bicycle paths, sidewalks etc. The segmented result can then be turned into a HD-Map and used for autonomous driving. Finally the hd-map needs to be updated it the scene changes. This is done when an autonomous vehicle detects a change between its current view and the provided map.

Open Bachelor/Master Theses and HiWi Positions

There is ALWAYS something to do!

So if you are interested about image segmentation of aerial images with neural networks, the generation of HD-maps from a segmented images or the verification of a HD-map with a autonomous vehicle send me a short mail and we can discuss what you would like to do and which task would fit you best. I always have tasks for both HiWis and Bachelor/Master theses. Just send a mail and we will find something just for you :)

  • Thesis/HiWi: Segmentation and Road Extraction of Aerial Images with Deep Learning
  • Thesis/HiWi: Improving Performance with Self-Supervised Learning when little data is available
  • Thesis/HiWi: Road Extraction with Diffusion Models
  • Thesis/HiWi: Validation of generated Map Material

Maschinelles Lernen, Deep Learning, Künstliche Intelligenz