Denoising Autoencoders for Time-Of-Flight Cameras
Advanced driver assistance and automated driving systems present new perspectives for the development of future vehicles. A new type of imaging sensor – the Time-of-Flight (ToF) Camera – offers new possibilities to monitor the surroundings of a vehicle by providing pixel level depth information in the image. The ToF Camera thus significantly reduces the computational effort to generate 3D environment maps.
The Virtual Vehicle Research Center (ViF), supported by the EU project IoSense and in collaboration with European partners, is currently equipping a demonstrator vehicle with ToF cameras. Initially, for the evaluation of the depth image processing for the ToF camera, common of-the-shelf computing hardware will be used. Thereafter, the functionality will be ported to dedicated embedded hardware components.
Your duties and responsibilities:
The task of the master thesis is the analysis of denoising autoencoders for depth images with machine learning methods to enable high quality and high frame rate depth image processing for automatic parking applications.
What we expect from you:
- Linear Algebra, Numerical Mathematics, Probability Theory
- Programming skills: Python and C/C++ on Linux or Windows
- Machine learning frameworks: Tensorflow, Keras
- Ongoing Master’s programme in Mathematics, Technical Mathematics
What we offer:
For technical questions contact:
Univ.-Ass. Mag.rer.nat. Dr.rer.nat. Manfred Liebmann
University of Graz / Institute for Mathematics and Scientific Computing
Dr.rer.nat. Jelena Rubesa-Zrim
Kompetenzzentrum – Das virtuelle Fahrzeug Forschungsgesellschaft mbH.
If you are interested, please contact:
Tel .: + 43- (0) 316-873-9028
We look forward to your details; Please use our online contact form to send your application to a VIRTUAL VEHICLE.Apply