M.Sc. Richard Fehler
- Wissenschaftlicher Mitarbeiter am FZI
- Tel.: +49 721 9654-261
- fehlerBrw4∂fzi de
FZI Forschungszentrum Informatik
Haid-und-Neu-Str. 10–14
76131 Karlsruhe, Germany
Zu vergebende Abschlussarbeiten
- Algorithmen + Deep Learning: Traffic Light to Lane Assignment für die Offline Anwendung in der Autolabeling Cloud (deutsch/english)
- Deep Learning & Software Engineering: Classification or Detection of Polylines with Aerial Images as CNN Input (deutsch/english)
- Applied Deep Learning: Training and Evaluation of a Large Panoptic Segmentation Neural Network Model on the Mapillary Vistas 2 Dataset (deutsch/english)
- Deep Learning and Mapping: Parametric Mapping on Argoverse2 Dataset with Deep Neural Networks and 3D Mapping (deutsch/english)
- Deep Learning: Consistent BEV Map Perception with Spatiotemporal NeuralMemoryBuffer for GlobalMap Perception (deutsch/english)
- Deep Learning: Long Range BEV Map Perception with: Virtual Telephoto Camera / Lidar Sensor Fusion / SD Map Priors (deutsch/english)
- Deep Learning: Shared ResNet50 Backbone for Mask2Former and MapTRv2 (like Tesla HydraNet) (deutsch/english)
- Applied Deep Learning & Software Engineering: Map Perception Extended to Bicycle Lanes (deutsch/english)
- Software Engineering: Optimal Graph Splitting and Merging of OpenStreetMap Road Networks for Highly ParallelHD-MappingforDeepLearning in Autonomous Driving (deutsch/english)