Lennart Evers
Wissenschaftlicher Mitarbeiter der
AG Optimierung und Optimale Steuerung,
Research Training Group π3
Research
My primary research interest is the computer vision based perception of autonomous vehicles. This includes real time solutions for semantic image segmentation and object detection based on deep convolutional neural networks, as well as their interaction during training. For that, I develop multi objective neural networks solving both tasks simultaneously, investigate convergence behavior depending on the training method used and evaluate the models on a real car in the traffic of Bremen.
Segmentation in urban driving scenarios
Object detection
Teaching (selection)
- Übung Analysis 3 (WiSe '18/19)
- Übung Maß- und Wahrscheinlichkeitstheorie (WiSe '17/18)
- Übung Analysis 3 (WiSe '16/17)
Lebenslauf
- 2019 M.Sc. Mathematics at University of Bremen
- Specialized in numerics, artificial intelligence and computer vision.
- Thesis: Applying deep learning to real time,
simultaneous detection and
recognition of traffic signs
- 2017 B.Sc. Mathematics at University of Bremen
- Thesis: Planung von Einparkmanövern mit Methoden der optimalen Steuerung
Projekte
-
OPA³L - Optimal Assistierte, hoch Automatisierte, Autonome und kooperative Fahrzeugnavigation und Lokalisation
(01.03.2019 - 14.02.2023)
- L. Evers.
Benchmarking pre-trained Encoders for real-time Semantic Road Scene Segmentation.
GAMM 92st Annual Meeting of the international Association of Applied Mathematics and Mechanics, online, 15.03.2021 - 19.03.2021.
Proceedings in Applied Mathematics & Mechanics, 20(1), WILEY-VCH, 2020. DOI: 10.1002/pamm.202000275
- A. Folkers, C. Wellhausen, M. Rick, X. Li, L. Evers, V. Schwarting, J. Clemens, P. Dittmann, M. Shubbak, T. Bustert, G. Zachmann, K. Schill, C. Büskens.
The OPA3L System and Testconcept for Urban Autonomous Driving.
25th IEEE International Conference on Intelligent Transportation Systems, 08.10.-12.10.2022.
Segmentation in urban driving scenarios
Object detection