Andreas Bär
PhD Student @
Technische Universität Braunschweig
Biography
I am a final-year PhD student in Electrical Engineering under supervision of Tim Fingscheidt, Professor at the Institute for Communications Technology, Technische Universität Braunschweig. My PhD thesis lies in the area of computer vision with focus on:
- Semantic Segmentation
- Out-of-Distribution Robustness, Adversarial Attacks
- Teacher-Student Learning
- Performance Prediction
In addition, I recently developed interest in:
- (Multimodal) large language models
- Vision foundation models
- Efficient transfer learning
- Data augmentation on frozen features
- Training at scale
Education
-
MSc in Electrical Engineering, 2018
Technische Universität Braunschweig, Braunschweig, Germany
-
BEng in Electrical Engineering, 2016
Ostfalia University of Applied Sciences, Wolfenbüttel, Germany
Awards
Best Paper Award 🏆
Serin Varghese, Sharat Gujamagadi, Marvin Klingner, Nikhil Kapoor, Andreas Bär, Jonas Schneider, Kira Maag, Peter Schlicht, Fabian Hüger, and Tim Fingscheidt. An Unsupervised Temporal Consistency (TC) Loss to Improve the Performance of Semantic Segmentation Networks (2021). In Proc. of CVPR - Workshops.
Best Paper Award 🏆
Andreas Bär, Marvin Klingner, Serin Varghese, Fabian Hüger, Peter Schlicht, and Tim Fingscheidt. Robust Semantic Segmentation by Redundant Networks With a Layer-Specific Loss Contribution and Majority Vote (2020). In Proc. of CVPR - Workshops.
Experience
Research Associate
- Internal project lead in research projects, e.g., KI Absicherung
- Write and share research code in Python, e.g., TUBSRobustCheck, PerfPredRec, …
- Write and publish scientific papers (Google scholar)
- Supervise student bachelor and master theses as well as student research assistants
- Peer review for conferences and journals, e.g., IEEE/CVF CVPR, ICRA, IROS, IJCV, IEEE RA-L, Neurocomputing, Pattern Recognition, …
Research Intern
- Research in computer vision under the guidance of my host Manoj Kumar and co-host Neil Houlsby.
- Implemented feature caching and corresponding training pipelines for several pretrained Vision Transformer models.
- Investigated transfer learning capabilities and the effect of data augmentation methods in the feature space.
- Contributed to Scenic core.
- The results are summarized in a paper which was presented at CVPR 2024!
Software Quality Assurance Engineer
- Work in a team of (software) engineers with 10+ years of experience
- Assist software quality assessments of OEM suppliers according to ISO 26262
Projects
Publications
(2024).
Frozen Feature Augmentation for Few-Shot Image Classification.
In Proc. of CVPR.
(2023).
A Novel Benchmark for Refinement of Noisy Localization Labels in Autolabeled Datasets for Object Detection.
In Proc. of CVPR - Workshops.
(2023).
Improvements to Image Reconstruction-Based Performance Prediction for Semantic Segmentation in Highly Automated Driving.
In Proc. of CVPR - Workshops.
(2022).
Detecting Adversarial Perturbations in Multi-Task Perception.
In Proc. of IROS.
(2022).
Adaptive Bitrate Quantization Scheme Without Codebook for Learned Image Compression.
In Proc. of CVPR - Workshops.
(2022).
Performance Prediction for Semantic Segmentation and by a Self-Supervised Image Reconstruction Decoder.
In Proc. of CVPR - Workshops.
(2022).
Improving Transferability of Generated Universal Adversarial Perturbations for Image Classification and Segmentation.
In Deep Neural Networks and Data for Automated Driving.
(2022).
Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety.
In Deep Neural Networks and Data for Automated Driving.
(2022).
Joint Optimization for DNN Model Compression and Corruption Robustness.
In Deep Neural Networks and Data for Automated Driving.
(2021).
From a Fourier-Domain Perspective on Adversarial Examples to a Wiener Filter Defense for Semantic Segmentation.
In Proc. of IJCNN.
(2021).
Detection of Collective Anomalies in Images for Automated Driving Using an Earth Mover’s Deviation (EMDEV) Measure.
In Proc. of IV - Workshops.
(2021).
An Unsupervised Temporal Consistency (TC) Loss To Improve the Performance of Semantic Segmentation Networks.
In Proc. of CVPR - Workshops.
(2021).
Improving Online Performance Prediction for Semantic Segmentation.
In Proc. of CVPR - Workshops.
(2021).
The Vulnerability of Semantic Segmentation Networks to Adversarial Attacks in Autonomous Driving: Enhancing Extensive Environment Sensing.
In IEEE SPM.
(2020).
Transferable Universal Adversarial Perturbations Using Generative Models.
In arXiv.
(2020).
Focussing Learned Image Compression to Semantic Classes for V2X Applications.
In Proc. of IV.
(2020).
Class-Incremental Learning for Semantic Segmentation Re-Using Neither Old Data Nor Old Labels.
In Proc. of ITSC.
(2020).
Improved Noise and Attack Robustness for Semantic Segmentation by Using Multi-Task Training with Self-Supervised Depth Estimation.
In Proc. of CVPR - Workshops.
(2020).
Robust Semantic Segmentation by Redundant Networks With a Layer-Specific Loss Contribution and Majority Vote.
In Proc. of CVPR - Workshops.
(2020).
Unsupervised Temporal Consistency Metric for Video Segmentation in Highly-Automated Driving.
In Proc. of CVPR - Workshops.
(2019).
On the Robustness of Redundant Teacher-Student Frameworks for Semantic Segmentation.
In Proc. of CVPR - Workshops.
(2019).
On Low-Bitrate Image Compression for Distributed Automotive Perception: Higher Peak SNR Does Not Mean Better Semantic Segmentation.
In Proc. of IV.
(2019).
Towards Corner Case Detection for Autonomous Driving.
In Proc. of IV.