Thomas Brox

Thomas Brox is a Professor for Pattern Recognition and Image Processing at the University of Freiburg, where he has headed the Computer Vision Group since 2010. He holds a PhD in computer science from Saarland University (2005) and has held postdoctoral positions at UC Berkeley, TU Dresden, and University of Bonn.

• Research Interests
• Education
• Publications & Awards
• Teaching & Leadership Roles

His research focuses on computer vision and deep learning, particularly visual representation learning, video analysis, and training deep networks without manual supervision. He investigates how deep networks learn capabilities and applies these insights to action-perception cycles in robotics.

Recent publications highlight trends in vision-language models, diffusion-based 3D generation, and anomaly detection. His work includes U-Net (seminal biomedical image segmentation) and FlowNet (optical flow estimation).

• Koenderink Prize (2014)
• ERC Starting Grant (2011)
• Longuet-Higgins Best Paper Award (2004)
• Multiple GCPR/ICCV/CVPR Best Paper Honors

He teaches courses in Optimization, Statistical Pattern Recognition, and Computer Vision. He has served as Dean of Studies, Director of Computer Science Department, and will become Dean in 2026. His lab mentors 25+ PhD/MSc students and collaborates with institutions like Amazon (2020-2024).