Carsten Rother
Microsoft Research (United Kingdom), TU Dresden, Microsoft (United States)
Papers
8
Total Citations
558
H-Index
6
About
Carsten Rother is a leading figure in computer vision, renowned for his pioneering work in interactive image segmentation and 3D reconstruction. His research fundamentally bridges the gap between automatic algorithms and human guidance, enabling precise, user-driven visual analysis. Rother’s most impactful contribution is the introduction of "Geodesic Star Convexity" for interactive segmentation (2010, 412 citations). This work revolutionized the field by extending simple shape constraints to complex, multi-star configurations, allowing users to intuitively guide algorithms to produce globally optimal segmentations—a critical advance for applications from medical imaging to photo editing. He further advanced user-centric vision by developing systems that learn from human interaction, as detailed in his works on learning interactive segmentation systems (2010, 24 citations; 2012, 26 citations). Beyond segmentation, Rother has made significant strides in 3D scene understanding, including multi-view reconstruction using reference planes (2003, 33 citations) and pose estimation for kinematic chains (2015, 38 citations), which are vital for robotics and augmented reality. His research consistently emphasizes practical, user-in-the-loop solutions, earning him a reputation for creating algorithms that are both theoretically elegant and deployable in real-world scenarios.
Research Focus
Key Achievements
Top Papers
- 1Geodesic star convexity for interactive image segmentation412 citations · 2010
- 2
- 3
- 4User-Centric Learning and Evaluation of Interactive Segmentation Systems26 citations · 2012
- 5Learning an interactive segmentation system24 citations · 2010
- 6
- 7
- 8Learning an Interactive Segmentation System3 citations · 2009