Carsten Rother
Carsten Rother is a computer-vision researcher whose work includes image segmentation, restoration, scene understanding and machine learning. He earned a diploma at Karlsruhe in 1999 and a doctorate at KTH in 2003. He worked at Microsoft Research Cambridge from 2003 to 2013, at TU Dresden from 2014 to 2017, and joined Heidelberg in September 2017. He coauthored GrabCut with Vladimir Kolmogorov and Andrew Blake in 2004, as well as TextonBoost, AutoCollage and later work on camera localization. His research spans deep learning, graphical models and methods that generate useful training data. These results are shared contributions.
- Now at
- Heidelberg University
- Works
- 343
- Citations
- 34,791
- h-index
- 83
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
1999–2003
Training in computer vision
Rother received his diploma at Karlsruhe in 1999 and his doctorate at KTH in 2003. Jan-Olof Eklundh and Stefan Carlsson supervised his doctorate.
-
2003–2013
Microsoft Research and GrabCut
He worked in Microsoft Research Cambridge’s computer-vision group from 2003 to 2013. He coauthored the 2004 GrabCut paper with Vladimir Kolmogorov and Andrew Blake, a shared method for interactive foreground extraction using graph cuts.
-
2014–2017
Dresden and Heidelberg
He was a full professor at TU Dresden from 2014 to 2017, then joined Heidelberg in September 2017. He also coauthored 2017 work on differentiable RANSAC for camera localization.
-
As of October 2026
Visual learning across applications
His research includes deep learning, graphical models and training-data generation. His work applies these methods to image editing, image matching, object-pose estimation and cell tracking.
Sources and credits
Photo: ellis.eu; Carsten Rother profile photograph; creator not stated. License: Not stated by source. Source: ellis.eu. Cropped to a square and resized. Ask us to remove or replace this photo.