Klaus Maier-Hein
Klaus Maier-Hein works on deep learning for medical image analysis and the software needed to use it in research. He earned a computer science doctorate at Heidelberg in 2010, followed by research at DKFZ and Harvard Medical School. He coauthored nnU-Net, a framework that adapts standard U-Net models to new segmentation datasets. He also leads work on medical image computing, scalable analysis and open research software. His work addresses both model design and the conditions needed to translate methods into clinical research.
- Works
- 519
- Citations
- 32,703
- h-index
- 70
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2010
Doctoral training and postdoctoral research
He studied computer science at Karlsruhe and EPFL, earned a 2010 Heidelberg doctorate and completed postdoctoral work at DKFZ and Harvard Medical School. His later deep learning research focuses on medical image analysis.
-
2018
Make segmentation adapt to the data
Maier-Hein coauthored nnU-Net with Fabian Isensee and other collaborators. The framework adapts preprocessing, training and inference around two- and three-dimensional U-Nets. It was tested in the Medical Segmentation Decathlon across different image types and tasks.
-
As of October 2026
Link images with clinical information
The DKFZ division analyzes images from different technologies and time points, linking them with clinical and biological parameters. Its methods include segmentation, object detection, probabilistic modeling and learning useful representations.
-
As of October 2026
Support open and scalable research
He coordinates the Helmholtz Imaging Platform and works on research software infrastructure. The division emphasizes open-source tools, uncertainty and scalable analysis across clinical research networks, identifying practical conditions for using image-analysis methods.
Sources and credits
Photo: www.klinikum.uni-heidelberg.de; creator not stated. License: Not stated by source. Source: klinikum.uni-heidelberg.de. Cropped to a square and resized. Ask us to remove or replace this photo.