Research interests

My primary research interest revolves around geometric models for the analysis of volumetric data. This includes volumetric segmentation (an example is a geometry-based segmentation of human peripheral nerves using this tool), tomographic segmentation, and methods based on deformable meshes. I work on applications in material science, energy materials, medicine, diagnostics, and food quality control.

Another interest involves 2D and 3D texture analysis — see examples demonstrating a probabilistic framework for curve evolution. I also work on 3D methods for digitizing cultural heritage. Take a look at a prize-winning motivational video prepared by a PhD student Dolores Messer, and optical scans of some skulls. Finally, I work with analyzing OCT and OCTA images for ophthalmology, see for example our video from ISBI 2020.

Some of my work is on display in my poster gallery. The segmentation tool InSegt is explained in this CVMI@CVPR video. Check Patrick's and Niels' videos from CVPR 2020, too. ICCV 2021 is also virtual, so another video by Niels. Our extensive work on min-cut/max-flow algorithms, collected in a PAMI paper, is summarized here.

I initiated the work on QIM data repository, and now we're open.

A few recent highlights and plans. In project STUDIOS, we'll develop easy-to-use, easy-to-train, simple and efficient algorithms for segmenting X-ray tomography data. In RENNER, we'll develop an ultra-fast system for measuring the microstructure of composite materials.


For a complete and updated list of publications see my Orbit page or my Google Scholar profile.