[ VIGRA Homepage | Publications | German Vision Groups | Cognitive Systems Group ]

Dr. rer. nat. Ullrich Köthe


Welcome to my homepage! I'm an assistant professor in the Cognitive Systems Group at the Computer Science Department of the University of Hamburg, Germany. Previously (until 1998) I had worked at the Fraunhofer-Institute for Computer Graphics in Rostock, where I did most of the work for my PhD thesis. My research interests include Computer Vision (especially image segmentation and feature extraction) and Reusable Software (reusable algorithms and generic programming) in the context of Computer Vision. Take a look at my publications. I'm also maintaining the semi-official list of German computer vision research groups and companies for the DAGM (German Association for Pattern Recognition).

Image Segmentation: Segmentation is the partitioning of an image into meaningful regions which correspond to (parts of) the objects to be seen. The better your initial segmentation the greater your chance to extract correct semantic information from the image. Among other topics, I've written papers on how to integrate edge and junction detection in order to get correct boundary topology, how to obtain and represent topological partitionings optimally in a computer, and on how sampling affects the correct representation of geometric shapes.
Reusable Software for Computer Vision: Software reuse has been a goal for a long time, but does not work as well as it should. I've written some design patterns which catalog some of the basic ideas of reusable architectures. At the level of algorithms, these ideas are best realized by generic programming (the programming style of the C++ Standard Library). Therefore, I developed a novel computer vision library VIGRA which is based on generic programming. A number of papers and, most comprehensively, my PhD thesis describe the ideas of this library and report our experiences with reusable algorithms.
The VIGRA Library: I'm the author of the free computer vision library VIGRA, which applies the above techniques to implement extremely flexible, yet efficient image processing algorithms. You can read the online documentation or download VIGRA 1.4.0.
3D Reconstruction: While I was at Fraunhofer-Institute for Computer Graphics in Rostock, I helped developing a system which allows to extract 3-dimensional object models from image sequences. I.e., to develop a virtual reality simulation containing natural objects (such as buildings), one takes some photos of each object and uses our system to generate realistic geometric models (including textures). Here are the official descriptions of the projects REKO and ERSO. A description of the ERSO project and the underlying concepts can be found in this paper.
Interesting Links: Cognitive Systems Group
Computer Science Department
University of Hamburg
German computer vision research groups and companies

PhD thesis: Generische Programmierung für die Bildverarbeitung
The VIGRA Image Processing Library

Contact: Fachbereich Informatik, Arbeitsbereich KOGS
Universität Hamburg
Vogt-Kölln-Straße 30
22527 Hamburg, Germany
koethe@informatik.uni-hamburg.de and u.koethe@computer.org
+49 (0)40 42883-2573
+49 (0)40 42883-2572