University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Summer School 2009

Photo by Böhringer

Our DFG-sponsored PhD program at the University of Konstanz conducts a weeklong

The summerschool contains tutorials and workshops on and poster/oral presentations of all participating students. The general language is English. About 25 of our doctoral students and some advisors will participate.
This year the summerschool is also open for interested female computer science students from all other universities, see below.

View some images



special opportunity only for female advanced students in computer science

This year our funding agency, the DFG, supports our program with funding specifically for advanced female master or diploma students in computer science, in order to encourage young women to consider a career in engineering by entering a PhD program like ours. Our funds will cover for about 20 such participants.

Therefore we invite the best and interested female students, that are advanced or have a completed CS master degree of 2009, to apply for participation at our summerschool this year. If you like to participate, please send us your application as a PDF to gksekr[at] covering the following topics: Dates:
The following students took advantage of this opportunity:

from left to right: Denise Hippler, Marina Litvinova, Anna Zubenko, Dina R. Khattab, Joselene Marques,
Emitza Guzman, Stefanie Mayer, Katarina Kanevceva, Gissel Velarde, and in the middle Prof. Dr. Deussen
not on the picture: Angelika Garz, Britta Weber

flyer "Invitation to summer school"

Flyer containg an overview of the informations as PDF.


tutorial descriptions

Kernel Methods for Classification: From Theory to Practice

by Prof. Dr. Michael Berthold, Prof. Dr. Ulrik Brandes, Martin Mader, Uwe Nagel

We will introduce the well-known kernel methods for classification tasks. With this method linear classification methods are extended to separate input data with non-linear separators based on nearly arbitrary object similarities. In a first part we give a short introduction to kernel theory and its use in classification problem. This theoretical knowledge will then be consolidated by direct application in the second part. In a hands-on session we will experiment with different types of kernels for a number of classification problems (in a state of the art data-mining environment). Interested attendees will have the opportunity to develop their own kernel(s) in a competition on a given classification task.

Winners of the Competition

from left to right: Andreas Stoffel, Denise Hippler, Roman Byshko

3D Scan and Image Processing Techniques

by Prof. Dr. Robert Sablatnig, Dr. Martin Kampel

This tutorial shows how image processing algorithms are used for practical, industrial applications. Basic algorithms and techniques are discussed and some of the major tasks of are shown. Image processing operations can be roughly divided into three major categories, Image Compression, Image Enhancement and Restoration, and Measurement Extraction.

The second major topic of this tutorial is called 3d machine vision. 3D scanners and image acquisition systems are rapidly becoming more affordable and allow to build highly accurate models of real 3D objects in a cost- and time-effective manner. This tutorial will present the potential of this technology, review the state of the art in model acquisition methods, and will discuss the 3D acquisition pipeline from physical acquisition until the final digital model. First, different optical scanning techniques (e.g. structured light triangulation, time-of-flight approaches) will briefly be presented. In the area of registration, we will consider both the problems of initially aligning individual scans, and of refining this alignment with variations of the Iterative Closest Point method. We will then look at various ways in which surface properties such as color and reflectance can be extracted from acquired imagery. Finally, we will examine techniques for the efficient management and rendering of very large, attribute-rich meshes, algorithms to represent the 3d-world in the computer are presented.

Furthermore for every image processing and 3d-acquisition strategy an industrial application is shown. The tutorial covers the development from the automatisation plan to the concrete system, from image acquisition to control structures and performance evaluation.

postersession by invited students

Texture based document layout analysis of glagolitic manuscripts
Angelika Garz, TU Vienna, Austria
Virtual News Presenter
Denise Hippler, State University of Campinas, Brazil, and WSI/GRIS University of Tübingen, Germany
High-Resolution mapping of the structural core of human cerebral cortex
Katarina Kanevceva, University Ss. Cyril and Methodius, Macedonia
Geometry compression of polygon mesh models
Dina R. Khattab, Ain Shams University, Egypt
Integrated network systems in a town
Marina Litvinova, Kharkov National University of Radioelectronic, Ukraine
Dimensionality reduction and relevance feedback: Powerful techniques on CBIR systems
Joselene Marques, University of Sao Paulo, Brazil
Anaglyph stereo for the realtime synthesis of images
Stefanie Mayer, Eberhard-Karls-Universität Tübingen, Germany
Enabling physical reasoning through 3D simulation
Emitza Guzman Ortega, TU Munich, Germany
Segmentation of microtubules from electron tomograms
Britta Weber, Free University Berlin, Germany
Database parameter optimization for welding machines
Gissel Velarde, Electrical R&D Department, Miebach GmbH
Application of tabu search to solving problems of non-linear programming
Anna Zubenko, Kyiv School of Economics, Ukraine

best poster awards

Denise Hippler, Britta Weber, Joselene Marques


Assessment of observation importance in large scale estimation problem
by Vladimir Bondarenko
Mathematical models for race-bike time trials
by Thorsten Dahmen
Blended library
by Mathias Heilig
Multi display environment
by Mahsa Jenabi
PDScan - "Photo Data Scan" density-based clustering algorithm and its applications
by Slava Kisilévich
Dynamic visualization of evolving document collections
by Miloš Krstajić
Frequency-based stippling
by Sören Pirk
Content based document structure analysis
by Andreas Stoffel
Document cards: A top trump visualization for documents
by Hendrik Strobelt
Quasi-semantic properties and their usage in document analysis
by Daniela Oelke
Connecting insect brain imaging to data mining tools - a neuroimage ectension for the KNIME platform
by Martin Strauch
Visualization of a tag cloud
by Iris Adä
Sparse color salient points for object retrieval and categorization
by Julian Stöttinger
Integrating data mining & data warehousing/databases
by Nafees Rehman

best presentation

Daniela Oelke