Bachelor and Master theses
Medical Image Processing
| Automatische Detektion des Ultraschalldatenbereichs in Screenshots verschiedener Ultraschallsysteme | C++ Kenntnisse, Grundlagenwissen in der Bildverarbeitung | Contact: Dr. Susanne Winter |
| MA/BA/Studienprojekte im Bereich der medizinischen Bildverarbeitung | Registrierung/Segmentierung/Klassifikation/Kalibrierung/medizinische Navigation/medizinische Modalitäten: CT, MRT, Ultraschall | Contact: Dr. Susanne Winter |
Adaptive Systems
| Bachelorarbeiten im Bereich Reinforcement Learning | Voraussetzungen: mindestends 70 Punkte in der Klausur "Reinforcement Learning" sowie gute C++ Kenntnisse und/oder mathematische Fähigkeiten | Contact: Jun.-Prof. Dr. Tobias Glasmachers |
| Master- und Diplomarbeiten im Bereich Machinelles Lernen (Reinforcement Learning, Support Vector Machines, Neuronale Netze, ...) | Voraussetzungen: solide mathematische Fähigkeiten | Contact: Jun.-Prof. Dr. Tobias Glasmachers |
Real-time Computer Vision
| Verschiedene BA- / MA-Arbeiten im Bereich Fahrerassistenzsysteme | Objekterkennung / Systemoptimierung / Kopplung von Modulen / Scene Flow / Szenenrepräsentation und -prädiktion / etc. Anforderungen: Programmierkenntnisse |
Contact: Jan Salmen |
| Computergestützte Sportanalyse: Data-Mining für Tracking-Daten im Fußball | Segmentierung / Bewegungsdetektion / Klassifikation / zeitliche Integration Anforderungen: Programmierkenntnisse |
Contact: Marc Schlipsing |
| Besondere Themen der Fahrerassistenz für Motorräder | Schätzung des Neigungswinkels / Objekterkennung Anforderungen: Programmierkenntnisse |
Contact: Marc Schlipsing |
| Fahererassistenz mit Fischaugenkameras | Objekterkennung / Kamerageometrie Anforderungen: Programmierkenntnisse |
Contact: Sebastian Houben |
| Implementierung von Bildverarbeitungs- und weiteren Verfahren auf GPU / Parallelisierung | Anforderungen: Programmierkenntnisse | Contact: Marc Schlipsing |
Theory of Neural Systems
Implementing Predictable Feature Analysis as an MDP-node |
Requirements: Basic programming skills and basic knowledge of Python or willingness to learn and program in it. Basic knowledge in mathematics, especially linear algebra (i.e. working with matrices, eigenvalues and eigenvectors, etc). Lacking the requirements does not matter, if the willingness to improve lacking skills is present. | Contact: Stefan Richthofer |
Solving multivariate polynomial systems  |
Requirements: Basic mathematical background in numerics and algebra and interest in computational and geometric algebra. General programming skills, ability to understand and adopt APIs from possibly unfamiliar programming languages. Lacking any of these requirements does not matter, if the willingness to improve it is present. | Contact: Stefan Richthofer |
Path integration based on the principles of slowness |
Requirements: A large part of this project consists of working with the robotic agent, which means you should be comfortable with extended periods of programming in C++ (you may also need to learn some basic Python as well). For the path integration part, a solid understanding of basic math will be required (an Eigenvector should not scare you away :-) or the willingness to learn missing skills that may be needed. You should be able to work independently and generally be enthusiastic about extending the horizon of your current studies. | Contact: Fabian Schönfeld |
| Slow Feature Analysis and the real world |
Requirements: As you will be working primarily with your chosen robot, this will mainly be a programming task. As such, you should be comfortable with extended periods of working with C++ (you may also need to learn some basic Python as well). Some general interest in math would be a bonus, but is not a requirement. You should be able to work independently and generally be enthusiastic about acquiring new skills you’re not familiar with yet. Ideally your work will be integrated into a larger project, so a fundamental requirement is the ability to write clean and well documented(!) code that enables others to pick up the project where you left off. | Contact: Fabian Schönfeld |
Development of an Improved Algorithm for Gender Estimation from Frontal Face Photographs |
Requirements: Intermediate programming skills and basic knowledge of Python or willingness to learn it are needed. Basic notions of machine learning or neural networks would be an advantage. |
Contact: Alberto Escalante |
Biologically plausible alternatives for back-propagation |
Requirements: Programming skills are obligatory. The knowledge of Python is of advance, but can be learned during the project as well. Biological knowledge is not necessary. The project can be done in German, too. | Contact: Torsten Neher |
| Biologically plausible attractor neural networks |
Requirements: Programming skills are obligatory. The knowledge of Python is of advance, but can be learned during the project as well. Biological knowledge is not necessary. The project can be done in German, too. | Contact: Torsten Neher |
Organic Computing
| BA/MA/Diplomarbeiten | Wir haben eine Reihe von Themen in den Bereichen Computersehen, Bildverstehen, Maschinelles Lernen und Robotik. Ein besonderer Schwerpunkt liegt auf Verfahren zum automatischen Lernen von Bildverarbeitungsroutinen aus Beispielen. Die konkreten Projekte werden den Vorkenntnissen und Interessen der Kandidaten angepasst. Anforderungen: Studium der Physik, Informatik, Elektrotechnik oder anderen relevanten Fächern; Interesse an Problemen des Computersehens oder der Mensch-Maschine-Interaktion. |
Contact: Dr. Rolf Würtz |