M.Sc. projects in the Pattern Recognition Laboratory focus on the fundamental pattern recognition questions. How can we learn from different types of data (measurements like video, images or sounds, but also more involved biomedical measurements, spectra, time series data)? How can we learn from poorly labeled data or poorly sampled data or how can we learn from objects for which we cannot invent good features? How do our pattern recognition tools perform on different types of classification problems, how can they be adapted to a new classification problem, in what situations do the tools fail?
We are not interested in solving one specific problem as such. Often the practical application question is the starting point of the real research, in which we try to answer a more fundamental questions (as mentioned above). It would be nice if at the end of the research a practical algorithm or method is developed that can solve the original application problem reasonably well. But if not, the method or algorithm should work well on another type of data. So it is very important to show the conditions under which an algorithm works well.
When you would like to do a graduation in the field of pattern recognition, please drop by on the 13th floor of the EWI building to have a chat with us (or make an appointment first). It might sound odd, but it is important that you like your graduation project and it therefore is certainly permitted, encouraged even, that you come with your own research question or problem. On the page with available projects some potential projects are mentioned, but feel free to invent your own, using the project page for inspiration. Also take a look at the finished projects if you want to get an idea of what projects other students have been doing. In further discussions, together we can come to a more specific scientific question and a better defined M.Sc. project.