My current research focuses on 3D modeling and reconstruction. In particular I'm interested in modeling low-textured objects/environments by incorporating (possibly learned) prior knowledge.
In order due to so I'm looking at both, reconstruction algorithms facilitating information from learned visual subspaces, and possible learning strategies itself.
Besides from that I'm very much interested in the combination of structural and sematic information; i.e. algorithms improving visual reconstruction by incorporating semantic labelings, or enabling scene under-standing by the fusion of texture and 3D data.
My broader research interest also includes object detection and recognition, and machine learning applied to vision problems.
For current and previous work please see my publication site.
Student theses
I'm always looking for motivated students, interested in pursuing a Bachelor, Master or semester thesis. Ideally you would already have some knowledge in computer vision and are familiar with coding in C++ and Matlab.
I'm happy to tell you about current options - just contact me by mail.
