Perception is a fundamental part of intelligence since perception is necessary to acquire knowledge and knowledge is necessary to understand perception. Therefore computer vision is one of the most important aspects in the realization of intelligent systems. My interest of research lies in computer vision and the combination with machine learning which, to my mind, will enable the realization of intelligent systems. Currently, I am working on optical flow and how to incorporate high-level information to alleviate this ill-posed problem.
Human brains can learn from data without an obvious supervision target and can afterwards use this knowledge in a variety of situations and even generate new content. On the other hand, the applicability of modern supervised machine learning algorithms is limited by the amount of available training data. Generative probabilistic models seem to provide a promising way to tackle this problem and in addition allow us to incorporate knowledge of the underlying process. My long term goal is to use such models to tackle ambiguities in 3D reconstruction due to non-Lambertian materials and to reason about global illumination.
Since November I am a PhD candidate under the supervision of Prof. Andreas Geiger. Currently I am working on deep generative models in collaboration with the ML team of the ETAS GmbH, a subsidiary of Bosch.
One of the most challenging tasks of Computer Vision is to endow computers with the ability to discover the underlying relationships between the objects in a scene. The large amount of available labeled data as well as the fast progress in deep learning has significantly advanced many Computer Vision tasks, such as object segmentation. optical flow estimation, action recognition etc. However, a truly intelligent system would ideally be able to infer high-level semantics underlying human actions such as motivation, intent and emotion. However, all human actions involve some uncertainty. To this end, I would like to either develop or to further enhance existing methodologies that incorporate such uncertainties. For now, I have worked on the 3D reconstruction task, by developing a model able to incorporate uncertainties in the image formation process..