Mathematical and Applied Visual Computing (MAVC)

The Mathematical and Applied Visual Computing group is led by Prof. Arjan Kuijper.

Visual Computing is image- and model-based information technology and includes computer graphics, computer vision, as well as virtual and augmented reality. It has a theoretical and an applied side: The challenge is to design good methods, that combine correctness and usefulness.

Mathematical Visual Computing deals with the underlying algorithms and models. One can, for instance, improve images using evolution equations (PDEs) that optimize a user-defined energy, but do these PDEs really do what we want? They can also be used for segmenting images - of course, the challenge is to do it faster and better than humans! A convenient way to describe shapes is by means of their skeleton, exploiting local symmetry. Interesting research questions here are which changes in the skeleton are caused by specific changes in the object, and if we can compare shapes by “simply” comparing the skeleton structures. Visualizing relevant data in a for a user pleasant way is far from trivial, especially when the amount of data is huge. Simplifying structure without throwing away relevant data is a research challenge - can we a priori decide what is relevant before we have investigated it at all?

Applied Visual Computing focuses on the practical use of VC methods. This is done in close collaboration with researchers from Fraunhofer IGD. Examples can be found in the list of Bachelor and Master theses below. In general, the challenge is to transfer models, ideas, and concepts to working prototypes. For instance: the Kinect generates images from a scene; how can we use these images to extract persons (skeleton-like lines!) and their gestures, and how can we use this to provide an interactive environment? Complicated models exist for modeling and rendering (visualizing) objects in a scene. How can we modify and optimize them to make them real-time and still get a high quality visualization. Or biometric systems: how can we identify people based on their biometric features (eye, fingerprint, ...) - and do it in a secure and safe way?

A A A | Drucken Print | Impressum Impressum | Sitemap Sitemap | Suche Search | Kontakt Contact | Webseitenanalyse: Mehr Informationen
zum Seitenanfangzum Seitenanfang