Vorträge

Organic Computing for autonomous

information uptake

Dr. Rolf P. Würtz
Institut für Neuroinformatik,
Ruhr-Universität Bochum

The talk starts by identifying hallmarks of Organic Computing, a new computer science philosophy with good potential to achieve progress on artificial vision systems. Important parts of this methodology are learning from nature, discretizing continuous dynamics, and the integration of submodalities. The holy grail of automatizing vision systems is autonomous learning of the necessary routines from examples.

After a short overview of the known technique of bunch graph matching examples of applications to the learning problem will be shown. Tracking of facial points is an important prerequisite for video compression and animation. Tracking in general is a difficult problem, which requires global constraints to function anywhere near stable. In the first learning example, constraints for face tracking are learned automatically from bunch graph matching on a large number of frontal images. The second learning example consists of the recognition of genetic diseases which can be diagnosed from the shape of the face by an expert. The learning technique of jet voting achieved high classification rates without explicit rule codeing.

A major drawback of correspondence-based recognition methods is the time-consuming matching operation. I finally present a rapid preselection based on a one-layer network, which lends itself to fusion with correspondence-based matching and tactile object exploration. Even in standalone mode, it achieves respectable recognition rates for objects and facial pose.

Zeit: Dienstag, den 13. Juli 2004, 17 Uhr c.t.

Ort: Seminarraum Karp - Raum 68,
        Neubau Informatik, Haus 64,
        Erdgeschoß