Person localization and posture recognition for human-robot interaction

Abstract

The development of a hybrid system for (mainly) gesture-based human-robot interaction is presented, thereby describing the progress in comparison to the work shown at the last gesture workshop (see [2]). The system makes use of standard image processing techniques as well as of neural information processing. The performance of our ar- chitecture includes the detection of a person as a potential user in an indoor environment, followed by the recognition of her gestural instruc- tions. In this paper, we concentrate on two major mechanisms: (i), the contour-based person localization via a combination of steerable lters and three-dimensional dynamic neural elds, and (ii), our rst experiences concerning the recognition of dierent instructional postures via a combination of statistical moments and neural classiers.