Hilfe Wegweiser Impressum Kontakt Einloggen





Perceptual Robotics


Bülthoff,  HH
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;

Wallraven,  C
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;

Giese,  MA
Max Planck Institute for Biological Cybernetics, Max Planck Society;

Externe Ressourcen
Es sind keine Externen Ressourcen verfügbar
Volltexte (frei zugänglich)
Es sind keine frei zugänglichen Volltexte verfügbar
Ergänzendes Material (frei zugänglich)
Es sind keine frei zugänglichen Ergänzenden Materialien verfügbar

Bülthoff, H., Wallraven, C., & Giese, M. (2008). Perceptual Robotics. In Springer Handbook of Robotics, Part G (pp. 1481-1498). Berlin, Germany: Springer.

Perceptual functions are central to many applications in robotics and for the construction of efficient human–robot interfaces. The study of perception in biological systems has revealed important information-processing principles that have been converted to powerful applications in robotics and computer vision. The chapter first discusses two central theories of object recognition: model- and exemplar-based theories. A review of experimental results from the study of object recognition in biological systems suggests that exemplar-based approaches capture important properties of object recognition in the brain. We then discuss how very similar principles have been realized in highly efficient technical systems for object recognition and detection, including realizations that are based on biologically inspired neural architectures. Principles for the efficient processing of complex shapes can be extended to the representation of complex movements and actions. We illustrate this by first reviewing some properties of the cortical mechanisms of the recognition of complex movements and actions, focusing on principles that are useful for robotics applications. Again, exemplar-based approaches seem to capture important properties of motion recognition in the brain, and at the same time provide a powerful approach for building technical movement recognition systems. Finally, it is shown that the example-based framework is not only useful for recognition, but also provides the basis for powerful synthesis methods. As one example we discuss the synthesis of photorealistic three-dimensional (3-D) models of faces, exploiting correspondencebetween training examples. Related approaches have been developed for spatiotemporal patterns. We review a class of algorithms that permit the accurate modeling of movements and movement styles by interpolation between example trajectories with high relevance for the synthesis of movements, e.g., in humanoid robotics.