Unencumbered Gestural Interaction

  • Authors:
  • Francis K. H. Quek

  • Affiliations:
  • -

  • Venue:
  • IEEE MultiMedia
  • Year:
  • 1996

Quantified Score

Hi-index 0.00

Visualization

Abstract

This article presents work on unencumbered hand-gesture interfaces. It encompasses both three-dimensional interaction and two-dimensional pointing. In three-dimensional gestures, the computational approach is motivated by a brief overview of human hand-gesture interaction. The model requires determining the gestural stroke, recognizing hand poses at the extrema of the stroke, and determining the dynamics of hand motion during the stroke. Work is presented on the inductive learning of hand-gesture poses using extended variable-valued logic and a rule-based induction algorithm. The author and his colleagues were able to attain a 94% recognition rate with their system. He discusses their work in the computation of the image flow fields representing the moving hand.Readers may contact Quek at the Electrical Engineering and Computer Science Dept., University of Illinois at Chicago, Chicago, Ill. 60607, e-mail quek@eecs.uic.edu.