Real-time 3D finger pointing for an augmented desk

  • Authors:
  • Le Song;Masahiro Takatsuka

  • Affiliations:
  • University of Sydney, Sydney, New South Wales;University of Sydney, Sydney, New South Wales

  • Venue:
  • AUIC '05 Proceedings of the Sixth Australasian conference on User interface - Volume 40
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The augmented desk is gaining popularity in recent HCI research. Its layout of a large horizontal screen on the desk enhances immersive and intense collaborative experiences. A responsive and unimpeded input interface is important for an efficient interaction in such an environment. In this paper, we developed a real-time stereo vision-based finger pointing interface for our augmented desk that supports drag-and-drop operation by the tap-and-move of the index finger. The core of our system is a 3D fingertip tracking system, which requires both careful calibration and efficient fingertip localization algorithm. To meet these requirements, we have designed a two-step calibration method that strikes a good balance between accuracy and convenience. Furthermore, based on the chain code representation of the contour, we propose the direction cancellation vector as a tool for fingertip localization. Our algorithm works efficiently with a time complexity of O(n) in term of the length of the chain code. Currently our system allows a user to select and move the displayed contents on the screen directly using his fingertip, and it is applied to the interactive graph drawing paradigm proposed by do Nascimento and Eades (2001). In this application, our real-time pointing interface enables the user to interact with a graph drawing program dynamically, which results in optimal layouts of graphs with maximum symmetry. In the last section, the strength and weakness of our system are discussed, and further suggestions to improve the system are also given.