Real Time Face and Object Tracking as a Component of a Perceptual User Interface

  • Authors:
  • Gary R. Bradski

  • Affiliations:
  • -

  • Venue:
  • WACV '98 Proceedings of the 4th IEEE Workshop on Applications of Computer Vision (WACV'98)
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

As a step towards a perceptual user interface, an objecttracking algorithm is developed and demonstratedtracking human faces. Computer vision algorithms that areintended to form part of a perceptual user interface mustbe fast and efficient. They must be able to track in realtime and yet not absorb a major share of computationalresources. An efficient, new algorithm is described herebased on the mean shift algorithm. The mean shiftalgorithm robustly finds the mode (peak) of probabilitydistributions. We first describe histogram based methodsof producing object probability distributions. In our case,we want to track the mode of an object's probabilitydistribution within a video scene. Since the probabilitydistribution of the object can change and movedynamically in time, the mean shift algorithm is modifiedto deal with dynamically changing probabilitydistributions. The modified algorithm is called theContinuously Adaptive Mean Shift (CAMSHIFT)algorithm. CAMSHIFT is then used as an interface forgames and graphics.