Machine Interpretation of Facial Expressions

  • Authors:
  • J. F. Baldwin;S. J. Case;T. Martin

  • Affiliations:
  • -;-;-

  • Venue:
  • BT Technology Journal
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

Facial expressions are an important source of information for human interaction. Therefore, it would be desirable if computers were able to use this information to interact more naturally with the user. However, facial expressions are not always unambiguously interpreted even by competent humans. Consequently, soft computing techniques in which interpretations are given some belief value would seem appropriate. This paper describes how the mass assignment approach to constructing fuzzy sets from probability distributions has been applied to the low-level classification of pixels into facial feature classes based on their colour. It will also describe how similar approaches can be used for the analysis of facial expressions themselves.