Facial analysis and synthesis scheme

  • Authors:
  • Ilse Ravyse;Hichem Sahli

  • Affiliations:
  • Department ETRO, Audio Visual Signal Processing (AVSP), Vrije Universiteit Brussel, Brussel;Department ETRO, Audio Visual Signal Processing (AVSP), Vrije Universiteit Brussel, Brussel

  • Venue:
  • ACIVS'06 Proceedings of the 8th international conference on Advanced Concepts For Intelligent Vision Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

We developed an algorithmic scheme to extract the semantical description of the face and the face motion from an image sequence, and to re-play this action in a 3-dimensional (3D) virtual world. The presented Facial Analysis and Synthesis Scheme combines new methods for detection and tracking of the face and facial features, for estimating the 3D face movements and the nonrigid facial expressions, and for extracting the MPEG4 facial animation parameters. In the scheme, the face is treated either as a 2D object that has specific color, shape and motion characteristics, either as a 3D model that is calibrated and moved using a natural displacement-based deformation model. A dynamic MPEG4 displacement table takes care of the semantical controls of the animations of the face model. As a result, this virtual face model mimics well the gestures of the person in the video.