A new approach to rapid image morphing for lip motion synthesis

  • Authors:
  • Anna Buttfield

  • Affiliations:
  • Department of Computer Science & Software Engineering, The University of Western Australia, 35 Stirling Highway, Crawley, W. A.

  • Venue:
  • ACSC '03 Proceedings of the 26th Australasian computer science conference - Volume 16
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Animating a human face with visual speech in a way that is accurate enough to look natural and aid both conscious and subconscious lipreading is a major goal of animation systems. One method of creating such animations is to collect a set of images of a human subject speaking, then combine these images together in a manner simulating natural speech. Image morphing techniques can be used to create transitions between the static images. A range of morphing techniques have been used in this context, covering traditional techniques and methods that reflect the specific problem domain. In this paper I present a morphing technique that combines the image warping techniques of traditional field morphing with a rapid morphing method developed specifically for facial animation. This new technique avoids complex computation and difficulties involved in optical flow calculation as is used in other methods, and instead uses simple geometrical correspondence to create smooth transitions. This allows for simplicity and flexibility in calculation while preserving the speed of image generation. The technique assumes that the motion of pixels between the two images can be approximated by a vector field; this vector field can then be calculated and stored. New intermediate images can then be generated rapidly by morphing the images along the calculated vector flows.