Automatic Masked Morphing for 3D Facial Animations


This paper presents an algorithm for a novel implementation of morphing 3D objects representing different facial expressions of the same head model, using masks. Specific selections of the face are morphed towards one specified target, whereas other selections of the same object are morphed after other targets. This way, novel facial expressions could be fastly processed, saving loads of the 3D modelers' time. Our algorithm also brings about a way to smooth out the unnatural aspect that appears on the model (after morphing) at the borders of the morphed selections, by partially applying the morphing transformation on a transition area surrounding the selection. Linear or non-linear functions are used to model the influence of morphing on this area. As a result, natural and lifelike new facial expressions of the model are saved with minimal human intervention (based on a set of manually modeled facial expressions). Most of all, we also present an algorithm that automatically calculates the transition area by determining the influence of the morphed selection on each row of vertex neighbors of each vertex on the selection border, thus even more reducing human intervention and improving the results. The test 3D head-model and its set of different facial expressions that were used as a basis for research in this paper were created (from concept scratch to actual 3D modeling) by the author of this paper. They served as a good work material for testing the efficiency of our algorithm. The implementation was made in C++, using OpenGL.

  • Abstract
  • Key Words
  • 1. Introduction
  • 2. Preliminaries
  • 3. Masked Morphing Algorithm
  • 4. Results and Conclusions
  • References

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In