Feature-Based Image Metamorphosis Optimization Algorithm

  • Authors:
  • Hussien Karam;A. Hassanien;M. Nakajima

  • Affiliations:
  • -;-;-

  • Venue:
  • VSMM '01 Proceedings of the Seventh International Conference on Virtual Systems and Multimedia (VSMM'01)
  • Year:
  • 2001

Quantified Score

Hi-index 0.01

Visualization

Abstract

Image metamorphosis using simple features provides flexible and easy control of visual effect. The feature-based image metamorphosis algorithm proposed by Beier and Neely suffer from two problems; speed and strange effects. This paper presents an optimization technique to recover these problems in a reasonable time. The idea is based on local decomposition of metamorphosis into piecewise approximation and then applying Beier's technique to establishe one-to-one correspondence to recover such effects. The combination of all mapping produces the result of foldover-free as well as dominates the ghost effect. The results show that our implementation is efficient in time complexity, smoothly interpolated metamorphosis images with only a remarkably small number of specified feature points, and the speed of computations makes image metamorphosis animation reliable for interactive machine.