An Image Warping Method Based on Level Set

An efficient image warping algorithm for gray and color images is proposed on the basis of level-set method. The image can be warped locally and globally through level-set dynamic evolution with a certain velocity field. We study the solution of the level-set function and define some velocity fields to control the deformation. In our scheme, the computational speed is very fast since the level-set function is not necessarily reinitialized and the CFL condition is not strictly satisfied. We exploit our approach to facial deformation. A lot of experiments were conducted to verify our algorithm, and the experimental results show that our algorithm can produce very complex and pleasurable image warping effects.