In this talk, we will present some new approaches in using operator splitting methods to get fast numerical algorithms for the minimization of Elastica energy for functions. The algorithms are fast and robust and nearly parameter free. Application to image restoration, segmentation, especially our recent efforts in adding convexity shape priori into geometry models, will be explained in details. Numerical experiments will be given to show the applications of the algorithms for real challenging applications.

This presentation is part of Minisymposium “MS15 - Nonlinear Diffusion: Models, Extensions and Algorithms (2 parts)”
organized by: Ke Chen (University of Liverpool) , Joachim Weickert (Saarland University) , Xue-Cheng Tai (Hong Kong Baptist University) .

Authors:

Xue-Cheng Tai (Hong Kong Baptist University)

Keywords:

computer vision, image enhancement, image reconstruction, image segmentation, partial differential equation models