A Fast Object Extraction Based on Semi-Implicit Scheme and Level Set
Qiang-Jun Xie · 2009
The paper proposed a novel semi-implicit scheme numerical iteration for an improved level set object extracting method. Firstly, an improved the C-V's PDE is given by adding a penalized energy term for no re-initializing and replacing the dirac function with the norm of level set function gradient for better globe optimization. Secondly, a new difference scheme is introduced for fast and more accurately segmenting results. In order to shorten the time of every loop, this paper constructed a new semi-implicit scheme, which is unconditional stable and superior to the AOS algorithms on the segmentation speed and accuracy. The third, the paper proposes an evolutional criterion or inequality for ending segmentation and searching the rule of parameters. The experimentations for synthesized and real images show that the new approach is faster and more accurate than the traditional level set methods. Moreover, the initial level set curve can be set freely and the parameters can be adjusted conveniently, and so the proposed approach can be applied in practice more flexibly and semi automatically.