Research And Application Of Chan-Vese Image Segmentation Model

In this paper, which based on the theory of partial differential equation for research and application of image segmentation model,then we do some research and improved the CV and LBF model. First chapter of this paper we mainly summarized the method of image segmentation and the image segmentation method based on variational partial differential equation, then overview several classical variational partial differential equation model for image segmentation. The second chapter mainly introduces some of the paper involves basical definitions and theorems. The third chapter we introduces an improved image segmentation model,namely, with the regularization of (the local binary fitting model, LBF) binary image segmentation model, putting forward a new model by adding the regularization of energy functional based on the LBF model, through theoretical analysis, and the existence and uniqueness of solution are given. Then using variational method and gradient descent method for new models of energy functional,then get the Euler-Lagrange equation correspondingly, the application of finite difference method solve the Euler-Lagrange equation. The numerical experiments show that the new model of the image segmentation effect is relative good for uneven grey value and a slight noise image, compared with the CV and LBF model, then the advantages of the new model can be seen. The fourth chapter we presents the second improved image segmentation model, namely, the punishment term is contained this new segmentation model, then the well-posedness of the solution of this model to do the theoretical analysis, then its numerical algorithm and the simulation results are given. Experiments show that the new model is better for the images of gray level values are not uniform and contrast is not high.