Based On The Multiphase Level Set Image Segmentation Method

Since image segmentation is an important aspect of image processing,its theoretical and practical significance cannot be ignored.Frequently,people only need to analyze or process some regions that interest them in an image.Image segmentation technology is to divide an image into some sub-regions according to the image's intensity,texture,color,and optical flow etc.Level set method is an important means for image segmentation.It can automatically deal with the changes in images' topologies and easily be extended from 2D segmentation to 3D segmentation.Some detailed research of variational level set method for image segmentation is done and an innovative multiphase image segmentation model which is based on both of the level set fimctions and regional characteristic functions is proposed in this paper.This paper comprises following aspects:Firstly,some traditional image segmentation methods are introduced. Secondly,the Curve Evolution Theory,Level Set method and its application in image segmentation are systematically studied.Thirdly,the general model for multiphase image segmentation which is based on the level set functions and regional characteristic functions is presented.This model well concerns the importance of the regional characteristic functions in image segmentation by adding an approach term which can represent the relationship between level set functions and regional characteristic functions.Experimental results show that the new model transcends the traditional C-V model in both velocity and accuracy and gets promising results using different region competition schemes and probability distribution models.Finally,future research in this field is suggested aiming at the problems and limitations of this model that have emerged in the experiments.