The Implicit Surface Piecewise Smooth Image Segmentation

Image segmentation is one of hot point problems in the research fields of image processing, computer vision etc. It has important applications in medical diagnoses, computer-aided surgery, machine vision, resource classification based on remote sensing and so on.Multiphase image segmentation is an active branch of image segmentation; level set method has topologic self-adaptive capacity. Moreover, variational method which can conveniently integrate image information, such as edge, region, shape, texture and motion field becomes the mainstream method of image Segmentation. The image on implicit surface is an extension of plane image. The models for plane image are not used for implicit surface. Since piecewise smooth image could be more closed to real image than piecewise constant image, the study of piecewise smooth image is more valuable. Some systematic research of variational level set method for multiphase piecewise smooth image segmentation on implicit surface is done. The paper comprises following aspects:Firstly, the Curve Evolution Theory. Secondly, level set method and Mumford-Shah Model are systematically researched and their application in image segmentation field is discussed. Then general regions representation model is brought out based on the region competition strategy of Chan-Vese without overlapping and vacuum problems. Thirdly, we extend the level set method from two-dimensional plane into three-dimensional implicit surface field and build the general model for multiphase piecewise smooth image segmentation on implicit surface. Fourthly, we write program for the new model, then test and verify in practice with artificial image.The model of multiphase piecewise smooth image segmentation on implicit surface using parametric estimation model and different region competition schemes is presented in this paper. Many experiments denote that we get the satisfying results. Finally, future research in this field is suggested aiming at the problems and limitations of this model that have emerged in the experiments.