Image Segmentation Based On Partial Differential Equation

Image segmentation is the most essential step from preprocess of image to imageanalysis, its goal of image segmentation is to divide an image into disjoint sub-domains,each of which has different character and homogeneous. Currently, none of genericmethod is suit to all of image in image segmentation community. Recently, theresearchers constantly committed to improve the existing method and propose manynovelty segmentation method combined with the theory of others relative subject.Among them, Partial differential equation based image segmentation method has beenextensively applied in the image processing field, and it has been the hot spots of mostscholars.This paper first gives a brief review for the image segmentation based on thepartial different equation and the existing method, and then focus on introducing tworegion based active contour model: Chan-Vese model and LBF model, lastly, studies theLGIF model specially. We obtain the following conclusion:Given the fault of higher computation complex and lower segmentation efficiency,this paper constructs a novel local fitting term, and combined with the global fittingterm, it compose a new model. The experiments show that the model is not only able todeal with intensity inhomogeneity, but also the evolving speed and the segmentationefficiency is more efficient than LGIF model.