Image Segmentation Method Based On Fractional Differential

Image segmentation has always been a fundamental and complex problem in thefield of image processing and computer vision.It’s purpose to separate the target fromthe background of interest in the image. So far, a large number of segmentationmethods have been proposed. Among them, the geometric active contour model is awidely concerned about the image segmentation method.Active contour models can be classified as parametric and geometric activecontour models according to the representation of the evolving curve. Geometriccontour active models can be classified into edge-based and region-based according tothe use of image feature.In recent years, fractional differential has attracted more and more attention. Andin many fields of application, for example: diffusion, electrochemical, mechanical,biological, image, signal processing and etc.and we obtained many importantresults.Fractional differential can be regarded as integer order differential of thepromotion, the fractional differential is also no uniform definition of forms, but theseare defined in the certain condition equivalence. Fractional differential and integraldifferential difference for fractional differential has global nature, and the integer orderdifferential is determined by its local point of the decision, which of local properties.This dissertation focuses on geometric active contour models; the main results aresummarized as follows:For edge-based active contour models, the choice of edge stopping function is veryimportant. The edge stopping function is typically defined by the image gradient whichonly depends on local properties of each point. Edge-based active contour models usingthis edge stopping function have the drawbacks of sensitivity to noise and edge leakage.This paper presents a new edge stopping function, which is based on fractionaldifferential. Experiments show that edge-based active contour models using this edgestopping function can obtain the good segmentation results for images with noiseand/or weak edges.