Image Segmentation Method Based on Fractional Varying-order Differential

Image segmentation plays a very important role in many applications. At present, the most difficult problem of image segmentation is segmenting intensity inhomogeneous images, some hybrid methods have better segmentation results than the traditional methods. In this paper, a new hybrid level set method based on fractional-varying-order differential curvature is proposed. We define a new curve evolution curvature calculated by fractional-varying-order differential according to the gradient of the level set function, so that the diverse order differential can be used in a whole curve evolution at the same time, which can describe the image edge more accurately. The energy functional for the proposed model consists of three terms: local term, length term and penalty term. The evolution of the level set function is the gradient flow that minimizes the overall energy functional. Experimental results for both synthetic and real images show desirable performance of our method. The Dice similarity coefficient are employed as the comparative quantitative measures for the segmented results.