An Improved Local or Global Active Contour Driven by Legendre Polynomials

Abstract

In the paper, an improved local or global active contour model driven by Legendre Polynomials(LGLP) is proposed. It implemented with a special method, which selectively penalizes the level set function and then uses a filter to regularize it. Firstly, utilizing Legendre Polynomials approximates region intensity. Secondly, an improved region-based signed pressure force (ISPF) function is proposed, which efficiently stop the contours at weak edges, especially for the segmented image with intensity inhomogeneity. Finally, an edge stopping function is added to robustly capture the boundaries of objects. Experimental results show that the improved method is faster and achieve higher accuracy than other models on real images with intensity inhomogeneity, noise and multiple objects.