Fast global segmentation based on the dual formulation of TV-norm

A fast global minimization segmentation model based on total variation is presented around Functional modeling and algorithm constructing. Firstly, a new active contour model is developed by maximum a-posterior probability (MAP), and a total variation model based on gradient information is constructed by the hint of geodesic active contour (GAC) model. So the improved M-S segmentation model is given by combining the upper two models. Secondly, we establish theorems on the existence of the global minimum of this model by equivalent conversion. Thirdly, a new numerical practical algorithm is given through a dual formulation of the total variation norm(TV-norm), which avoids the usual drawback of initializing and re-initializing in the active contour model. We apply our segmentation algorithms on many synthesized and real-world images, and the results show the efficiency by assigning only one or two parameters for melanoma segmenting.