A combined gradient vector flow and mean shift approach to image segmentation

Classical gradient vector flow (GVF) based segmentation has been shown to work less well when other significant edges are present adjacent to the real boundary. To counter this, in this paper, we propose an improved energy function by consistently reducing the Euclidean distance between the inspected centroid of the real boundary and the estimated one of the snake. Experimental results show that our new method outperforms the classical GVF algorithm.