Research on Image Segmentation Using Active Contour Driven by Local Region Force
Caikou Chen · Jisuanji fangzhen · 2011
Research the segmentation effects of the active contour model.Focus on the following problems: geodesic is sensitive to the initialization,and sometimes obtains false segmentation under the cross boundary initializations,and the coupled curve active contour model is prone to produce false segmentation with heterogeneous grey distribution of the objective image.We proposed a novel geodesic active contour model combined with an external force field based on region information,which made the proposed model driven by the gradient-based and region-based external force to converge to the boundaries of the objects.Experiments demonstrate that the proposed model can get better performance in both synthetic images and real images,due to the combination of the geodesic model and region information.