Robust Active Contours for Fast Image Segmentation


A robust active contour model is proposed for fast image segmentation. By introducing the intensity fitting energy in a local region, the proposed model can segment the images with intensity inhomogeneity efficiently. Because the local fitting functions are computed before curve evolution, the proposed model is insensitive to initialization and has a high segmentation efficiency. Experiments on several synthetic and real images have proved the effectiveness of the proposed model.