Interactive image segmentation using power watershed and active contour model
Quan Sun, Hui Tian · 2012
A novel approach to interactively segmenting a foreground object from its surrounding background in an image is proposed. The segmentation is done by evolving an initial object contour computed by the power watershed algorithm. The contour evolution is based on a novel scheme using the active contour model. Experiments show that the given approach is very effective for segmenting objects in gray images or images with noises and when the contrast between objects and the background is low.