A New Marker Extraction Method Based on Skeleton of Gradient Image

A new marker-extraction method is proposed to extract the marker which is needed in the preprocessing of watershed algorithm. Firstly, the proposed method marks all the valley points in the gradient image to get a connected skeleton; then, picks out and eliminates the saddle points whose gray values are larger than the mean value of the whole gradient image. After that, the skeleton is separated into some little discrete skeletons, each of them is a marker and correspond a segmented area. The experimental results show that the proposed method can not only avoid the difficult of parameter initialization, but also decrease the problem of over-segmentation, and a satisfactory segmentation result can be obtained.