Comparative study of image segmentation algorithms
I. V. Gribkov, P. P. Koltsov, Н. В. Котович, A. A. Kravchenko, A. S. Koutsaev, Andrey S. Osipov, A. V. Zakharov · 2008
A considerable amount of image segmentation methods gives rise to the problem of method's choice, most adequate for practical purposes. In this paper we study some properties of four digital image segmentation methods with the aid of our PICASSO (PICture Algorithms Study Software, [1]-[3]) program system. PICASSO's datasase accumulates artificial image samples both typical for the real images and difficult for prosessing by image processing methods. Like in the PICASSO general approach, the comparative study of operating quality of segmentation methods is fulfiled using artificial test images with known true segmentation. The description of test images and testing procedures are given. Our approach allows to clear up specific features and applicability limits of the segmentation methods under examination.