Improving Watersheds Image Segmentation Method with Graph Theory

This paper presents a new image segmentation method - FWTN (first watershed then normalized cut) based on watersheds and graph theory to solve the over-segmentation problem of watersheds. FWTN firstly uses normalized cut to segment between regions after applying watersheds, and then generates the final segmented images. The algorithm can successfully solve over-segmentation problem, and at the same time improve the NP-hard problem of normalized cut. Experimental results show that FWTN is efficient and practical for image segmentation.