Image Segmentation Method Based on Hierarchical Region Merging's NCUT

Abstract

Image segmentation is the foundation of image analysis and scene understanding. In this paper, a novel image segmentation method based on hierarchical region merging's Ncut is proposed. The image is divided into several regions using SLIC firstly. The region represented with superpixels are clustered by Graph-based Ncut algorithm, and then the clustered regions are further merged by hierarchical region merging algorithm. The Experiment shows that our method can reduce the cost of segmentation time and improve the segmentation results.