COMBINATORIAL APPROACH FOR IMAGE SEGMENTATION

Abstract

A novel algorithm called sub-Markov random walk (subRW) algorithm is used for image segmentation along with prior nodes. Various random walk methods are used for clustering of objects and a combination of certain random walk methods will produce the sub-Markov random process. It is used to solve twigs segmentation problem by adding auxillary nodes along with label prior. The random method solves certain segmentation issues among thin and elongated objects. The proposed method solves the complex texture issues in natural images. It performs better than other random walk algorithms