An Improved Normalized Cut Image Segmentation Algorithm with k-Means Cluster

Abstract

Image segmentation with low computational burden has been highly regarded as important goal for researchers. One of the popular image segmentation methods is normalized cut algorithm. But it is unfavorable for high resolution image segmentation because the amount of segmentation computation is very huge [1]. To solve this problem, we propose a novel approach for high resolution image segmentation based on the Normalized Cuts. The proposed method preprocesses an image by using the normalized cut algorithm to form segmented regions, and then use k-Means clustering on the regions. The experimental results verify that the proposed algorithm behaves an improved performance comparing to the normalized cut algorithm.