The process of image segmentation is one of the most important steps in computer vision forimage retrieval, visual summary, image based modeling and in many other processes. The goalof segmentation is typically to locate certain objects of interest. In this paper, we have studiedand investigated graph based normalized cut segmentation methods and proposed a techniquefor adding flexibility to the parameters for performance improvement. These methods areexamined analytically and tested their performance for the standard images. The resultsobtained for the important metrics show that these methods perform better than othersapproaches and are computationally efficient, and useful for precise image segmentation.