DETAILED IMAGE SEGMENTATION USING NORMALISED CUTS AND WEIGHTED CO- EFFICIENT
Najmuzzama Zerdi, Subhash Kulkarni, V. D. Mytri · 2014
Image segmentation is one of the vital steps in th e processing of image which includes dividing the image pixels in to salient image regions. These may be the regions cor responding to individual surfaces, objects, or natural parts of o bjects which could be used for object recognition, occlusion bou ndary estimation within motion or stereo systems, image c ompression or image editing. Normalized cut method of image segmentation is one of the intensively used image segmentation a pproach which deals with clustering a set of elements based only on the values of a similarity measure between all possible pairs of elements. Along with this method if weighted graph based analysis of clustered image is used, that would help in reducing the comp utational complexity and provides new vision in image segmentation. In this paper weighted graphical co-efficient and normalized cut methods are combined to get the optimal segmentation results. Once the image is segmented using normalized cut method, clusters are analyzed using weighted pixel paramete rs and by taking their co- efficient into consideration.