Technique of Mean Shift is used for segmentation of present image and it can be clubbed together with numerous other techniques plus applications. As one of its feature of vast time consumption other algorithms are considered over mean shift. But feature of discontinuity preservation and edge preservation of an image is a plus point of this technique. The work emphasis on technique of Mean Shift in detail with some alteration, so that disadvantage of Standard Mean shift can be eliminated. An alteration like K- distance based technique is used for estimating bandwidth of the kernel. To limit the size of the cluster a termination condition is defined. Extracting the feature using GLCM technique to calculate features of present image and EDISON Mean Shift due to which time taken to calculate Mean Shift is reduced. The result observed due to Mean Shift is an over segmented image and it is further grouped or merge to form segmented image using merging technique.