Motivation – Same as above with specialty that region to be identified may large in terms of area. Thus visiting each neighborhood pixel (an exhaustive approach is costly). Optimization is possible if we try a bottom up approach. Method – divide an image into regions of arbitrary size. Then these regions are merged or split depending upon the requirement for Magniwork Review. Split the region if all pixels don’t satisfy the criteria of common property and merge regions that are adjacent and satisfy the criteria. Splitting is done for regions that are greater than a threshold. The algorithm may be exemplified as Input – an image and Output – segmented image.
