Segmentation of an image of defected leaf involves distribution of the pixel intensity values of an image into clusters, such that the defected portions or rust-spots in the leaf falls in one such cluster, and the test of the leaf (which is supposed to be normal) contributes to other clusters. Apparently, the process is simple, but confusion creeps in when we deal with the pixels that belong to the boundary of the rust-spots, whether to consider them as a part of the rust and hence draft it into the corresponding cluster, or we consider it to be a part of the normal portion and hence, accommodate it in some other cluster.