A region is obtained when none of the neighboring pixels satisfy the chosen criterion. Further another region is grown and so on until every pixel belongs to a region. The differences between the existing regions growing methods are based on the selection method for the seed pixel/pixels and
the mathematical implementation of the chosen similarity criterion. Many of these methods set the initial seed as the firs pixels of the image, corresponding to the upper left corner, and choose the following seed as the first pixel of the image that doesn’t belong to a region given that the image is processed from left to right and top to bottom. In split and merge the basic mechanism uses two stages for region based segmentation. The first stage is splitting the image into uniform macro blocks of pixels using a homogeneity criterion. It’s a top down process starting from the entire image and going down to pixel level. First the homogeneity of the entire image is evaluated. If the testing shows an inhomogeneous image then the image is split into 4 equal and non-overlapping parts and for every obtained part the same homogeneity criterion is verified and the splitting is applied if necessary. At the end of the first stage a set of homogeneous pixel blocks of various dimensions are obtained. The second stage is merging the neighboring macro blocks with the same characteristics using the same or a different similarity criterion [7].