In medical imaging, the algorithm data is classified based on thresholds, clustering techniques and deformable models. The algorithms based on thresholds assert that the more interesting structures or organs have distinctive quantifiable features which include image intensity or the gradient magnitude.  Clustering techniques are the most popular ones for the medical image segmentation.  The two leading members of these algorithms are supervised classification algorithms and unsupervised classification algorithms.  The algorithms based on deformable models are more flexible and used for complex segmentations.