The Application of Energy-Driven Watershed Transform Algorithm to Hippocampus Segmentation in MRI

Traditional watershed transform algorithm is flexible,but some regional characteristics in the boundary are seldom used.So,over-segmentation is serious in segmentation results.A method is presented,which employs energy-driven watershed transform algorithm to segment hippocampus in human brain MRI.For this method,the watershed transform model is used.Water floods to each pixel along the shortest path from seed points,then computes the cost of flooding across the pixel.The cost can be used as the energy of this pixel.Driven by energy minimizing,the class to which the points in the contour belong is modified to make the contour develop into the ultimate result.Some inner characteristics and boundary condition are used to limit over-segmentation.The segmentation results show this method can be applied to the segmentation of some complex structures,such as hippocampus.