Medical Image Segmentation by Water Flow

We present a new image segmentation technique based on the paradigm of water flow and apply it to medical images. The force field analogy is used to implement the major water flow attributes like water pressure, surface tension and adhesion so that the model achieves topological adaptability and geometrical flexibility. A new snake-like force functional combining edge- and region-based forces is introduced to produce capability for both range and accuracy. The method has been assessed qualitatively and quantitatively, and shows decent detection performance as well as ability to handle noise.