Medical image segmentation via coupled curve evolution equations with global constraints

In this work the authors modify the coupled curve evolution approach to snakes presented by the authors in previous work for bimodal and trimodal imagery through the introduction of global constraints. The key idea, as before, is to derive curve evolution equations which “pull apart” the values of one or more statistics within the image. However, by imposing a new constraint on the evolution of these statistics, the authors are able to segment a larger class of medical imagery for which their original model would fail