The improvement of a variational level set formulation for image segmentation

This paper shows a new improvement variational formulation for geometric active contours that make sure the level set function to be close to a signal distance function. The variational formulation consists of an internal energy term that penalizes the deviation of the level set function from a signal distance function. An external energy term that drives the motion of the zero level set toward the object boundaries. Therefore eliminates the need of the costly re-initialization procedure. Upon simulation experiments present that the method is fast and applicable way for application in image segmentation. This method not only simplifies the calculation, but also the iteration can be set longer than the traditional method in time-step so that the evolution of the curve faster. Its flexibility in initializing level set function makes the selection of initial contour has more freedom, and calculations are a lot simpler.