Seeing the Unseen: Segmenting with Distributions

An efficient method for separating an object from the background in an image is presented. The segmenting curve, corresponding to the object boundary, is represented as the zero level set of a signed distance function. Most existing region based methods in the geometric active contour framework perform segmentation by maximizing the separation of intensity moments inside and outside the evolving contour. We generalize these methods by minimizing the Bhattacharyya distance so that it separates regions of different distributions. Preliminary results show that the proposed method can segment low contrast, complex images with a very simple curve flow equation.