Abstract
A novel scheme of external force for detecting the object boundary of medical image based on Snakes (active contours) is introduced in the paper. In our new method, an electrostatic field on a template plane above the original image plane is designed to from the map of the external force. Compared with the method of Gradient Vector Flow (GVF), our approach has clear physical meanings. It has strong ability to conforme to boundary concavities, is simple to implement, and reliable for shape segmenting. Additionally, our method has larger capture range for the external force and is useful for medical image preprocessing in various applications. Finally, by adding the balloon force to the electrostatic field model, our Snake is able to represent long tube-like shapes or shapes with significant protrusions or bifurcations, and it has speciatlty to prevent Snake leaking from large gaps on image edge by using a two-stage segmentation technique introduced in this paper. The test of our models proves that our methods are robust, precise in medical image segmentation.
∗Supported by National Natural Science Foundation of China (Grant Nos. 10275003 and 10175004), and Natural Science Foundation of Beijing (Grant No. 3011002)
Notes
∗Supported by National Natural Science Foundation of China (Grant Nos. 10275003 and 10175004), and Natural Science Foundation of Beijing (Grant No. 3011002)