Image segmentation based on novel adaptive bidirectional balloon force model

摘要

A novel adaptive bidirectional balloon force model for image segmentation is proposed in this paper. Active contour model based on a balloon force field has been widely used to detect object boundaries. However, the unidirectional movement property demands that the active contour should be completely placed inside or outside the boundary, and this can easily cause leaking through weak edges and seriously limit the application of the balloon force model. Therefore, the adaptive bidirectional balloon force model based on magnetostatic field is studied as a new force model. Experimental results show that, compared with gradient vector flow model (GVF) and vector field convolution model (VFC), the proposed method significantly improves the active contour in capturing complex geometries and dealing with difficult initializations.