A Method of Image Segmentation Based on Adaptive Bidirectional Balloon Force Model

A novel image segmentation method is proposed using geometric active contour model based on an adaptive bidirectional balloon force (ABBF). Since active contour model was used extensively in image segmentation. great progress has been achieved. The evolved active contour model combined with balloon force could effectively detect object boundaries, however, this model require that we must put the initial active contour inside or outside the boundary completely, also it is easy to cause leaking through weak edges. In order to solve these problems, we propose a new ABBF-snake model based on magetostatic field. According to our experiment, it is obvious that our method make great improvement in convergence ability and broken edges, furthermore. compare with GVF and VCF, our method achieved more satisfactory results in segmentation of complex shape objects and real images.