一种向量场卷积外力加速的GAC模型

A GAC Model Accelerated by Vector Field Convolution External Force

结合参数活动轮廓模型和几何活动轮廓模型的优势,提出一种向量场卷积(VFC)外力加速的侧地活动轮廓(GAC)模型。利用外力场优势,通过引入基于VFC力场的双向边界吸引力和自适应膨胀力,自适应地调整模型曲线的演化方向。实验结果表明,该模型可以克服参数活动轮廓模型不能处理拓扑结构变化的问题,避免传统GAC模型曲线单边演化的现象,能够提取深度凹陷的目标边界,对初始轮廓不敏感,对图像噪声和弱边界具有较高的鲁棒性。

This paper proposes a Geodesic Active Contours（GAC） model accelerated by the Vector Field Convolution（VFC） external force, which takes advantage of parameters active contour models and geometric ones. Based on the advantage of the external forces field, the bi-directional boundary attractive force and the adaptive expansive force are introduced in the model, so it can adaptively adjust the direction of the evolution curves＇. Experimehtal results show that the model overcomes the defect that parameters active contour models can not handle topology change of target structures, and the weakness of one-side evolution of the traditional GAC model. Moreover, the model can also extract the concavity boundaries, and is robust to the image noise, weak edges and the initial contour of the model.