一种改进的VFC Snake模型

Improved VFC Snake model

为了克服VFC(vector field convolution)Snake模型对图像弱边界的泄露问题,对其作了两点改进:a)重新定义了向量场核的模,有效降低了VFC外力对向量模参数的敏感性;b)综合利用VFC外力和图像势能力,给出动态VFC外力。随着模型曲线的形变,不断调整外力,使得曲线精确定位到目标边界上。最后通过实验证明了改进后的VFC Snake模型对噪声具有鲁棒性、对参数变化不敏感,且能够收敛到图像弱边界处。

In order to alleviate the weak edges leakage problem of the VFC（vector field convolution） Snake model, improved two points. First, redefined the magnitude of the vector field kernel, which reduced the sensitivity between the capture range of the external forces and the change of parameter. Second, proposed a dynamic force, the combination of the VFC force field with the potential force field. At last, exampled the segmentation of a head MRI image and compared the performance about the GVF、the VFC and the improved VFC. Experiments demonstrate the improved VFC model is superior noise robustness and parameters bluntness and converge to the object boundary accurately.