一种改进的Snake模型与MRI图像分割

Snake模型分割图像时要求初始轮廓线位于图像特征附近,且难以处理深度凹陷区域。该文在快速Snake模型的目标函数中增加局部面积能量项,扩大了算法捕获图像特征的范围;以边缘增强后的BPV图像的梯度为参数,计算梯度向量流场以代替MRI图像的梯度,提高了算法处理弱边界和凹陷区域边界的能力,优化算法的时间复杂度仍然为O(nm)。实验结果表明,该算法能够有效地分割左心室MRI图像。

An Improved Snake Model and Left Ventricle MRI Segmentation

An initial contour must be set near the feature of the image when using the parametric active contour model(Snake)to segment images.Moreover,the model can not deal with the concave regions in the image properly.A local area energy term is thus added to the target function for capturing the image feature in a wider region.On the basis of edge enhancement algorithm,a block pixel variation(BPV)image is achieved for calculating gradient vector flow(GVF).By replacing the gradient field with GVF,the model can segment the weak edges and concave regions effectively.The complexity of the optimization model is still O(nm).The experiment results demonstrate that this algorithm can segment left ventricle MRI images effectively.