【摘要】 针对部分模型存在轮廓初始化敏感、迭代次数多的问题,本文提出了一个新的基于区域的图像分割模型。该模型在水平集演化过程中增加了能量惩罚项,使得水平集在演化过程中保持近似的符号距离函数,进而不需要重新初始化。通过实验结果可以看出:本模型在不需要初始轮廓的情况下,不仅具有速度更快、迭代更稳定等优点,而且对轮廓初始化是更鲁棒的。更多还原

【Abstract】 In order to solve the problem that some models are sensitive to contour initialization and have many iterations,a new region based image segmentation model is proposed in this paper.In this model,the energy penalty is added to the evolution process of the level set,so that the level set keeps the approximate symbolic distance function during the evolution process,and thus does not need to be reinitialized.The experimental results show that this model not only has the advantages of faster speed and more stable iteration,but also is more robust to contour initialization without initial contour.更多还原