基于PDE的快速自适应图像分割算法

Fast Adaptive Image Segmentation Algorithm Based on PDE

在水平集理论和Chan-Vese模型的基础上,详细分析了水平集函数的初始位置和分割速度的关系,指出如果初始位置不好,将会影响曲线的演化速度。针对该问题提出了利用阈值分割技术对水平集函数进行初始化,通过优化水平集函数的初始位置来加快Chan-Vese模型的演化速度。实验结果表明,本文所提出的方法不受初始轮廓位置的限制,具有鲁棒性,分割效果良好,有很好的实际意义。

Based on the level set theory and the Chan-Vese model,the relationship between the initial position of the level set function and the speed of division is analyzed in detail,and the poor initial position can affect the speed of the curve evolution.For the problem,initializing the level set functions by thresholding segmentation technique,and speeding up the evolution of the Chan-Vese model by optimizing the initial position of the level set function.Experimental results show that the proposed method is not restricted by the location of the initial contour.Besides,it has robustness,good segmentation results and practical significance.