水平集在图像分割中的应用研究*

水平集方法的诞生有效解决了以前算法不能解决的在曲线演化过程中的拓扑变化问题,其核心是利用水平集这一数学理论来对能量函数进行极小值求解的曲线演化过程,通过求解极小值最终获取目标轮廓从而达到图像分割的目的。为了解决不同应用领域的图像处理问题,各种相应的基于水平集方法的图像分割算法已被提出,大量的研究者仍在不断地改进和提高这些算法的效率和有效性。对现有的用于部分图像分割的水平集方法进行了综述,主要介绍传统水平集方法、无重新初始化水平集方法、连续水平集方法以及最近相关的改进方法,并简要讨论了各种方法的优缺点以及应用情况,最后指出了水平集方法进一步研究的方向。

Survey on application of level set in image segmentation

The level set method effectively solve the problem of topology changes in curve evolution that the other algorithms can’t solve.The core of the method is to use the mathematical theory of level set to solve the minimal value of energy function,then get the target contour.In order to solve image processing problem of different fields,there have proposed various relevant solutions based on level set method,and a large number of researchers will continue to improve and enhance the efficiency and effectiveness of the algorithms.This paper presented an overview of existing methods about level set used to partial image segmentation,mainly introduced the traditional level set method,level set method without reinitialization,continuous level set method,and recent related improving methods.And it briefly discussed the advantages and application of various methods.At last,it pointed out the further research direction of methods.