基于期望最大化的水平集分割算法  

Level set image segmentation based on EM algorithm

作　　者：高尚兵[1,2] 严云洋[1,2] 

机构地区：[1]淮阴工学院计算机工程学院,江苏淮安223001 [2]南京理工大学计算机学院,江苏南京210094

出　　处：《计算机工程与设计》2011年第7期2436-2438,2466,共4页Computer Engineering and Design

基　　金：江苏省高校自然科学计划基金项目(08KJB520001);江苏省"青蓝工程"基金项目

摘　　要：针对经典的水平集算法(比如Chan-Vese模型算法)在迭代过程中要重新初始化和容易受噪声和模棱两可的边界的影响的缺点,增加一项内部能量泛函达到不需重新初始化的目的,并结合贝叶斯决策理论,利用图像先验知识,提出了一个改进的能量函数,根据符号距离函数来不断调整水平集函数的偏差。该函数是利用期望最大化算法来得到的。实验结果表明,该算法分割精度和运行准确率上都优于经典算法。Aimed at the shortcomings of the classical level set methods such as the Chan-Vese model algorithm in the iteration process to re-initialize and easily affected by noise and ambiguous boundaries, an internal energy functional is added to achieve the purpose of without re-initialization, and prior knowledge of image is combined with Bayesian decision theory to propose an improved energy function to tackle this problem by continuously rectifying the deviation of the level set function according to the signed distance function. This is achieved using an expectation-maximisation algorithm. Experimental results shows the proposed algothim is better than the classical image segmentation on precision and accuracy.

关 键 词：图像分割 期望最大化算法 递归 水平集 贝叶斯决策 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]