图像分割中的交叉熵和模糊散度算法  

lmage Segmentation Algorithms Based on Cross Entropy and Fuzzy Divergence

作　　者：薛景浩[1] 章毓晋[1] 林行刚[1] 

机构地区：[1]清华大学电子工程系,北京100084

出　　处：《电子学报》1999年第10期131-134,共4页Acta Electronica Sinica

基　　金：国家自然科学基金

摘　　要：本文将交叉熵和模糊散度应用于图像分割中，提出了四种最优灰度阈值选取算法．其一是基于均匀分布假设的最小交叉熵算法，其二是利用后验概率的最大类间交叉熵算法，其三是类间最大模糊散度的改进算法，其四是最小模糊散度算法．针对图像阈值化分割的要求，在后两种算法中构造了一种新的模糊隶属度函数．本文采用均匀测度和形状测度比较各算法的性能．利用多种类型测试图像得到的分割结果，显示了所提算法的有效性和通用性．Cross entropy measures information discrepancy between two probability distributions.Induced by cross entropy,fuzzy divergence measures dissimilarity between two fuzzy sets,as fruit of both information theory and fuzzy set theory,In this paper,in the light of different criteria we present four new algorithms of optimal gray scale threshold selection for image segmentation,integrating cross entropy and fuzzy divergence with image histogram.The first algorithm is based on minimum cross entropy with the hypothesis of uniform probability distribution. The second algorithm maximizes between classcross entropy using posterior probability.The third one is a modified version of existing method based on maximum betweenclass fuzzy divergence.The last one is a minimum fuzzy divergence algorithm.According to the requirement of image thresholding,we construct a new fuzzy membership function in the last two algorithms.The effectiveness and generality of our new algorithms are shown by applying them to various test images and by evaluating the results with uniformity measure and shape measure.

关 键 词：图像分割 交叉熵 模糊录属度函数 模糊散度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN919.8[自动化与计算机技术—计算机科学与技术]