基于最大交叉熵的PCNN简化模型在图像分割中的应用  

Application of the Simplified PCNN Model Based on Max Cross Entropy in Image Segmentation

作　　者：夏辉[1] 穆希辉[1] 马振书[1] 兰箭[1] 杜峰坡[1] 

机构地区：[1]军械工程学院军械技术研究所,河北石家庄050000

出　　处：《军械工程学院学报》2007年第1期58-60,72,共4页Journal of Ordnance Engineering College

基　　金：“863”计划项目（2001AA422420）

摘　　要：用PCNN作图像分割时，选择不同的模型参数值和停止判决准则，图像分割的差异很大。分别选择最大熵和最大交叉熵作为PCNN图像分割准则，并对PCNN模型进行合理简化。实验结果表明，基于最大交叉熵准则PCNN迭代次数更少，图像分割效果明显优于最大熵准则。When carrying through the image segmentation, the different model parameters and ending criterions will lead to the great differences of the image segmentation results. In the text the max entropy and the max cross entropy are respectively selected as the image segmentation criterions and the PCNN model has been rationally simplified. The experiment results prove that the PCNN model based on the max cross entropy criterion has less iterative times and better image segmentation effects than the one selecting the max entropy dose.

关 键 词：最大熵 最大交叉熵 PCNN 图像分割 

分 类 号：TN911.73[电子电信—通信与信息系统]