一种基于Jeffrey散度的鲁棒图像分割方法  

A Robust Image Segmentation Method Based on the Jeffrey s Divergence

作　　者：单晓英 任迎春 Shan Xiaoying;Ren Yingchun(Pinghu Normal College,Jiaxing University,Jiaxing,Zhejiang 314000;College of Data Science,Jiaxing University,Jiaxing,Zhejiang 314000)

机构地区：[1]嘉兴学院平湖师范学院,浙江嘉兴314000 [2]嘉兴学院数据科学学院,浙江嘉兴314000

出　　处：《嘉兴学院学报》2022年第6期21-30,共10页Journal of Jiaxing University

基　　金：浙江省自然科学基金(LQ20F020027);浙江省教育厅科研项目(Y202044497)。

摘　　要：由于成像设备不完善或外界干扰等因素,图像经常有灰度不均现象,一种基于Jeffrey散度相似性度量的鲁棒图像分割算法具有更高的分割精度和分割效率:建立基于Jeffrey散度的灰度拟合项,以提升算法分割灰度不均图像的能力;整合长度正则项、符号距离函数惩罚项和灰度拟合项构建总能量泛函;通过最小化能量泛函实现偏置场修正和灰度不均图像分割.为提高算法的分割效率和对初始轮廓的鲁棒性,可引入一种新的偏置场初始化方法.该算法的有效性在一些自然图像、医学图像和合成图像的实验中得到验证.Due to factors such as imperfections of imaging devices or external interference,images are often inhomogeneous.Therefore,a robust image segmentation method based on the Jeffrey s divergence similarity measure with higher precision and efficiency is proposed in this paper.First,to improve the ability of the algorithm to segment the inhomogeneous images,an image intensity fitting term is constructed based on the Jeffrey s divergence.Then,the total energy functional is constructed by integrating the length regularization term,the symbolic distance function penalty term and the intensity fitting term.Finally,by minimizing the energy functional,the proposed method can correct the bias field and segment the inhomogeneous image.Besides,a new bias field initialization is introduced to improve the segmentation efficiency and the robustness to the initial contour.The experimental results on some nature images,medical images and synthetic images can prove the effectiveness of the proposed model.

关 键 词：Jeffrey散度 偏置场 初始化方法 主动轮廓模型 图像分割 

分 类 号：TP391.413[自动化与计算机技术—计算机应用技术]