《计算机应用研究》|Application Research of Computers

Bregman全散度水平集图像分割方法

Level set image segmentation model based on total Bregman divergence

免费全文下载 (已被下载 次)  
获取PDF全文
作者 李红蕾,王翊
机构 1.重庆电子工程职业学院 人工智能与大数据学院,重庆 401331;2.重庆大学 计算机学院,重庆 400044
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)06-065-1916-05
DOI 10.19734/j.issn.1001-3695.2019.01.0027
摘要 由于图像来源的广泛性和成像条件的复杂性,使得图像分割是一个极具挑战性的问题。针对传统活动轮廓模型不适用于噪声、弱边缘图像分割的问题,提出基于Bregman全散度的全局优化分割方法。首先用Bregman全散度替换传统模型中的<i>l</i><sub>2</sub>测度,构造能量泛函;然后构造全局最优解求解方法,交替迭代求解最优解,得到最终的目标边界;最后在模拟图像、医学图像和自然图像场景下进行实验对比。对比实验结果表明,该分割方法具有较高的鲁棒性和抗噪能力,能准确地分割出具有噪声、弱边缘的目标区域。
关键词 图像分割; 活动轮廓模型; Bregman全散度; 能量泛函; 全局优化
基金项目 国家自然科学基金资助项目(61672120)
本文URL http://www.arocmag.com/article/01-2020-06-065.html
英文标题 Level set image segmentation model based on total Bregman divergence
作者英文名 Li Honglei, Wang Yi
机构英文名 1.Artificial Intelligence & Big Data College,Chongqing College of Electronic Engineering,Chongqing 401331,China;2.School of Computer Science,Chongqing University,Chongqing 400044,China
英文摘要 Image segmentation plays an important role in the research field of computer vision and it is the foundation of high-level semantic analysis. However, image segmentation is a very challenging problem due to the wide range of image sources and the complexity of imaging conditions. Focusing on the problems that the traditional active contour models are not fully suitable for image segmentation with noise and weak edge, this paper proposed a global optimal segmentation method based on TBD. Firstly, the TBD replaced the original <i>l</i><sub>2</sub> measure in the traditional models to construct the segmentation energy functional. Then it applied a solution to obtain the global optimum alternately. The optimal solution represented the final target boundaries. Finally, experimental results on synthetic, medical and nature images validated the high robustness and noise immunity of the proposed model. It can achieve the object boundaries in noise and weak edge accurately and efficiently.
英文关键词 image segmentation; active contour model; total Bregman divergence(TBD); energy functional; global optimization
参考文献 查看稿件参考文献
 
收稿日期 2019/1/3
修回日期 2019/3/27
页码 1916-1920
中图分类号 TP391.41
文献标志码 A