您的位置:首页  > 论文页面

基于超像素的快速图像分割

发表时间:2015-01-15  浏览量:1946  下载量:1069
全部作者: 程再兴,刘双龙,马尽文
作者单位: 北京大学数学科学学院,数学及其应用教育部重点实验室
摘 要: 采用超像素的思想和方法提出一种快速图像分割方法。首先采用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法对图像进行一定的过分割而形成超像素,然后以超像素为基本单元构造无向加权图,最后对超像素进行自适应谱聚类得到一个较为均衡的图像处理分割结果。实验结果表明:这种基于超像素的快速图像分割新方法不仅拥有很高的运行效率,还可获得较高水平的分割结果。
关 键 词: 应用数学;图像分割;超像素;无向加权图;谱聚类;竞争学习
Title: Fast image segmentation through super-pixels analysis
Author: CHENG Zaixing, LIU Shuanglong, MA Jinwen
Organization: Key Lab of Mathematics and Applied Mathematics, Ministry of Education, School of Mathematical Sciences, Peking University
Abstract: In this paper, a fast image segmentation method is proposed with the help of super-pixels analysis. Specifically, the simple linear iterative clustering (SLIC) algorithm is firstly implemented on the image for a proper over-segmentation to form a set of super-pixels, and then these super-pixels are regarded as the basic nodes to construct an undirected weighted graph, and finally we utilize an adaptive spectral clustering algorithm to classify the super-pixels as well as the corresponding pixels and obtains a balanced image segmentation. It is demonstrated by the experimental results that this proposed fast image segmentation method based on super-pixels not only runs very fast but also keep a relatively good segmentation accuracy as well.
Key words: applied mathematics; image segmentation; super-pixels; undirected weight graph; spectral clustering; competitive learning
发表期数: 2015年1月第1期
引用格式: 程再兴,刘双龙,马尽文. 基于超像素的快速图像分割[J]. 中国科技论文在线精品论文,2015,8(1):19-30.
 
2 评论数 0
暂无评论
友情链接