- 0
- 0
- 约1.5万字
- 约 3页
- 2019-04-12 发布于广西
- 举报
2014 年1月15 日 现代电子技术 Jan. 2014
第37 卷第2 期 Modern Electronics Technique Vol. 37 No. 2
45
45
基于像素相似性与Graph⁃Cut 的图像自动分割
胡 焦,陈 辉
(安徽理工大学,安徽 淮南 232000 )
摘 要:采用Graph cut 的图像分割分为交互式分割与自动分割,为了弥补交互式分割需要用户的参与、分割结果依赖种
子点的选取与模型的建立、需要用户的进一步修正才能得到满意结果等不足,介绍基于相似性的自动图像分割。该方法采用
Mean⁃shift 平滑图像、转换成YCbCr 空间后将图像分割成若干区域块等方法,建立合适的能量函数后利用最大流/ 最小割定理
求最小割。并通过实验证明,该方法在自动性、准确性、时间效率方面都获得较好结果。
关键字:交互式分割;自动分割分割;Graph cut ;像素相似性
中图分类号:TN919⁃34 文献标识码:A 文章编号:1004⁃373X (2014 )02⁃0045⁃03
Image automatic segmentation based on pixel similarity and Graph cut
HU Jiao ,CHEN Hui
(Anhui University of Science and Technology ,Huainan 232000 ,China )
Abstract :The image segmentation based on Graph cut can be classified into interactive segmentation and automatic segmenta⁃
tion. In order to make up the shortages of interactive segmentation which needs interaction with users ,segmentation result which
relies on selection of the seed points and the model establishment ,and further revision done by users ,an automatic image seg⁃
mentation based on pixel similarity is introduced in this paper. Mean ⁃shift is used in this method to smooth the image. The image
is divided into several blocks as the image is converted into YcbCr. The proper energy function is establ
原创力文档

文档评论(0)