东北大学学报（自然科学版）

东北大学学报(自然科学版) ›› 2008, Vol. 29 ›› Issue (11): 1540-1543.DOI: -

• 论著 • 上一篇    下一篇

基于Pixon的规格化切与图像分割

崔华;高立群;   

  1. 东北大学信息科学与工程学院;
  • 收稿日期:2013-06-22 修回日期:2013-06-22 出版日期:2008-11-15 发布日期:2013-06-22
  • 通讯作者: Cui, H.
  • 作者简介:-
  • 基金资助:
    国家自然科学基金资助项目(60475036);;

Pixon-based normalized cuts and image segmentation

Cui, Hua (1); Gao, Li-Qun (1)   

  1. (1) School of Information Science and Engineering, Northeastern University, Shenyang 110004, China
  • Received:2013-06-22 Revised:2013-06-22 Online:2008-11-15 Published:2013-06-22
  • Contact: Cui, H.
  • About author:-
  • Supported by:
    -

RichHTML

1

PDF (PC)

461

摘要： 针对规格化切过程需要求解大规模矩阵的特征向量,使得这一图像分割方法需要很大的运算量的缺点,将基于Pixon的图像模型与规格化切相结合,提出了先获取基于Pixon的图像模型,然后再使用规格化切对基于Pixon的图像模型进行递归分割的图像分割方法.实验结果表明,该方法显著降低了使用规格化切进行图像分割时的运算量和内存消耗,提高了该方法的效率和实用性.

关键词: 图像分割, 图切, 规格化切, 各向异性扩散, Pixon

Abstract: Normalized cut, as an image segmentation technique, requires a great deal of operation since an eigenvector is needed to solve the large-scale matrix in operation process. To get rid of such a trouble, a new approach is proposed combining the normalized cut with a Pixon-based image model, i.e., the model should be obtained first, then the normalized cut technique is applied to the model so as to segment images recursively. Experimental results showed that the approach proposed can reduce greatly the operational quantity and memory consumption as using the normalized cut singly in image segmentation, thus improving the efficiency and practicability.

中图分类号: