J4 ›› 2013, Vol. 40 ›› Issue (5): 86-91.doi: 10.3969/j.issn.1001-2400.2013.05.014

• 研究论文 • 上一篇    下一篇



  1. (1. 西安电子科技大学 理学院,陕西 西安  710071;
    2. 西安工业大学 理学院,陕西 西安  710012)
  • 收稿日期:2013-01-13 出版日期:2013-10-20 发布日期:2013-11-27
  • 通讯作者: 张文娟
  • 作者简介:张文娟(1980-),女,讲师,西安电子科技大学博士研究生,E-mail:girl-zwj@163.com.
  • 基金资助:

    国家自然科学基金资助项目(61271294, 60872138, 61105011, 11101292)

Image super-pixels segmentation method based on the  non-convex low-rank and sparse constraints

ZHANG Wenjuan1,2;FENG Xiangchu1   

  1. (1. School of Science, Xidian Univ., Xi'an  710071, China;
    2. School of Science, Xi'an Technological Univ., Xi'an  710012, China)
  • Received:2013-01-13 Online:2013-10-20 Published:2013-11-27
  • Contact: ZHANG Wenjuan


将图像超像素分割看作子空间聚类问题.给出一个约束条件,等价于以干净数据为字典.利用系数矩阵的非凸迫近p范数作为稀疏约束,利用系数矩阵奇异值的非凸迫近p范数作为低秩约束,建立非凸极小化模型.运用增广拉格朗日方法和交替极小化方法给出数值计算方法.数值实验表明,笔者提出的约束条件下的分割效果优于原始数据作为字典; 非凸迫近p范数的分割效果优于凸的核范数和l1范数.

关键词: 图像分割, 超像素, 稀疏, 低秩, 非凸


Image super-pixels segmentation is considered as the subspace clustering problem. A new constraint condition is presented to be equivalent to using the clean data as the dictionary. The non-convex proximal p-norm of the coefficients matrix is used for the sparse constraint, and, the non-convex proximal p-norm of the singular values of the coefficients matrix is used for the low-rank constraint. Then a non-convex minimization model is proposed. The augmented Lagrangian method and the AM (alternating minimization) method are applied for solving the unknown matrices. The results of numerical experiments show that the constraint condition presented in this paper is better than using the original data as the dictionary, and that the non-convex proximal p-norm has a better segmentation result than the convex nuclear norm and l1 norm.

Key words: image segmentation, super-pixels, sparse, low-rank, non-convex