J4 ›› 2010, Vol. 37 ›› Issue (5): 825-828.doi: 10.3969/j.issn.1001-2400.2010.05.009

• 研究论文 • 上一篇    下一篇



  1. (1. 西安电子科技大学 数学科学系,陕西 西安  710071;
    2. 河南科技大学 数学与统计学院,河南 洛阳  471003)
  • 收稿日期:2010-10-23 出版日期:2010-10-20 发布日期:2010-10-11
  • 通讯作者: 郝岩
  • 作者简介:郝岩(1977-),女,西安电子科技大学博士研究生,E-mail: haoyan_@126.com.
  • 基金资助:


Image inpainting model of nonlocal diffusion

HAO Yan1;FENG Xiang-chu1;XU Jian-lou1,2   

  1. (1. Dept. of Mathematical Sciences, Xidian Univ., Xi'an  710071, China;
    2. Dept. of Mathematics and Statistics, Henan Univ. of Sci. and Tech., Luoyang  471003, China)
  • Received:2010-10-23 Online:2010-10-20 Published:2010-10-11
  • Contact: HAO Yan


利用非局部算子的定义,提出了一种新的非局部扩散的修复模型.该模型在扩散的过程中充分利用了图像的全局信息对损坏的区域进行修复,克服了偏微分方程模型在修复图像时产生的模糊和不能很好地保护图像边缘及纹理的不足. 实验结果表明,该模型是行之有效的.

关键词: 图像修复, 扩散, 纹理, 非局部算子


We present a new inpainting model of nonlinear diffusion using the definitions of nonlocal operators. This model can inpaint those damaged domains by making use of the image global information. Furthermore, it can overcome the shortcomings that PDE can lead to some blur in diffusion procession and cannot preserve the edges of the image better. Experimental results show that this new model is feasible.

Key words: image inpainting, diffusion, texture, nonlocal operators