计算机应用 ›› 2011, Vol. 31 ›› Issue (02): 362-365.

• 图形图像处理 • 上一篇    下一篇

新的彩色图像分层修复方法

周姗姗1,赵知劲2,叶学义2   

  1. 1. 杭州电子科技大学
    2.
  • 收稿日期:2010-07-21 修回日期:2010-09-11 发布日期:2011-02-01 出版日期:2011-02-01
  • 通讯作者: 周姗姗
  • 基金资助:
    国家自然科学青年基金;浙江省科技计划重点项目

New color image multi-level-inpainting method

  • Received:2010-07-21 Revised:2010-09-11 Online:2011-02-01 Published:2011-02-01
  • Contact: SHAN SHANZHOU

摘要: 基于全变分的彩色图像修复(CTV)模型,可以在保持边缘的同时达到去噪目的,但对纹理较复杂的区域修复效果不理想。利用分层修复的思想在对整幅彩色图像进行CTV修复后,再对有明显修复痕迹的局部待修复区域使用基于块的纹理合成算法。实验结果表明,这种分层修复的方法在彩色图像的结构修复和纹理修复两方面实现较好的统一,既能很好地降噪,又能保持彩色图像的清晰边界,而且在较大区域图像修复上表现出良好的效果。

关键词: 彩色图像修复, CTV去噪模型, 分层修复方法, 纹理合成

Abstract: The Color Total Variation (CTV) model can denoise while maintaining the edge of the damaged images, but it should be improved in domains of complicated texture. In this paper, a new image multilevelinpainting method for color images was proposed. The main research topic is to improve the visual result of inpainted images of complicated texture. Damaged images were calculated on the basis of CTV model, and then the patchbased texture synthesis method was used to improve the inpainting results of complicated texture domain. The experimental results show that the inpainting performance is better both in the structure and the texture of color images. Moreover, the new image multilevelinpainting method can be used in the image inpainting with bigger damaged domains.

Key words: color image inpainting, CTV denoising model, multi-level-inpainting method, texture synthesis