计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (26): 180-182.DOI: 10.3778/j.issn.1002-8331.2008.26.055

• 图形、图像、模式识别 • 上一篇    下一篇

一种基于整体变分的图像修补算法及其应用

田艳艳,祝 轩,彭进业,郭彦伶   

  1. 西北大学 信息科学与技术学院,西安 710127
  • 收稿日期:2007-11-07 修回日期:2008-02-27 出版日期:2008-09-11 发布日期:2008-09-11
  • 通讯作者: 田艳艳

Implementation for digital image inpainting based on total variation

TIAN Yan-yan,ZHU Xuan,PENG Jin-ye,GUO Yan-ling   

  1. School of Information Science and Technology,Northwest University,Xi’an 710127,China
  • Received:2007-11-07 Revised:2008-02-27 Online:2008-09-11 Published:2008-09-11
  • Contact: TIAN Yan-yan

摘要: 针对污损数字图像修复这一问题,重点考虑了图像的两个重要特征:梯度和等照度线的曲率,探讨了两者对图像修补效果及速度的影响。从而将两者同时纳入图像修补模型,提出了一种新的基于整体变分的图像修补算法。实验表明,相比于CDD算法,该算法能够在短时间内较好地修复污损区域,且能将断裂的水平线连接得更为光滑、自然。

Abstract: In order to restore destroyed images,this paper mainly observes the two important characters,the gradient and the curvature of the isophotes,and discusses their function in inpainting process.Furthermore,a novel algorithm is introduced for digital inpainting of destroyed images based on total variation.The results show that in contrast with previous approaches,CDD for example,the new model restores the information of the destroyed area better and connects the broken level sets smoother and natural.