纹理和边缘特征相结合的图像修复算法  

An Image Inpainting Algorithm Combined with Texture and Edge Features

作　　者：任澍[1] 唐向宏[1,2] 康佳伦[1] 

机构地区：[1]亢州电子科技大学通信工程学院,杭州310018 [2]杭州电子科技大学信息工程学院,杭州310018

出　　处：《计算机辅助设计与图形学学报》2013年第11期1682-1693,共12页Journal of Computer-Aided Design & Computer Graphics

摘　　要：针对Criminisi算法在图像修复过程中易将纹理部分误认为边缘部分,从而造成修复顺序偏差而影响修复效果的不足,以及由于在已知区域中不存在合适匹配块而造成的修复效果失真问题,提出一种基于纹理和边缘特征相结合的图像修复算法.首先利用图像的纹理特征和边缘特征引入差别因子,改进优先权模型来增强对边缘部分的辨别能力,以克服修复过程中易造成纹理延伸的不足;通过改进原算法的搜索方式,在提高算法效率的同时去除修复过程中的瑕疵点;然后对修复过程中所产生的不良匹配区域进行定位,采用基于图像分解的图像修复算法进行修复.实验结果证明,改进算法较好地克服了Criminisi算法存在的缺点,能获得较好的视觉效果.Since Criminisi algorithm is easy to mistakenly treat texture parts as edge parts, which disorders the inpainting sequence and leads to poor inpainting effect. Besides, it also produces artifacts because of lacking appropriate match pieces in the known area. This paper proposes a new image inpainting algorithm based on the combination of texture and edge features. First, the proposed algorithm introduces differential factor by making use of the texture and edge features to improve the ability of distinguishing texture parts from edge parts, which overcomes the shortcoming of texture extension. Then it uses a new search approach to avoid flaws and improve the efficiency. At last, it makes use of image inpainting algorithm based on image decomposition to inpaint the area of bad match pieces. Experiment results show that the proposed algorithm overcomes the shortcomings of Criminisi algorithm and achieves better visual effect.

关 键 词：图像块差别因子 优先权模型 不良匹配块 图像分解 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]