摘要
针对块匹配图像修复算法容易出现结构不连贯与块效应问题,提出一种两阶段图像修复方法.该方法通过增加相邻修复块在重叠区域的关联性约束,利用图像块关联匹配修复算法实现对降采样受损图像的粗修复,以保证图像主体结构的完整性;依据不同分辨率图像中像素点与其8邻域的线性相关性,引入带噪低秩矩阵填充算法对图像细节进行超分辨率精修复,使修复后图像的纹理与色彩变化具有连续性.实验结果表明,文中方法对包含渐变特征的自然图像修复具有更好的视觉效果.
A novel two stage method of image completion is proposed to handle structure discontinuities andblock artifacts caused by traditional exemplar-based image completion algorithms. With the relevance of adjacentpatches taken into consideration, a patch associated matching restoration algorithm, which is able to fix mainstructure in image, is performed on the down-sampling image. Meanwhile, according to the linear correlation ofpixels and their eight neighbors in different resolution images, a super-resolution method based on noisy low-rankmatrix filling is proposed and employed to refine the target image. This operation ensures the continuity of thechanges in the detail information of texture and color. Experimental results demonstrate that the proposed methodhas more outstanding performance on natural images with gradual features.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2015年第2期271-278,共8页
Journal of Computer-Aided Design & Computer Graphics
基金
中央高校基本科研业务费专项资金(13XS01)
关键词
图像修复
块关联匹配
低秩矩阵
超分辨率
image completion
patch associated matching
low-rank matrix
super resolution
