摘要
研究了传统整体变分去噪算法和图像修补算法,提出了一种基于轮廓-纹理分解的图像修补算法.算法首先将待修补图像分解为轮廓结构图像和纹理细节图像.再对轮廓结构图像的空缺进行轮廓结构修补,并对纹理细节图像的空缺进行纹理合成.最后将修补后的轮廓结构图像及纹理合成后的纹理细节图像进行合成,得到需要的修补图像.这种方法能够很好地修补图像空缺部分的轮廓结构及纹理细节.实验结果表明,该算法比纯结构图像修补法或纯纹理合成法要好.
By studying the principle and application of total variation de - noising model and image inpainting,an image first deco inpainting method based on contour - texture decomposition is proposed. In this method, an image is mposed into two parts： contour and texture image, which are inpainted respectively. Finally, the two inpainted images will be synthesized into one image. It is proved through the results that this method can better inpaint images.
出处
《昆明理工大学学报（理工版）》
2008年第3期57-61,共5页
Journal of Kunming University of Science and Technology(Natural Science Edition)
基金
云南省教育厅自然科学基金
云南民族大学青年基金科研项目资助(项目编号:6Y0006D,200610)
关键词
图像修补
图像分解
纹理合成
image inpainting
image decomposition
texture synthesis
