基于图像分解的图像修复技术-计算机工程.PDFVIP

36 10 2010 5 Vol.36 No.10 Computer Engineering May 2010 ·· 2010 A TP319.41 1 1 1 2 (1. 510640 2. 510640) (TV) TV TV Matlab Image Inpainting Technology Based on Image Decomposition 1 1 1 2 LIN Yun-li , ZHAO Jun-hong , ZHU Xue-feng , HU Yong-jian (1. College of Automation Science and Engineering, South China University of Technology, Guangzhou 510640; 2. School of Electronic and Information, South China University of Technology, Guangzhou 510640) AbstractAiming at the drawbacks of Total Variation(TV) model for image impainting, this paper proposes an improved inpainting model based on image decomposition. Using image decomposition technology, it extracts the structure information and texture information from the image. The improved TV model is applied to structure part of image to effectively avoid step effect in TV model on smooth region. On the process of iterative, it respectively analyzes the feature points and non-feature points to avoid the fuzziness of feature points during the inpainting. Image texture information uses improved exemplar-based inpainting technology. Matlab simulation experimental results show that the inpainting effect and the results of peak signal to noise ratio are better than that of original algorithm. Key wordsimage inpainting; Total Variation(TV) model; image decomposition; structure information; texture information 1 Criminisi [1] (PDE) 2 TV [2]