Inpainting based on total variation

Inpainting, the technique of modifying an image in an undetectable form, is as ancient as art itself. With the release of the new image compression standard JPEG2000, people pay more attention to the wavelet inpainting problems. We consider the problems of filling in missing or damaged coefficients in both spatial and wavelet domain. First, we propose an algorithm in order to restore long and narrow blank automatically in spatial domain. It introduces different weights into the algorithm according to features of image and we can achieve better inpainting results with much less computation time. Second, we present another new algorithm for the wavelet based inpainting problems. The algorithm is simple and has good recovery effect even if as high as 90% wavelet coefficients have been lost. Moreover, convergence rate is enormous speed up. ? 2007 IEEE.EI