*M.E. 1st year Computer Engineering,
**Associate Professor,
In this study, a new image inpainting approach using multiscale salient structure propagation is proposed. The proposed approach consists of four stages, namely, (1) detection of salient structure(s), (2) inpainting of salient structure(s), (3) inpainting of surrounding areas of salient structure(s) by modified ant colony optimization (ACO), and (4) inpainting of remaining missing regions. Based on the experimental results obtained in this study, as compared with four comparison approaches, the proposed approach provides the better image inpainting results. Image inpainting can be roughly defined as: for an image containing missing regions, we want to fill the missing regions so that a visually plausible outcome is obtained. Image inpainting has developed very quickly and becomes one of the important contents of Digital Image Processing. Image inpainting can be employed in various applications, such as repairing aged images and multimedia editing. It is highly valuable in the protection of culture relic, movie and video special effect making.
Image inpainting, multiscale salient structure propagation, patch priority computation, ant colony optimization (ACO), pheromone update rule