Interactive Image Completion with Direction Empirical Mode

Abstract

An interactive image completion method is proposed based on Direction Empirical Mode Decomposition (DEMD). Blocked area or area with loss of information in a target image is completed with DEMD combined with texture synthesis in an interactive way. The target image is decomposed by DEMD into levels of Intrinsic Mode Functions (IMF) images, while the user is allowed to indicate the structural image edge to recover in the unknown image regions. For each level of IMF image, first the target image patches along user-specified curves in the unknown region are completed. Then the remaining target image patches are prioritized to complete according to image gradient feature. The target image patches are completed based on the combination frequency feature values from DEMD and the texture synthesis. Finally, levels of completed IMF images are compounded into the result image. Experiments prove that the proposed algorithm is not only able to recover the damaged structural and texture information of large scale, but also completely blocked image structures.