A Scale Consistent Approach to Image Completion∗

Most patch based algorithms for completing missing parts of images fill in the absent regions by copying patches from the known part of the image into the unknown part, somewhat like plastic surgery. The criterion for deciding which patch to copy is compatibility of the copied patch with the vicinity of the region being completed. In this paper we propose introducing a new dimension to this compatibility criterion, namely, scale. The patch is thus chosen by evaluating its consistency with respect to a hierarchy of smoothed (less detailed) versions of the image, as well as its surroundings in the current version. Applied recursively, this approach results in a multi-scale framework that is shown to yield a dramatic improvement in the robustness of patch based image completion.