Multi-image Interpolation based on Graph-Cuts and Symmetric Optic Flow

Abstract

Multi-image interpolation in space and time has recently received considerable attention. Typically, the interpolated image is synthesized by adaptively blending several forward-warped images. Blending itself is a low-pass filtering operation: the interpolated images are prone to blurring, even if correspondences are perfect. Furthermore, ghosting artifacts appear as soon as the underlying correspondence fields are imperfect. We address both issues and propose a multi-image interpolation algorithm that avoids blending. Instead, we cast multi-image interpolation as a labeling problem and decide for each pixel in the synthesized view from which input image to sample. Combined with a symmetrical long-range optical flow formulation for correspondence field estimation, our approach yields crisp interpolated images without ghosting artifacts.