Computing the optical flow in sequences with missing parts

From LRDE

The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Abstract

Computing optical flow can be a first step towards video inpainting. In such cases, one has to deal with video sequences where some parts, the ones to be inpainted, are missing. On the other hand, the optical flow can be computed either locally or globally. Global methods generally show better results. In the case of sequences with missing areas, the computation with global methods is not straightforward due to the lack of information in some regions. In this technical report, we introduce a method combining both global and local algorithms in order to compute the optical flow in such sequences, leading to an efficient and simple way to compute video inpainting.