DAVIS and FVI datasets
Video inpainting aims to fill spatio-temporal corrupted regions with plausible content. To achieve this goal, it is necessary to find correspondences from neighboring frames to faithfully hallucinate the unknown content.
BibTex:
