CVPR2020
FeatureFlow: Robust Video Interpolation via Structure-to-Texture Generation
Shurui Gui, Chaoyue Wang, Qihua Chen, Dacheng Tao
Abstract
Overlayed inputs Texture refinement Ground-truth Structure-guided interpolation Overlayed inputs Texture refinement Ground-truth Structure-guided interpolation Figure 1: Examples of the proposed structure-to-texture generation for video interpolation. The whole framework splits the video interpolation task into two stages: structure-guided interpolation and texture refinement. The first row gives a challenging interpolation example of complicated dynamic scenes. The second row is a typical example to explain.