Divergence-Based Adaptive Extreme Video Completion
This addresses video reconstruction challenges for applications like cheap sampling, but it is incremental as it builds on existing image completion methods.
The paper tackles extreme video completion with only 1% of pixels retained, extending a state-of-the-art image method to video by using a color KL-divergence approach for color-motion estimation and adaptive filtering, achieving validated results on 50 videos with PSNR and mean opinion scores.
Extreme image or video completion, where, for instance, we only retain 1% of pixels in random locations, allows for very cheap sampling in terms of the required pre-processing. The consequence is, however, a reconstruction that is challenging for humans and inpainting algorithms alike. We propose an extension of a state-of-the-art extreme image completion algorithm to extreme video completion. We analyze a color-motion estimation approach based on color KL-divergence that is suitable for extremely sparse scenarios. Our algorithm leverages the estimate to adapt between its spatial and temporal filtering when reconstructing the sparse randomly-sampled video. We validate our results on 50 publicly-available videos using reconstruction PSNR and mean opinion scores.