Flow-Guided Video Inpainting with Scene Templates

We consider the problem of filling in missing spatiotemporal regions of a video. We provide a novel flow-based solution by introducing a generative model of images in relation to the scene (without missing regions) and mappings from the scene to images. We use the model to jointly infer the scene template, a 2D representation of the scene, and the mappings. This ensures consistency of the frame-to-frame flows generated to the underlying scene, reducing geometric distortions in flow based inpainting. The template is mapped to the missing regions in the video by a new (L

-L

) interpolation scheme, creating crisp inpaintings and reducing common blur and distortion artifacts. We show on two benchmark datasets that our approach out-performs state-of-the-art quantitatively and in user studies.

Citation

Lao, D., Zhu, P., Wonka, P., & Sundaramoorthi, G. (2021). Flow-Guided Video Inpainting with Scene Templates. 2021 IEEE/CVF International Conference on Computer Vision (ICCV). https://doi.org/10.1109/iccv48922.2021.01433

Version History

You are currently viewing version 2 of the item.

Now showing 1 - 2 of 2

Version	Date	Summary
2*	2022-03-08 07:38:32	Conference Paper
1	2021-09-01 06:23:18

* Selected version