# Towards Unified Keyframe Propagation Models

arXiv: https://arxiv.org/abs/2205.09731 · HTML: https://mlsanity.com/paper/2205.09731/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 31 of its 41 references that are in the map.
- First posted on arXiv 19 May 2022.

## Builds on (most cited first, 30 of 31)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Taming Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2012.09841/) (2020)
- [RAFT: Recurrent All-Pairs Field Transforms for Optical Flow](https://mlsanity.com/paper/2003.12039/) (2020)
- [Rethinking Attention with Performers](https://mlsanity.com/paper/2009.14794/) (2020)
- [Linformer: Self-Attention with Linear Complexity](https://mlsanity.com/paper/2006.04768/) (2020)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [Transformers in Vision: A Survey](https://mlsanity.com/paper/2101.01169/) (2021)
- [Resolution-robust Large Mask Inpainting with Fourier Convolutions](https://mlsanity.com/paper/2109.07161/) (2021)
- [Free-Form Image Inpainting with Gated Convolution](https://mlsanity.com/paper/1806.03589/) (2018)
- [Video Object Segmentation using Space-Time Memory Networks](https://mlsanity.com/paper/1904.00607/) (2019)
- [How Do Vision Transformers Work?](https://mlsanity.com/paper/2202.06709/) (2022)
- [Generating Images with Sparse Representations](https://mlsanity.com/paper/2103.03841/) (2021)
- [One-Shot Video Object Segmentation](https://mlsanity.com/paper/1611.05198/) (2016)
- [Large Scale Image Completion via Co-Modulated Generative Adversarial Networks](https://mlsanity.com/paper/2103.10428/) (2021)
- [Rethinking Space-Time Networks with Improved Memory Coverage for Efficient Video Object Segmentation](https://mlsanity.com/paper/2106.05210/) (2021)
- [EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning](https://mlsanity.com/paper/1901.00212/) (2019)
- [Learning Joint Spatial-Temporal Transformations for Video Inpainting](https://mlsanity.com/paper/2007.10247/) (2020)
- [Real-Time User-Guided Image Colorization with Learned Deep Priors](https://mlsanity.com/paper/1705.02999/) (2017)
- [Deep Flow-Guided Video Inpainting](https://mlsanity.com/paper/1905.02884/) (2019)
- [Towards An End-to-End Framework for Flow-Guided Video Inpainting](https://mlsanity.com/paper/2204.02663/) (2022)
- [Deep Video Inpainting](https://mlsanity.com/paper/1905.01639/) (2019)
- [Flow-edge Guided Video Completion](https://mlsanity.com/paper/2009.01835/) (2020)
- [Aggregated Contextual Transformations for High-Resolution Image Inpainting](https://mlsanity.com/paper/2104.01431/) (2021)
- [Copy-and-Paste Networks for Deep Video Inpainting](https://mlsanity.com/paper/1908.11587/) (2019)
- [TransFill: Reference-guided Image Inpainting by Merging Multiple Color and Spatial Transformations](https://mlsanity.com/paper/2103.15982/) (2021)
- [Onion-Peel Networks for Deep Video Completion](https://mlsanity.com/paper/1908.08718/) (2019)
- [Deep Video Color Propagation](https://mlsanity.com/paper/1808.03232/) (2018)
- [Region-wise Generative Adversarial ImageInpainting for Large Missing Areas](https://mlsanity.com/paper/1909.12507/) (2019)

## Abstract (arXiv)

Many video editing tasks such as rotoscoping or object removal require the propagation of context across frames. While transformers and other attention-based approaches that aggregate features globally have demonstrated great success at propagating object masks from keyframes to the whole video, they struggle to propagate high-frequency details such as textures faithfully. We hypothesize that this is due to an inherent bias of global attention towards low-frequency features. To overcome this limitation, we present a two-stream approach, where high-frequency features interact locally and low-frequency features interact globally. The global interaction stream remains robust in difficult situations such as large camera motions, where explicit alignment fails. The local interaction stream propagates high-frequency details through deformable feature aggregation and, informed by the global interaction stream, learns to detect and correct errors of the deformation field. We evaluate our two-stream approach for inpainting tasks, where experiments show that it improves both the propagation of features within a single frame as required for image inpainting, as well as their propagation from keyframes to target frames. Applied to video inpainting, our approach leads to 44% and 26% improvements in FID and LPIPS scores. Code at https://github.com/runwayml/guided-inpainting