Unsupervised Flow-Aligned Sequence-to-Sequence Learning for Video Restoration

TL;DR

This paper introduces an unsupervised flow-aligned sequence-to-sequence model for video restoration that leverages optical flow estimation and sequence modeling to improve tasks like deblurring and super-resolution.

Contribution

It pioneers the use of sequence-to-sequence models in video restoration and integrates unsupervised optical flow estimation to enhance frame correspondence.

Findings

Outperforms existing methods in multiple VR tasks

Effectively captures long-range dependencies among frames

Improves optical flow accuracy with unsupervised distillation

Abstract

How to properly model the inter-frame relation within the video sequence is an important but unsolved challenge for video restoration (VR). In this work, we propose an unsupervised flow-aligned sequence-to-sequence model (S2SVR) to address this problem. On the one hand, the sequence-to-sequence model, which has proven capable of sequence modeling in the field of natural language processing, is explored for the first time in VR. Optimized serialization modeling shows potential in capturing long-range dependencies among frames. On the other hand, we equip the sequence-to-sequence model with an unsupervised optical flow estimator to maximize its potential. The flow estimator is trained with our proposed unsupervised distillation loss, which can alleviate the data discrepancy and inaccurate degraded optical flow issues of previous flow-based methods. With reliable optical flow, we can establish accurate correspondence among multiple frames, narrowing the domain difference between 1D language and 2D misaligned frames and improving the potential of the sequence-to-sequence model. S2SVR shows superior performance in multiple VR tasks, including video deblurring, video super-resolution, and compressed video quality enhancement. Code and models are publicly available at https://github.com/linjing7/VR-Baseline