NAFSSR: Stereo Image Super-Resolution Using NAFNet

TL;DR

This paper introduces NAFSSR, a stereo image super-resolution method based on NAFNet with added cross attention for view fusion, achieving state-of-the-art results and winning a major challenge.

Contribution

The paper extends NAFNet with cross attention modules for stereo super-resolution and proposes training/testing strategies, setting new benchmarks in the field.

Findings

NAFSSR outperforms existing methods on multiple datasets.

NAFSSR achieved 1st place in NTIRE 2022 challenge.

The approach effectively exploits binocular information for super-resolution.

Abstract

Stereo image super-resolution aims at enhancing the quality of super-resolution results by utilizing the complementary information provided by binocular systems. To obtain reasonable performance, most methods focus on finely designing modules, loss functions, and etc. to exploit information from another viewpoint. This has the side effect of increasing system complexity, making it difficult for researchers to evaluate new ideas and compare methods. This paper inherits a strong and simple image restoration model, NAFNet, for single-view feature extraction and extends it by adding cross attention modules to fuse features between views to adapt to binocular scenarios. The proposed baseline for stereo image super-resolution is noted as NAFSSR. Furthermore, training/testing strategies are proposed to fully exploit the performance of NAFSSR. Extensive experiments demonstrate the effectiveness of our method. In particular, NAFSSR outperforms the state-of-the-art methods on the KITTI 2012, KITTI 2015, Middlebury, and Flickr1024 datasets. With NAFSSR, we won 1st place in the NTIRE 2022 Stereo Image Super-resolution Challenge. Codes and models will be released at https://github.com/megvii-research/NAFNet.