Robust Reference-based Super-Resolution via C2-Matching

TL;DR

This paper introduces C2-Matching, a novel method for reference-based super-resolution that explicitly addresses transformation and resolution gaps, improving texture transfer accuracy and robustness, validated on new and standard datasets.

Contribution

We propose C2-Matching, a new approach that explicitly handles transformation and resolution gaps in reference-based super-resolution, along with a new dataset for realistic evaluation.

Findings

Outperforms state-of-the-art by over 1dB on CUFED5

Demonstrates robustness to large scale and rotation transformations

Shows strong generalization on the WR-SR dataset

Abstract

Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g. scale and rotation) and the resolution gap (e.g. HR and LR). To tackle these challenges, we propose C2-Matching in this work, which produces explicit robust matching crossing transformation and resolution. 1) For the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) For the resolution gap, we adopt a teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue. In addition, to faithfully evaluate the performance of Ref-SR under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. Extensive experiments demonstrate that our proposed C2-Matching significantly outperforms state of the arts by over 1dB on the standard CUFED5 benchmark. Notably, it also shows great generalizability on WR-SR dataset as well as robustness across large scale and rotation transformations.