Enhanced Super-Resolution Training via Mimicked Alignment for Real-World Scenes

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Elezabi, Omar, Wu, Zongwei, Timofte, Radu
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914967040557056
author Elezabi, Omar
Wu, Zongwei
Timofte, Radu
author_facet Elezabi, Omar
Wu, Zongwei
Timofte, Radu
contents Image super-resolution methods have made significant strides with deep learning techniques and ample training data. However, they face challenges due to inherent misalignment between low-resolution (LR) and high-resolution (HR) pairs in real-world datasets. In this study, we propose a novel plug-and-play module designed to mitigate these misalignment issues by aligning LR inputs with HR images during training. Specifically, our approach involves mimicking a novel LR sample that aligns with HR while preserving the degradation characteristics of the original LR samples. This module seamlessly integrates with any SR model, enhancing robustness against misalignment. Importantly, it can be easily removed during inference, therefore without introducing any parameters on the conventional SR models. We comprehensively evaluate our method on synthetic and real-world datasets, demonstrating its effectiveness across a spectrum of SR models, including traditional CNNs and state-of-the-art Transformers. The source codes will be publicly made available at https://github.com/omarAlezaby/Mimicked_Ali .
format Preprint
id arxiv_https___arxiv_org_abs_2410_05410
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhanced Super-Resolution Training via Mimicked Alignment for Real-World Scenes
Elezabi, Omar
Wu, Zongwei
Timofte, Radu
Computer Vision and Pattern Recognition
Image and Video Processing
Image super-resolution methods have made significant strides with deep learning techniques and ample training data. However, they face challenges due to inherent misalignment between low-resolution (LR) and high-resolution (HR) pairs in real-world datasets. In this study, we propose a novel plug-and-play module designed to mitigate these misalignment issues by aligning LR inputs with HR images during training. Specifically, our approach involves mimicking a novel LR sample that aligns with HR while preserving the degradation characteristics of the original LR samples. This module seamlessly integrates with any SR model, enhancing robustness against misalignment. Importantly, it can be easily removed during inference, therefore without introducing any parameters on the conventional SR models. We comprehensively evaluate our method on synthetic and real-world datasets, demonstrating its effectiveness across a spectrum of SR models, including traditional CNNs and state-of-the-art Transformers. The source codes will be publicly made available at https://github.com/omarAlezaby/Mimicked_Ali .
title Enhanced Super-Resolution Training via Mimicked Alignment for Real-World Scenes
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2410.05410