集美大学学报（自然科学版）2023，Vol.28Issue(5)：442-453,12.DOI:10.19715/j.jmuzr.2023.05.08
基于参考图像的超分辨率重建算法综述
A Review of Reference-Based Image Super-Resolution Reconstruction Algorithms
摘要
Abstract
Single image super-resolution reconstruction refers to the process of reconstructing high-resolu-tion images from low-resolution images,which is a typical ill-posed problem.Recently,a new reconstruction i-dea has emerged:Taking the existing high-resolution image as a reference,the rich texture in the reference im-age is used to compensate for the missing detail information of the low-resolution image,so as to alleviate the ill-posedness of single image super-resolution reconstruction.This new method is called reference-based image super-resolution reconstruction.Different from ordinary single image super-resolution reconstruction,the input of this method also needs one or more high-resolution reference images with similar content or texture to the low-resolution image.Although the related research on this reconstruction method has been carried out for a short period of time,many advanced reconstruction ideas and algorithms have emerged,and a comprehensive review is urgently needed.Therefore,on the basis of analyzing and drawing on relevant research results about a-lignment-based and patch-matching-based methods,this paper gives a general introduction to the new methods emerging in recent years,and discusses the future development direction.It is hoped that this will be a useful reference for subsequent research in this field.关键词
图像超分辨率重建/参考图/图像对齐/图像块匹配Key words
image super-resolution reconstruction/reference image/image alignment/image patch matching分类
信息技术与安全科学引用本文复制引用
张东晓,唐妮..基于参考图像的超分辨率重建算法综述[J].集美大学学报（自然科学版）,2023,28(5):442-453,12.基金项目
福建省自然科学基金项目(2020J01710) （2020J01710）
集美大学国家基金培育计划项目(ZP2020063) （ZP2020063）
