深度学习图像超分辨率重建:关键技术解析与前沿进展  

Deep Learning Image Super-Resolution Reconstruction:Key Technology Analysis and Frontier Progress

作　　者：于会昌 刘士远 YU Huichang;LIU Shiyuan(School of Health Sciences and Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;Radiodiagnostic Department,Affiliated Changzheng Hospital,Naval Medical University,Shanghai 200003,China)

机构地区：[1]上海理工大学健康科学与工程学院,上海200093 [2]海军军医大学附属长征医院放射诊断科,上海200003

出　　处：《软件导刊》2025年第2期211-220,共10页Software Guide

基　　金：国家自然科学基金项目(81930049)。

摘　　要：图像超分辨率重建是计算机视觉领域的一个重要研究方向,主要目标是从低分辨率图像中恢复高分辨率图像、提升图像质量和清晰度,使图像在视觉感知和信息提取方面更具可用性。为此,全面总结了基于深度学习的图像超分辨率重建技术,对已有模型进行分析和比较。首先,概述图像超分辨率重建的背景;其次,详细回顾了典型深度学习模型及其原理;再次,分析现有研究不足,并概述当前的重点研究,介绍了研究者通过预训练策略、轻量级网络设计、优化创新等方式,在降低计算成本、增强模型适应性等方面取得的成绩;最后,对未来研究趋势进行总结展望。以期为研究深度学习图像超分辨率重建领域的学者提参考与借鉴。Image super-resolution reconstruction is an important research direction in the field of computer vision,with the main goal of restoring high-resolution images from low resolution images,improving image quality and clarity,and making images more usable in visual perception and information extraction.To this end,a comprehensive summary of image super-resolution reconstruction techniques based on deep learning was conducted,and existing models were analyzed and compared.Firstly,provide an overview of the background of image super-resolution reconstruction;Secondly,a detailed review was conducted on typical deep learning models and their principles;Next,analyze the shortcomings of existing research and outline the current key research.Introduce the achievements of researchers in reducing computational costs and enhancing model adaptability through pre training strategies,lightweight network design,optimization innovation,and other methods;Finally,a summary and outlook on future research trends are provided,in order to provide reference and inspiration for scholars in the field of deep learning image super-resolution reconstruction.

关 键 词：深度学习 超分辨率重建 神经网络 生成对抗网络 注意力机制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]