基于深度学习的图像去模糊方法研究概述  

A Survey on Deep Learning Based Image Deblurring

作　　者：秦婧[1] QIN Jing(Suzhou Chien-Shiung Institute of Technology,Suzhou Jiangsu 215400)

机构地区：[1]苏州健雄职业技术学院,江苏苏州215400

出　　处：《软件》2024年第12期168-170,共3页Software

摘　　要：随着深度学习的迅猛发展,基于深度学习的图像去模糊取得了重大进展与突破。本文围绕基于深度学习的图像去模糊技术展开,概述了图像去模糊技术的背景,包括其定义、数学模型及发展历程;阐述了基于深度学习的不同图像去模糊算法,如编码—解码网络模型、生成对抗网络模型、多尺度网络模型、级联网络模型和融合物理退化模型等;探讨了图像去模糊技术的发展趋势。With the superior performance achieved by deep learning in computer vision,image restoration also focuses on the methods based on deep neural networks.This paper focuses on image deblurring technology based on deep learning,summarizes the background of image deblurring technology,including its definition,mathematical model and development history;expounds different image deblurring algorithms based on deep learning,such as the encoder and the decoder,coding-decoding network model,generative adversarial network model,multi-scale network model,cascade network model and fusion physical degradation model.Finally,this paper discusses the current technical challenges faced by image deblurring research,and makes some prospects for future development.

关 键 词：图像去模糊 深度学习 网络模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]