超分辨率图像重建算法综述

doi:10.3778/j.issn.1673-9418.2111126

摘要/Abstract

摘要：

在人类视觉感知系统中,高分辨率（HR）图像是图像清晰表达其空间结构、细节特征、边缘纹理等信息的重要媒介,在医学、刑侦、卫星等领域有着极为广泛的实用价值。超分辨率图像重建（SRIR）旨在从给定的低分辨率（LR）图像中,重建含有清晰细节特征的高分辨率图像,是计算机视觉和图像处理领域中的一项重点研究任务。首先,对超分辨率图像重建的概念和数学模型进行阐述,并对图像重建方法进行系统分类,将其系统地分为基于插值、基于重构、基于学习（深度学习前、后）三类超分辨率图像重建方法;其次,对三类方法中典型的、常用的、最新的算法及其研究进行全面回顾和综述,并从网络结构、学习机制、适用场景、优势和局限性等方面对所列的图像重建算法进行了梳理;然后,归纳总结了超分辨率图像重建算法所用的数据集和图像质量评价指标,重点比较基于深度学习的各种超分辨率图像重建算法的特点与性能;最后,从四方面对超分辨率图像重建问题未来的研究方向或角度进行展望。

关键词: 图像处理, 超分辨率重建, 深度学习, 图像质量评估

Abstract:

In human visual perception system, high-resolution (HR) image is an important medium to clearly express its spatial structure, detailed features, edge texture and other information, and it has a very wide range of practical value in medicine, criminal investigation, satellite and other fields. Super-resolution image reconstruction (SRIR) is a key research task in the field of computer vision and image processing, which aims to reconstruct a high-resolution image with clear details from a given low-resolution (LR) image. In this paper, the concept and mathematical model of super-resolution image reconstruction are firstly described, and the image reconstruction methods are systematically classified into three kinds of super-resolution image reconstruction methods：based on interpolation, based on reconstruction, based on learning (before and after deep learning). Secondly, the typical, commonly used and latest algorithms among the three methods and their research are comprehensively reviewed and summarized, and the listed image reconstruction algorithms are combed from the aspects of network structure, learning mechanism, application scenarios, advantages and limitations. Then, the datasets and image quality evaluation indices used for super-resolution image reconstruction algorithms are summarized, and the characteristics and performance of various super-resolution image reconstruction algorithms based on deep learning are compared. Finally, the future research direction or angle of super-resolution image reconstruction is prospected from four aspects.

Key words: image processing, super-resolution reconstruction, deep learning, image quality assessment

中图分类号:

TP391.41

钟梦圆, 姜麟. 超分辨率图像重建算法综述[J]. 计算机科学与探索, 2022, 16(5): 972-990.

ZHONG Mengyuan, JIANG Lin. Review of Super-Resolution Image Reconstruction Algorithms[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(5): 972-990.

图/表 14

图1 超分辨率图像重建技术图解

Fig.1 Illustration of super-resolution image reconstruction technology

图2 图像重建方法分类

Fig.2 Classification of image reconstruction algorithms

表1 基于插值的图像重建算法比较

Table 1 Comparison of image reconstruction algorithms based on interpolation

算法	原理	运算复杂度	运算速度	算法灵活性	图像质量
最近邻域插值法	线性插值	低	快	强	差
双线性插值法	线性插值	较低	较快	较强	较差
双三次线性插值法	线性插值	中	慢	弱	一般
边缘导向插值法	非线性插值	中	慢	较强	高
梯度引导插值法	非线性插值	高	慢	较弱	中
小波变换插值法	非线性插值	高	较慢	中	高

表2 基于重构的图像重建算法比较

Table 2 Comparison of image reconstruction algorithms based on reconstruction

算法	先验信息	可行解	运算复杂度	运算速度	算法灵活性	图像质量
频域法	依赖性弱	唯一	低	慢	较差	差
非均匀内插法	依赖性强	唯一	较低	较慢	差	中
迭代反投影法	依赖性较强	不唯一	中	中	差	中
凸集投影法	依赖性较弱	不唯一	高	较慢	较强	较高
最大后验概率法	依赖性较弱	唯一	较高	较快	较强	较高
MAP/POCS法	依赖性弱	唯一	中	快	强	高

图3 基于学习的图像重建算法思想

Fig.3 Idea of image reconstruction algorithm based on learning

图4 稀疏表示法

Fig.4 Sparse representation

图5 深度学习背景下的图像重建网络结构基本图

Fig.5 Image reconstruction network structure diagram under background of deep learning

表3 深度学习背景下的部分网络结构

Table 3 Partial structure of network under deep learning background

类型	作用	代表算法
残差学习	提高网络收敛速度,学习丰富的复杂特征等	VDSR^[56]、DRCN^[67]、RDN^[71]、EDSR^[78]等
递归学习	缓解梯度爆炸或消失,多路径递归学习等	DRCN^[67]、DRRN^[73]、SRFBN^[74]、CDC^[89]等
密集连接	加强不同层之间的图像传播、利用、融合等	SRFBN^[74]、SRDenseNet^[75]、ESRGAN^[79]等
跳跃连接	增强层间联系以及特征信息流的传递等	DRCN^[67]、SRDenseNet^[75]、EDSR^[78]等
注意力机制	标定图像重点与非重点学习重建区域,充分挖掘层内特征信息等	RCAN^[63]、CBAM^[64]、SMSR^[83]、CDC^[89]、SA-SR-GAN^[91]、HAN^[92]等
连续记忆机制	全局性图像特征融合,连续记忆性传递低频、高频特征信息等	RDN^[71]、DRRN^[73]、SRFBN^[74]、SRGAN^[77]等
反馈机制	共享权重值,确保高级信息与低级信息间的表达与交流等	SRFBN^[74]、SRGAN^[77]、ESRGAN^[79]、SRFeat^[84]等

图6 SRCNN网络结构

Fig.6 Network structure of SRCNN

图7 SRGAN网络结构

Fig.7 Network structure of SRGAN

表4 深度学习后的各算法特点对比

Table 4 Comparison of features of each algorithm after deep learning

算法	上采样方式	上采样方法	残差学习	递归学习	密集连接	注意力机制	VGG网络	损失
SRCNN^[53]	预上采样	双三次插值	—	—	—	—	—	$L 2$ 损失
FSRCNN^[54]	后上采样	反卷积	—	—	—	—	—	$L 2$ 损失
FSRCNN*^[55]	后上采样	反卷积	—	—	—	—	—	$L 2$ 损失
VDSR^[56]	后上采样	双三次插值	√	—	—	—	—	MSE损失
ESPCN^[57]	预上采样	亚像素卷积	—	—	—	—	—	$L 2$ 损失
RCAN^[63]	后上采样	亚像素卷积	√	—	—	√	—	$L 1$ 损失
DRCN^[67]	预上采样	双三次插值	√	√	—	—	—	$L 2$ 损失
RDN^[71]	后上采样	亚像素卷积	—	√	√	—	—	$L 1$ 损失
DRRN^[73]	预上采样	双三次插值	√	√	—	—	—	$L 2$ 损失
SRFBN^[74]	后上采样	反卷积	√	√	√	—	—	$L 1$ 损失
SRDenseNet^[75]	后上采样	亚像素卷积	√	—	—	—	—	$L 2$ 损失
SRGAN^[77]	后上采样	亚像素卷积	√	—	—	—	—	对抗+内容损失
EDSR^[78]	后上采样	亚像素卷积	√	—	—	—	—	$L 1$ 损失
ESRGAN^[79]	后上采样	亚像素卷积	—	√	√	—	—	$L 1$ 损失
RFB-ESRGAN^[80]	后上采样	亚像素卷积	√	√	√	—	—	$L 1$ 损失
NatSR^[81]	后上采样	反卷积	√	—	√	—	√	重建+对抗+自然度损失
SMSR^[83]	后上采样	双三次插值	√	—	√	√	—	稀疏正则化损失
SRFeat^[84]	逐步上采样	亚像素卷积	√	—	√	—	—	感知+图像损失
LatticeNet^[85]	后上采样	反卷积	√	—	—	√	—	MAE损失
IRN^[88]	逐步上采样	可逆双射变换	—	—	√	—	—	LR制导+重构+匹配+感知损失
CDC^[89]	逐步上采样	反卷积	√	√	—	√	—	梯度加权损失
HAN^[92]	后上采样	亚像素卷积	√	—	—	√	—	$L 1$ 损失
MASA-SR^[93]	迭代上采样	反卷积	√	—	—	—	—	重建+感知+对抗损失
CF-Net^[94]	迭代上采样	反卷积	√	—	—	—	—	MSSIM损失
Class-SR^[96]	后上采样	反卷积	—	—	—	—	—	$L 1$ 损失+平均损失+类损失
LIIF^[97]	后上采样	亚像素卷积	√	—	—	—	—	$L 1$ 损失
SRwarp^[98]	自适应重采样	双三次插值	—	—	—	—	—	$L 1$ 损失
Meta-SR^[99]	后上采样	Meta Upscale	—	√	√	—	—	$L 1$ 损失

表4 深度学习后的各算法特点对比

Table 4 Comparison of features of each algorithm after deep learning

算法	上采样方式	上采样方法	残差学习	递归学习	密集连接	注意力机制	VGG网络	损失
SRCNN^[53]	预上采样	双三次插值	—	—	—	—	—	$L 2$ 损失
FSRCNN^[54]	后上采样	反卷积	—	—	—	—	—	$L 2$ 损失
FSRCNN*^[55]	后上采样	反卷积	—	—	—	—	—	$L 2$ 损失
VDSR^[56]	后上采样	双三次插值	√	—	—	—	—	MSE损失
ESPCN^[57]	预上采样	亚像素卷积	—	—	—	—	—	$L 2$ 损失
RCAN^[63]	后上采样	亚像素卷积	√	—	—	√	—	$L 1$ 损失
DRCN^[67]	预上采样	双三次插值	√	√	—	—	—	$L 2$ 损失
RDN^[71]	后上采样	亚像素卷积	—	√	√	—	—	$L 1$ 损失
DRRN^[73]	预上采样	双三次插值	√	√	—	—	—	$L 2$ 损失
SRFBN^[74]	后上采样	反卷积	√	√	√	—	—	$L 1$ 损失
SRDenseNet^[75]	后上采样	亚像素卷积	√	—	—	—	—	$L 2$ 损失
SRGAN^[77]	后上采样	亚像素卷积	√	—	—	—	—	对抗+内容损失
EDSR^[78]	后上采样	亚像素卷积	√	—	—	—	—	$L 1$ 损失
ESRGAN^[79]	后上采样	亚像素卷积	—	√	√	—	—	$L 1$ 损失
RFB-ESRGAN^[80]	后上采样	亚像素卷积	√	√	√	—	—	$L 1$ 损失
NatSR^[81]	后上采样	反卷积	√	—	√	—	√	重建+对抗+自然度损失
SMSR^[83]	后上采样	双三次插值	√	—	√	√	—	稀疏正则化损失
SRFeat^[84]	逐步上采样	亚像素卷积	√	—	√	—	—	感知+图像损失
LatticeNet^[85]	后上采样	反卷积	√	—	—	√	—	MAE损失
IRN^[88]	逐步上采样	可逆双射变换	—	—	√	—	—	LR制导+重构+匹配+感知损失
CDC^[89]	逐步上采样	反卷积	√	√	—	√	—	梯度加权损失
HAN^[92]	后上采样	亚像素卷积	√	—	—	√	—	$L 1$ 损失
MASA-SR^[93]	迭代上采样	反卷积	√	—	—	—	—	重建+感知+对抗损失
CF-Net^[94]	迭代上采样	反卷积	√	—	—	—	—	MSSIM损失
Class-SR^[96]	后上采样	反卷积	—	—	—	—	—	$L 1$ 损失+平均损失+类损失
LIIF^[97]	后上采样	亚像素卷积	√	—	—	—	—	$L 1$ 损失
SRwarp^[98]	自适应重采样	双三次插值	—	—	—	—	—	$L 1$ 损失
Meta-SR^[99]	后上采样	Meta Upscale	—	√	√	—	—	$L 1$ 损失

表5 超分辨率图像重建的开源数据集

Table 5 Open source dataset for super-resolution image reconstruction

数据集	格式	张数	大小	内容	用途	来源
Set5	PNG	5	1.6 MB	baby、bird、butterfly、head、woman	测试	2012BMVC
Set14	PNG	14	4.2 MB	人、动植物、漫画、幻灯片等	测试	2014CVPR
BSD100	PNG	100	142.0 MB	人、动植物、建筑、自然景观、环境等	测试	—
BSD300	JPG	300	93.0 MB	人、动植物、食物、建筑、自然景观等	训练/验证	2001ICCV
BSD500	JPG	500	155.0 MB	人、动植物、食物、建筑、自然景观等	训练/验证	2011IEEE
DIV2K2017	PNG	900	3.7 GB	人、手工制品、环境、风景等	训练/验证	2017CVPR
Urban100	PNG	100	143.0 MB	建筑	测试	2015CVPR
Manga109	PNG	109	—	漫画图	测试	2019ICML
SunHay80	PNG	80	77.0 MB	建筑、自然景观	测试	—
91-Image	PNG	91	13.8 MB	人、植物、汽车等	训练	2010IEEE
Flickr2K	PNG	2 650	10.8 GB	人、动植物、建筑、自然景观等	训练	2017CVPR
Real SR	JPG	1 352	11.7 GB	相机照片	训练/验证	2019CVPR
Waterloo	PNG	99 624	1.3 GB	人、动植物、景观、交通等	训练	2017IEEE
Outdoor-Scenes	PNG	337 189	4.4 GB	动植物、建筑、山水、天空等	训练/测试	2018CVPR
PIRM	PNG	200	59.8 MB	人、物、环境、植物、自然风景等	验证/测试	2018ECCV
W2S	PNG	144 000	1.9 GB	显微镜图像	训练/测试	2020ECCV
PIPAL	PNG	250	82.0 MB	人、动植物、建筑等	测试	2020ECCV
L20	PNG	20	55.9 MB	人、动植物、建筑、景观等	测试	2016CVPR
General-100	BMP	100	31.0 MB	人、动植物、日用品、食物等	测试	2016ECCV

表6 影响不同图像重建效果的客观因素

Table 6 Objective factors affecting different image reconstruction effects

类别	特点	过程	适用场景	优势	局限性
全参考图像	HR图像（真值图像）	$H R → L R → H R$	相机图像、光学图像、医学图像等	高适用性和灵活性	训练时间长、经济成本高
半参考图像	LR+HR （真值+失真图像）	$H R → L R → H R$ 或 $L R → H R$	遥感图像、医学图像、交通监控图像、SAR图像等	高对比性与参考性	图像差异较大,影响图像质量评价
盲参考图像	LR图像（失真图像）	$L R → H R$	遥感图像、医学图像、交通监控图像、SAR图像等	图像来源不受限制	图像质量评价受限

表6 影响不同图像重建效果的客观因素

Table 6 Objective factors affecting different image reconstruction effects

类别	特点	过程	适用场景	优势	局限性
全参考图像	HR图像（真值图像）	$H R → L R → H R$	相机图像、光学图像、医学图像等	高适用性和灵活性	训练时间长、经济成本高
半参考图像	LR+HR （真值+失真图像）	$H R → L R → H R$ 或 $L R → H R$	遥感图像、医学图像、交通监控图像、SAR图像等	高对比性与参考性	图像差异较大,影响图像质量评价
盲参考图像	LR图像（失真图像）	$L R → H R$	遥感图像、医学图像、交通监控图像、SAR图像等	图像来源不受限制	图像质量评价受限

表7 深度学习后各算法的重建效果对比

Table 7 Comparison of reconstruction effects of each algorithm after deep learning

算法	大小	Set5		Set14		Urban100		BSD100		Manga109		DIV2K
算法	大小	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
SRCNN^[53]	$× 2$	36.7	0.95	32.5	0.91	29.5	0.89	31.4	0.89	35.6	0.97	—	—
FSRCNN^[54]		37.1	0.96	32.7	0.91	29.9	0.90	31.5	0.89	36.7	0.97	—	—
FSRCNN*^[55]		37.3	0.96	32.8	0.91	30.2	0.90	—	—	—	—	—	—
VDSR^[56]		37.5	0.96	33.0	0.91	30.8	0.91	31.9	0.90	—	—	37.2	0.98
RCAN^[63]		38.3	0.96	34.2	0.92	33.5	0.94	37.2	0.97	39.6	0.98	—	—
DRCN^[67]		37.6	0.96	33.0	0.91	30.8	0.91	—	—	—	—	—	—
RDN^[71]		38.3	0.96	34.1	0.92	33.1	0.94	27.7	0.74	39.4	0.98	—	—
DRRN^[73]		37.7	0.96	33.2	0.91	31.2	0.92	32.1	0.90	—	—	—	—
SRFBN^[74]		38.2	0.96	33.9	0.92	32.8	0.93	—	—	39.3	0.98	—	—
SRGAN^[77]		38.9	0.97	35.6	0.94	35.2	0.96	33.8	0.93	38.3	0.96	34.4	0.97
EDSR^[78]		38.2	0.96	34.0	0.92	33.1	0.94	27.7	0.74	39.1	0.98	35.1	0.97
ESRGAN^[79]		38.2	0.96	34.0	0.92	33.1	0.94	32.4	0.90	—	—	36.1	0.96
SMSR^[83]		38.0	0.96	33.6	0.91	32.2	0.93	32.2	0.90	38.8	0.98	—	—
SRFeat^[84]		32.3	0.89	28.7	0.78	—	—	27.6	0.74	—	—	—	—
IRN^[88]		44.0	0.98	40.8	0.97	39.9	0.98	41.3	0.98	—	—	44.3	0.99
HAN^[92]		38.3	0.96	34.2	0.92	33.5	0.93	32.5	0.90	39.6	0.98	—	—
Meta-SR^[99]		34.0	0.92	32.3	0.90	—	—	—	—	39.2	0.98	35.2	0.95
SRCNN^[53]	$× 3$	32.8	0.91	29.3	0.82	26.2	0.80	28.4	0.79	30.5	0.91	—	—
FSRCNN^[54]		33.2	0.91	29.4	0.82	26.4	0.81	28.5	0.79	31.1	0.92	—	—
FSRCNN*^[55]		33.3	0.92	29.6	0.83	26.7	0.82	—	—	—	—	—	—
VDSR^[56]		33.7	0.92	29.8	0.83	27.1	0.83	36.7	0.97	32.0	0.93	—	—
RCAN^[63]		34.9	0.93	30.8	0.85	29.3	0.87	29.3	0.81	34.8	0.95	—	—
DRCN^[67]		33.8	0.90	29.8	0.83	27.2	0.83	—	—	—	—	—	—
RDN^[71]		34.9	0.93	30.7	0.85	29.0	0.87	27.7	0.74	33.4	0.94	—	—
DRRN^[73]		34.0	0.92	30.0	0.83	27.5	0.84	29.0	0.80	—	—	—	—
SRFBN^[74]		34.8	0.93	30.6	0.85	28.9	0.87	—	—	34.4	0.95	—	—
SRGAN^[77]		33.7	0.93	30.2	0.87	26.9	0.84	29.6	0.84	31.0	0.94	30.9	0.93
EDSR^[78]		34.8	0.93	30.7	0.85	29.0	0.87	29.3	0.81	34.1	0.95	31.4	0.93
ESRGAN^[79]		36.2	0.95	32.7	0.90	31.4	0.92	31.6	0.88	—	—	35.1	0.93
SMSR^[83]		34.4	0.93	30.3	0.84	28.3	0.86	29.1	0.81	33.7	0.94	—	—
HAN^[92]		34.9	0.93	30.8	0.85	29.3	0.87	29.4	0.81	34.8	0.95	—	—
Meta-SR^[99]		30.6	0.85	29.3	0.81	—	—	—	—	34.1	0.95	31.4	0.89
SRCNN^[53]	$× 4$	30.5	0.86	27.5	0.75	24.5	0.72	26.9	0.71	27.6	0.86	—	—
FSRCNN^[54]		30.7	0.87	27.6	0.76	24.6	0.73	27.0	0.72	27.9	0.86	—	—
FSRCNN*^[55]		31.0	0.88	27.8	0.76	27.1	0.72	—	—	—	—	—	—
VDSR^[56]		31.4	0.88	28.0	0.77	25.2	0.75	27.9	0.86	28.8	0.89	—	—
RCAN^[63]		32.7	0.90	29.0	0.79	27.1	0.81	28.8	0.89	31.7	0.92	—	—
DRCN^[67]		31.5	0.89	28.0	0.77	25.1	0.75	—	—	—	—	—	—
RDN^[71]		32.6	0.90	28.9	0.79	26.8	0.81	27.7	0.74	31.4	0.92	—	—
DRRN^[73]		31.7	0.89	28.2	0.77	25.4	0.76	27.4	0.73	—	—	—	—
SRFBN^[74]		32.6	0.90	28.9	0.79	26.7	0.80	—	—	31.4	0.92	—	—
SRGAN^[77]		33.9	0.91	30.3	0.84	29.3	0.87	29.2	0.80	32.8	0.88	28.9	0.90
EDSR^[78]		32.5	0.90	28.8	0.79	26.6	0.80	27.7	0.74	31.0	0.91	29.4	0.90
ESRGAN^[79]		32.7	0.90	29.0	0.79	27.8	0.75	27.0	0.82	—	—	30.9	0.85
NatSR^[81]		31.0	0.86	27.4	0.73	25.5	0.76	26.4	0.68	—	—	—	—
SMSR^[83]		32.1	0.89	28.6	0.78	26.1	0.79	27.6	0.74	30.5	0.91	—	—
SRFeat^[84]		32.3	0.89	28.7	0.78	—	—	27.6	0.74	—	—	—	—
IRN^[88]		36.2	0.95	32.7	0.90	31.4	0.92	31.6	0.88	—	—	35.1	0.93
HAN^[92]		32.8	0.90	29.0	0.79	27.0	0.81	27.9	0.75	31.7	0.92	—	—
Meta-SR^[99]		28.8	0.79	27.8	0.74	—	—	—	—	31.0	0.92	29.4	0.85

表7 深度学习后各算法的重建效果对比

Table 7 Comparison of reconstruction effects of each algorithm after deep learning

算法	大小	Set5		Set14		Urban100		BSD100		Manga109		DIV2K
算法	大小	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
SRCNN^[53]	$× 2$	36.7	0.95	32.5	0.91	29.5	0.89	31.4	0.89	35.6	0.97	—	—
FSRCNN^[54]		37.1	0.96	32.7	0.91	29.9	0.90	31.5	0.89	36.7	0.97	—	—
FSRCNN*^[55]		37.3	0.96	32.8	0.91	30.2	0.90	—	—	—	—	—	—
VDSR^[56]		37.5	0.96	33.0	0.91	30.8	0.91	31.9	0.90	—	—	37.2	0.98
RCAN^[63]		38.3	0.96	34.2	0.92	33.5	0.94	37.2	0.97	39.6	0.98	—	—
DRCN^[67]		37.6	0.96	33.0	0.91	30.8	0.91	—	—	—	—	—	—
RDN^[71]		38.3	0.96	34.1	0.92	33.1	0.94	27.7	0.74	39.4	0.98	—	—
DRRN^[73]		37.7	0.96	33.2	0.91	31.2	0.92	32.1	0.90	—	—	—	—
SRFBN^[74]		38.2	0.96	33.9	0.92	32.8	0.93	—	—	39.3	0.98	—	—
SRGAN^[77]		38.9	0.97	35.6	0.94	35.2	0.96	33.8	0.93	38.3	0.96	34.4	0.97
EDSR^[78]		38.2	0.96	34.0	0.92	33.1	0.94	27.7	0.74	39.1	0.98	35.1	0.97
ESRGAN^[79]		38.2	0.96	34.0	0.92	33.1	0.94	32.4	0.90	—	—	36.1	0.96
SMSR^[83]		38.0	0.96	33.6	0.91	32.2	0.93	32.2	0.90	38.8	0.98	—	—
SRFeat^[84]		32.3	0.89	28.7	0.78	—	—	27.6	0.74	—	—	—	—
IRN^[88]		44.0	0.98	40.8	0.97	39.9	0.98	41.3	0.98	—	—	44.3	0.99
HAN^[92]		38.3	0.96	34.2	0.92	33.5	0.93	32.5	0.90	39.6	0.98	—	—
Meta-SR^[99]		34.0	0.92	32.3	0.90	—	—	—	—	39.2	0.98	35.2	0.95
SRCNN^[53]	$× 3$	32.8	0.91	29.3	0.82	26.2	0.80	28.4	0.79	30.5	0.91	—	—
FSRCNN^[54]		33.2	0.91	29.4	0.82	26.4	0.81	28.5	0.79	31.1	0.92	—	—
FSRCNN*^[55]		33.3	0.92	29.6	0.83	26.7	0.82	—	—	—	—	—	—
VDSR^[56]		33.7	0.92	29.8	0.83	27.1	0.83	36.7	0.97	32.0	0.93	—	—
RCAN^[63]		34.9	0.93	30.8	0.85	29.3	0.87	29.3	0.81	34.8	0.95	—	—
DRCN^[67]		33.8	0.90	29.8	0.83	27.2	0.83	—	—	—	—	—	—
RDN^[71]		34.9	0.93	30.7	0.85	29.0	0.87	27.7	0.74	33.4	0.94	—	—
DRRN^[73]		34.0	0.92	30.0	0.83	27.5	0.84	29.0	0.80	—	—	—	—
SRFBN^[74]		34.8	0.93	30.6	0.85	28.9	0.87	—	—	34.4	0.95	—	—
SRGAN^[77]		33.7	0.93	30.2	0.87	26.9	0.84	29.6	0.84	31.0	0.94	30.9	0.93
EDSR^[78]		34.8	0.93	30.7	0.85	29.0	0.87	29.3	0.81	34.1	0.95	31.4	0.93
ESRGAN^[79]		36.2	0.95	32.7	0.90	31.4	0.92	31.6	0.88	—	—	35.1	0.93
SMSR^[83]		34.4	0.93	30.3	0.84	28.3	0.86	29.1	0.81	33.7	0.94	—	—
HAN^[92]		34.9	0.93	30.8	0.85	29.3	0.87	29.4	0.81	34.8	0.95	—	—
Meta-SR^[99]		30.6	0.85	29.3	0.81	—	—	—	—	34.1	0.95	31.4	0.89
SRCNN^[53]	$× 4$	30.5	0.86	27.5	0.75	24.5	0.72	26.9	0.71	27.6	0.86	—	—
FSRCNN^[54]		30.7	0.87	27.6	0.76	24.6	0.73	27.0	0.72	27.9	0.86	—	—
FSRCNN*^[55]		31.0	0.88	27.8	0.76	27.1	0.72	—	—	—	—	—	—
VDSR^[56]		31.4	0.88	28.0	0.77	25.2	0.75	27.9	0.86	28.8	0.89	—	—
RCAN^[63]		32.7	0.90	29.0	0.79	27.1	0.81	28.8	0.89	31.7	0.92	—	—
DRCN^[67]		31.5	0.89	28.0	0.77	25.1	0.75	—	—	—	—	—	—
RDN^[71]		32.6	0.90	28.9	0.79	26.8	0.81	27.7	0.74	31.4	0.92	—	—
DRRN^[73]		31.7	0.89	28.2	0.77	25.4	0.76	27.4	0.73	—	—	—	—
SRFBN^[74]		32.6	0.90	28.9	0.79	26.7	0.80	—	—	31.4	0.92	—	—
SRGAN^[77]		33.9	0.91	30.3	0.84	29.3	0.87	29.2	0.80	32.8	0.88	28.9	0.90
EDSR^[78]		32.5	0.90	28.8	0.79	26.6	0.80	27.7	0.74	31.0	0.91	29.4	0.90
ESRGAN^[79]		32.7	0.90	29.0	0.79	27.8	0.75	27.0	0.82	—	—	30.9	0.85
NatSR^[81]		31.0	0.86	27.4	0.73	25.5	0.76	26.4	0.68	—	—	—	—
SMSR^[83]		32.1	0.89	28.6	0.78	26.1	0.79	27.6	0.74	30.5	0.91	—	—
SRFeat^[84]		32.3	0.89	28.7	0.78	—	—	27.6	0.74	—	—	—	—
IRN^[88]		36.2	0.95	32.7	0.90	31.4	0.92	31.6	0.88	—	—	35.1	0.93
HAN^[92]		32.8	0.90	29.0	0.79	27.0	0.81	27.9	0.75	31.7	0.92	—	—
Meta-SR^[99]		28.8	0.79	27.8	0.74	—	—	—	—	31.0	0.92	29.4	0.85

参考文献 110

[1]	PARK S C, PARK M K, KANG M G. Super-resolution image reconstruction: a technical overview[J]. IEEE Signal Processing Magazine, 2003, 20(3): 21-36. DOI URL
[2]	TSAI R, HUANG T. Multi-frame image restoration and regi-stration[J]. Computer Vision and Image Processing, 1984, 1(2): 317-339.
[3]	钟宝江, 陆志芳, 季家欢. 图像插值技术综述[J]. 数据采集与处理, 2016, 31(6): 1083-1096.
	ZHONG B J, LU Z F, JI J H. A review of image interpolation techniques[J]. Data Acquisition and Processing, 2016, 31(6): 1083-1096.
[4]	马书红. 几种经典插值算法的放大结果比较[J]. 信息通信, 2014(5): 34.
	MA S H. Comparison of magnification results of several classical interpolation algorithms[J]. Information and Communication, 2014(5): 34.
[5]	郑璐, 王保云, 杨昆, 等. 基于深度学习的超分辨率图像重建研究综述[J]. 计算机与数字工程, 2021, 49(4): 804-808.
	ZHENG L, WANG B Y, YANG K, et al. A review of research on super resolution image reconstruction based on deep learning[J]. Computer and Digital Engineering, 2021, 49(4): 804-808.
[6]	陈剑涛. 基于深度学习的图像超分辨率重建算法[J]. 现代计算机, 2020(30): 52-55.
	CHEN J T. Image super resolution algorithm based on deep learning[J]. Modern Computers, 2020(30): 52-55.
[7]	李赛赛. 基于深度学习的图像超分辨率重建研究[J]. 电脑知识与技术, 2020, 16(29): 191-192.
	LI S S. Research on image super-resolution reconstruction based on deep learning[J]. Computer Knowledge & Technology, 2020, 16(29): 191-192.
[8]	王子扬, 章义来. 图像超分辨率重建SRGAN算法的探讨[J]. 福建电脑, 2020, 36(5): 38-40.
	WANG Z Y, ZHANG Y L. Discussion on SRGAN algorithm of image super-resolution reconstruction[J]. Fujian Computer, 2020, 36(5): 38-40.
[9]	李佳星, 赵勇先, 王京华. 基于深度学习的单幅图像超分辨率重建算法综述[J]. 自动化学报, 2021, 47(10): 2341-2363.
	LI J X, ZHAO Y X, WANG J H. A review of single image super-resolution reconstruction algorithms based on deep lear-ning[J]. Acta Automatica Sinica, 2021, 47(10): 2341-2363.
[10]	龙超. 图像超分辨率重建算法综述[J]. 科技视界, 2015(13): 88-89.
	LONG C. A review of image super-resolution reconstruction algorithms[J]. Science and Technology Vision, 2015(13): 88-89.
[11]	王春霞, 苏红旗, 范郭亮. 图像超分辨率重建技术综述[J]. 计算机技术与发展, 2011, 21(5): 124-127.
	WANG C X, SU H Q, FAN G L. Review of image super-resolution reconstruction techniques[J]. Computer Technology and Development, 2011, 21(5): 124-127.
[12]	苏衡, 周杰, 张志浩. 超分辨率图像重建方法综述[J]. 自动化学报, 2013, 39(8): 1202-1213.
	SU H, ZHOU J, ZHANG Z H. Review of super resolution image reconstruction methods[J]. Acta Automatica Sinica, 2013, 39(8): 1202-1213. DOI URL
[13]	张芳, 赵东旭, 肖志涛, 等. 单幅图像超分辨率重建技术研究进展[J/OL]. 自动化学报(2021-11-29)[2021-12-20]. https://cnki.net/kcms/detail/11.2109.TP.20211129.1735.002.html.
	ZHANG F, ZHAO D X, XIAO Z T, et al. Research progress of single image super-resolution reconstruction technology[J/OL]. Acta Automatica Sinica (2021-11-29)[2021-12-20]. https://cnki.net/kcms/detail/11.2109.TP.20211129.1735.002.html.
[14]	董银丽, 任翠萍. 超分辨率图像重建解析[J]. 电脑知识与技术, 2020, 16(17): 188-190.
	DONG Y L, REN C P. Image reconstruction with super resolution[J]. Computer Knowledge and Technology, 2020, 16(17): 188-190.
[15]	LI K, YANG S, DONG R, et al. Survey of single image super-resolution reconstruction[J]. IET Image Processing, 2020, 14(11): 2273-2290. DOI URL
[16]	BLU T, THEVENAZ P, UNSER M. Linear interpolation revitalized[J]. IEEE Transactions on Image Processing, 2004, 13(5): 710-719. DOI URL
[17]	TAO H J, TAN G X, LIU J, et al. Super-resolution remote sensing image processing algorithm based on wavelet transform and interpolation[J]. Journal of Wuhan University of Technology, 2002, 24(8): 63-66.
[18]	NAYAK R, PATRA D. Image interpolation using adaptive P-spline[C]// Proceedings of the 2015 Annual IEEE India Conference, New Delhi, Dec 17-20, 2015. Piscataway: IEEE, 2015: 1-6.
[19]	KEYS R G. Cubic convolution interpolation for digital image processing[J]. IEEE Transactions on Acoustics, Speech, and Signal Processing, 1981, 29(6): 1153-1160. DOI URL
[20]	ZHOU D, SHEN X, DONG W. Image zooming using directional cubic convolution interpolation[J]. IET Image Processing, 2012, 6(6): 627-634. DOI URL
[21]	彭燕, 胡丹屏, 刘宇红, 等. 基于FPGA实现的Ferguson双三次曲面插值图像缩放算法[J]. 贵州大学学报(自然科学版), 2019, 36(6): 68-72.
	PENG Y, HU D P, LU Y H, et al. Implementation of Ferguson bicubic surface interpolation image scaling algorithm based on FPGA[J]. Journal of Guizhou University (Natural Science), 2019, 36(6): 68-72.
[22]	LI X, ORCHARD M T. New edge-directed interpolation[J]. IEEE Transactions on Image Processing, 2001, 10(10): 1521-1527. DOI URL
[23]	DAI S Y, HAN M, XU W, et al. Soft edge smoothness prior for alpha channel super resolution[C]// Proceedings of the 2007 IEEE Conference on Computer Vision and Pattern Classification, Minneapolis, Jun 17-22, 2007. Piscataway: IEEE, 2007: 1-8.
[24]	ZHANG X, WU X. Image interpolation by adaptive 2D auto-regressive modeling and soft-decision estimation[J]. IEEE Transactions on Image Processing, 2008, 17(6): 887-896. DOI URL
[25]	JING G M, CHOI Y K, WANG J Y, et al. Gradient guided image interpolation[C]// Proceedings of the 2014 IEEE International Conference on Aerospace Electronics and Remote Sensing Technology, Paris, Oct 27-30, 2014. Piscataway: IEEE, 2014: 1822-1826.
[26]	SUN J, XU Z, SHUM H. Image super-resolution using gradient profile prior[C]// Proceedings of the 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Alaska, Jun 24-26, 2008. Piscataway: IEEE, 2008: 1-8.
[27]	RASTI P, LUSI I, DEMIREL H, et al. Wavelet transform based new interpolation technique for satellite image resolution enhancement[C]// Proceedings of the 2014 IEEE International Conference on Aerospace Electronics and Remote Sensing Technology, Paris, Oct 27-30, 2014. Piscataway: IEEE, 2014: 185-188.
[28]	宋明煜, 张靖, 俞一彪. 多向滤波结合小波逆变换的图像超分辨率重建[J]. 现代信息科技, 2020, 4(15): 5-8.
	SONG M Y, ZHANG J, YU Y B. Multi-directional filtering combined with inverse wavelet transform for image super-resolution reconstruction[J]. Modern Information Technology, 2020, 4(15): 5-8.
[29]	FORD C, ETTER D M. Wavelet basis reconstruction of nonuniformly sampled data[J]. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, 1998, 45(8): 1165-1168. DOI URL
[30]	NGUYEN N, MILANFAR P. A wavelet-based interpolation-restoration method for super resolution[J]. Circuits, Systems & Signal Processing, 2000, 19(4): 321-338.
[31]	段立娟, 武春丽, 恩擎, 等. 基于小波域的深度残差网络图像超分辨率算法[J]. 软件学报, 2019, 30(4): 941-953.
	DUAN L J, WU C L, EN Q, et al. Deep residual network in wavelet domain for image super-resolution[J]. Journal of Software, 2019, 30(4): 941-953.
[32]	PATTI A J, SEZAN M I, TEKALP A M. Super-resolution video reconstruction with arbitrary sampling lattices and nonzero aperture time[J]. IEEE Transactions on Image Processing, 1997, 6(8): 1064-1076. DOI URL
[33]	NASONOV A V, KRYLOV A S. Fast super-resolution using weighted median filtering[C]// Proceedings of the 2010 20th International Conference on Pattern Recognition, Istanbul, Aug 23-26, 2010. Washington: IEEE Computer Society, 2010: 2230-2233.
[34]	IRANI M, PELEGM S. Improving resolution by image registration[J]. CVGIP: Graphical Models and Image Processing, 1991, 53(3): 231-239. DOI URL
[35]	STARK H, OSKOUI P. High-resolution image recovery from image-plane arrays, using convex projections[J]. Journal of the Optical Society of America A Optics & Image Science, 1989, 6(11): 1715-1726.
[36]	BANHAM M R, KATSAGGELOS A K. Digital image restoration[J]. IEEE Signal Processing Magazine, 1997, 14(2): 24-41.
[37]	SCHUITZ R R, STEVENSON R L. Extraction of high-resolution frames from video sequences[J]. IEEE Transactions on Image Processing, 1996, 5(6): 996-1011. DOI URL
[38]	陈光盛, 李树涛. MAP和POCS算法实现超分辨率图像的重建[J]. 科学技术与工程, 2006, 6(4): 396-399.
	CHEN G S, LI S T. Reconstruction of super-resolution image using MAP and POCS algorithms[J]. Science Technology and Engineering, 2006, 6(4): 396-399.
[39]	FREEMAN W T, PASZTOR E C, CARMICHAEL O T. Learning low level vision[J]. International Journal of Computer Vision, 2000, 40(1): 25-47. DOI URL
[40]	FREEMAN W T, JONES T R, PASZTOR E C. Example-based super-resolution[J]. IEEE Computer Graphics and Applications, 2002, 22(2): 56-65.
[41]	GLASNER D, BAGON S, IRANI M. Super-resolution from a single image[C]// Proceedings of the 2009 IEEE 12th International Conference on Computer Vision, Kyoto, Sep 27-Oct 4, 2009. Washington: IEEE Computer Society, 2009: 349-356.
[42]	CHANG H, YEUNG D Y, XIONG Y. Super-resolution through neighbor embedding[C]// Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Washington, Jun 27-29, 2004. Washington: IEEE Computer Society, 2004: 275-282.
[43]	王银玲, 王昕. 基于稀疏表示的遥感图像超分辨率重建[J]. 长春工业大学学报(自然科学版), 2021, 42(2): 175-180.
	WANG Y L, WANG X. Super-resolution reconstruction of sparse remote sensing images[J]. Journal of Changchun University of Technology (Natural Science Edition), 2021, 42(2): 175-180.
[44]	胡浩, 张建林, 李斌成, 等. 基于稀疏优化的图像压缩感知重建算法[J]. 半导体光电, 2021, 42(2): 269-274.
	HU H, ZHANG J L, LI B C, et al. Compressive sensing image reconstruction algorithm based on optimized sparse representation[J]. Semiconductor Optoelectronics, 2021, 42(2): 269-274.
[45]	YANG J C, WRIGNT J, HUANG T, et al. Image super-resolution via sparse representation[J]. IEEE Transactions on Image Processing, 2010, 19(11): 2861-2873. DOI URL
[46]	TIMOFTE R, DESMET V, VANGOOL L. A+: adjusted anchored neighborhood regression for fast super-resolution[C]// LNCS 9006: Proceedings of the 12th Asian Conference on Computer Vision, Singapore, Nov 1-5, 2014. Cham: Springer, 2014: 111-126.
[47]	LI J, WU J, DENG H, et al. A self-learning image super-resolution method via sparse representation and non-local similarity[J]. Neurocomputing, 2016, 184(5): 196-206. DOI URL
[48]	檀结庆, 蔡蒙琪, 朱星辰, 等. 基于局部结构相似与稀疏表示的超分辨率图像重建[J]. 合肥工业大学学报(自然科学版), 2021, 44(8): 1045-1050.
	TAN J Q, CAI M Q, ZHU X C, et al. Super-resolution image reconstruction based on local structure similarity and sparse representation[J]. Journal of Hefei University of Technology (Natural Science), 2021, 44(8): 1045-1050.
[49]	沈瑜, 刘成, 杨倩. 利用子空间稀疏特征的超分辨率图像重建算法[J]. 计算机工程与应用, 2021, 57(5): 173-182.
	SHEN Y, LIU C, YANG Q. A super-resolution image reconstruction algorithm using subspace sparse features[J]. Computer Engineering and Applications, 2021, 57(5): 173-182.
[50]	曾台英, 杜菲. 基于层次聚类的图像超分辨率重建[J]. 光学学报, 2018, 38(4): 130-137.
	ZENG T Y, DU F. Image super-resolution reconstruction based on hierarchical clustering[J]. Acta Optica Sinica, 2018, 38(4): 130-137.
[51]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]// Proceedings of the 26th Annual Conference on Neural Information Processing Systems 2012, Lake Tahoe, Dec 3-6, 2012. Red Hook: Curran Associates, 2012: 1097-1105.
[52]	GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]// Proceedings of the Annual Conference on Neural Information Processing Systems 2014, Montreal, Dec 8-13, 2014. Red Hook: Curran Associates, 2014: 2672-2680.
[53]	DONG C, LOY C C, HE K, et al. Learning a deep convolutional network for image super-resolution[C]// LNCS 8692: Proceedings of the 13th European Conference on Computer Vision, Zurich, Sep 6-12, 2014. Cham: Springer, 2014: 184-199.
[54]	DONG C, LOY C C, TANG X. Accelerating the super-resolution convolutional neural network[C]// LNCS 9906: Proceedings of the 14th European Conference on Computer Vision, Amsterdam, Oct 8-16, 2016. Cham: Springer, 2016: 391-407.
[55]	LEE W, LEE J, KIM D, et al. Learning with privileged information for efficient image super-resolution[C]// LNCS 12369: Proceedings of the 16th European Conference on Computer Vision, Glasgow, Aug 23-28, 2020. Cham: Springer, 2020: 465-482.
[56]	KIM J, LEE J K, LEE K M. Accurate image super-resolution using very deep convolutional networks[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 1646-1654.
[57]	SHI W, CABALLERO J, HUSZAR F, et al. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 1874-1883.
[58]	HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 770-778.
[59]	MONIZ J, PAL C. Convolutional residual memory networks[J]. arXiv:1606.05262v, 2016.
[60]	ZHANG K, SUN M, HAN T X, et al. Residual networks of residual networks: multilevel residual networks[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2017, 28(6): 1303-1314. DOI URL
[61]	李现国, 冯欣欣, 李建雄. 多尺度残差网络的单幅图像超分辨率重建[J]. 计算机工程与应用, 2021, 57(7): 215-221.
	LI X G, FENG X X, LI J X. Sigle image super-resolution reconstruction based on multi-scale residual network[J]. Computer Engineering and Applications, 2021, 57(7): 215-221.
[62]	SHA F, ZANDAVI S M, CHUNG Y Y. Fast deep parallel residual network for accurate super resolution image processing[J]. Expert Systems with Applications, 2019, 128(8): 157-168. DOI URL
[63]	ZHANG Y L, LI K P, LI K, et al. Image super-resolution using very deep residual channel attention networks[C]// LNCS 11211: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 294-310.
[64]	WOO S, PARK J, LEE J Y, et al. CBAM:convolutional block attention module[C]// LNCS 11211: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 3-19.
[65]	徐永兵, 袁东, 余大兵, 等. 多注意力机制引导的双目图像超分辨率重建算法[J]. 电子测量技术, 2021, 44(15): 103-108.
	XU Y B, YUAN D, YU D B, et al. Binocular image super-resolution reconstruction algorithm guided by multi-attention mechanism[J]. Electronic Measurement Technology, 2021, 44(15): 103-108.
[66]	卢正浩, 刘丛. 多尺度特征复用混合注意力网络的图像重建[J]. 中国图象图形学报, 2021, 26(11): 2645-2658.
	LU Z H, LIU C. Multiscale feature reuse mixed attention network for image reconstruction[J]. Journal of Image and Graphics, 2021, 26(11): 2645-2658.
[67]	KIM J, LEE J K, LEE K M. Deeply-recursive convolutional network for image super-resolution[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 1637-1645.
[68]	程德强, 郭昕, 陈亮亮, 等. 多通道递归残差网络的图像超分辨率重建[J]. 中国图象图形学报, 2021, 26(3): 605-618.
	CHENG D Q, GUO X, CHEN L L, et al. Image super-resolution reconstruction from multi-channel recursive resi-dual network[J]. Journal of Image and Graphics, 2021, 26(3): 605-618.
[69]	HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 4700-4708.
[70]	程玉, 郑华, 陈晓文, 等. 基于密集残差注意力网络的图像超分辨率算法[J]. 计算机系统应用, 2021, 30(1): 135-140.
	CHENG Y, ZHENG H, CHEN X W, et al. Image super-resolution algorithm based on residual dense attention networks[J]. Computer Systems & Applications, 2021, 30(1): 135-140.
[71]	ZHANG Y L, TIAN Y P, KONG Y, et al. Residual dense network for image super-resolution[C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, Jun 18-22, 2018. Washington: IEEE Computer Society, 2018: 2472-2481.
[72]	满开亮, 汪友生, 刘继荣. 基于稠密残差网络的图像超分辨率重建算法[J]. 图学学报, 2021, 42(4): 556-562.
	MAN K L, WANG Y S, LIU J R. Image super-resolution reconstruction algorithm based on dense residual network[J]. Journal of Graphics, 2021, 42(4): 556-562.
[73]	TAI Y, YANG J, LIU X M. Image super-resolution via deep recursive residual network[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2017: 2790-2798.
[74]	LI Z, YANG J L, LIU Z, et al. Feedback network for image super-resolution[C]// Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 3867-3876.
[75]	TONG T, LI G, LIU X J, et al. Image super-resolution using dense skip connections[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 4799-4807.
[76]	SONG D H, WANG Y H, CHEN H T, et al. AdderSR: towards energy efficient image super-resolution[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, Jun 21-24, 2021. Piscataway: IEEE, 2021: 15648-15657.
[77]	LEDIG C, THEIS L, HUSZAR F, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 105-114.
[78]	LIM B, SON S, KIM H, et al. Enhanced deep residual networks for single image super-resolution[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 1132-1140.
[79]	WANG X T, YU K, WU S X, et al. ESRGAN: enhanced super-resolution generative adversarial networks[C]// LNCS 11133: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 63-79.
[80]	SHANG T Z, DAI Q J, ZHU S C, et al. Perceptual extreme super-resolution network with receptive field block[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, Jun 16-21, 2020. Piscataway: IEEE, 2020: 1778-1787.
[81]	SOH J W, PARK G Y, JO J, et al. Natural and realistic single image super-resolution with explicit natural manifold discrimination[C]// Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 8122-8131.
[82]	VU T, LUU T M, YOO C D. Perception-enhanced image super-resolution via relativistic generative adversarial networks[C]// LNCS 11133: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 98-113.
[83]	WANG L G, DONG X Y, WANG Y G, et al. Exploring sparsity in image super-resolution for efficient inference[C]// Proceedings of the 2021 IEEE Conference on Computer Vision and Pattern Recognition, Nashville, Jun 21-24, 2021. Piscataway: IEEE, 2021: 4915-4924.
[84]	PARK S J, SON H, CHO S, et al. SRFeat: single image super-resolution with feature discrimination[C]// LNCS 11220: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 455-471.
[85]	LUO X T, XIE Y, ZHANG Y L, et al. LatticeNet: towards lightweight image super-resolution with lattice block[C]// LNCS 12367: Proceedings of the 16th European Conference on Computer Vision, Glasgow, Aug 23-28, 2020. Cham: Springer, 2020: 272-289.
[86]	姜玉宁, 李劲华, 赵俊莉. 基于生成式对抗网络的图像超分辨率重建算法[J]. 计算机工程, 2021, 47(3): 249-255.
	JIANG Y N, LI J H, ZHAO J L. Image super-resolution reconstruction algorithm based on generative adversarial network[J]. Computer Engineering, 2021, 47(3): 249-255.
[87]	KIM H, KIM G. Single image super-resolution using fire modules with asymmetric configuration[J]. IEEE Signal Processing Letters, 2020, 27: 516-519. DOI URL
[88]	XIAO M, ZHENG S, LIU C, et al. Invertible image rescaling[C]// LNCS 12346: Proceedings of the 16th European Conference on Computer Vision, Glasgow, Aug 23-28, 2020. Cham: Springer, 2020: 126-144.
[89]	WEI P, XIE Z, LU H, et al. Component divide-and-conquer for real-world image super-resolution[C]// LNCS 12353: Proceedings of the 16th European Conference on Computer Vision, Glasgow, Aug 23-28, 2020. Cham: Springer, 2020: 101-117.
[90]	LAI W S, HUANG J B, AHUJA N, et al. Fast and accurate image super-resolution with deep Laplacian pyramid networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 41(11): 2599-2613. DOI URL
[91]	蒋明峰, 支明豪, 李杨, 等. 基于自注意力机制生成对抗网络的超分辨率磁共振图像重建[J]. 中国科学: 信息科学, 2021, 51(6): 959-970.
	JIANG M F, ZHI M H, LI Y, et al. Super-resolution reconstruction of MR image with self-attention based generate adversarial network algorithm[J]. Science in China: Information Sciences, 2021, 51(6): 959-970.
[92]	NIU B, WEN W, REN W, et al. Single image super-resolution via a holistic attention network[C]// LNCS 12357: Proceedings of the 16th European Conference on Computer Vision, Glasgow, Aug 23-28, 2020. Cham: Springer, 2020: 191-207.
[93]	LU L, LI W, TAO X, et al. MASA-SR: matching acceleration and spatial adaptation for reference-based image super-resolution[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, Jun 19-25, 2021. Piscataway: IEEE, 2021: 6368-6377.
[94]	DENG X, ZHANG Y, XU M, et al. Deep coupled feedback network for joint exposure fusion and image super-resolution[J]. IEEE Transactions on Image Processing, 2021, 30: 3098-3112. DOI URL
[95]	WANG L, WANG Y, DONG X, et al. Unsupervised degradation representation learning for blind super-resolution[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, Jun 19-25, 2021. Piscataway: IEEE, 2021: 10581-10590.
[96]	KONG X, ZHAO H, QIAO Y, et al. ClassSR: a general framework to accelerate super-resolution networks by data characteristic[C]// Proceedings of the 2021 IEEE Conference on Computer Vision and Pattern Recognition, Nashville, Jun 21-24, 2021. Los Alamitos: IEEE Computer Society Press, 2021: 12016-12025.
[97]	CHEN Y, LIU S, WANG X. Learning continuous image representation with local implicit image function[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, Jun 21-24, 2021. Piscataway: IEEE, 2021: 8628-8638.
[98]	SON S, LEE K M. SRWarp: generalized image super-resolution under arbitrary transformation[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Jun 21-24, 2021. Piscataway: IEEE, 2021: 7782-7791.
[99]	HU X, MU H, ZHANG X, et al. Meta-SR: a magnification-arbitrary network for super-resolution[C]// Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 1575-1584.
[100]	NASROLLAHI K, MOESLUND T B. Super-resolution: a comprehensive survey[J]. Machine Vision and Applications, 2014, 25(6): 1423-1468. DOI URL
[101]	SHEKIH H R, BOVIK A C. Image information and visual quality[J]. IEEE Transactions on Image Processing, 2006, 15(2): 430-444. DOI URL
[102]	SAJJADI M S, SCHOLKOPF B, HIRSH M. EnhanceNet: single image super-resolution through automated texture synthesis[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 4491-4500.
[103]	WANG Z H, CHEN J, HOI S C H. Deep learning for image super-resolution: a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 43(10): 3365-3387. DOI URL
[104]	GAO R, HUANG Z Q, LIU S G. Multi-task deep learning for no-reference screen content image quality assessment[C]// LNCS 12572: Proceedings of the 27th International Conference on Multimedia Modeling, Prague, Jun 22-24, 2021. Cham: Springer, 2021: 213-226.
[105]	SAKOWICZ B, KAMINSKI M, RITTER R, et al. Methods of 3D images quality assessment[C]// Proceedings of the 2015 IEEE International Conference on Mixed Design of Integrated Circuits and Systems, Boston, Jun 7-12, 2015. Piscataway: IEEE, 2015: 123-128.
[106]	SHEKICH H R, SABIR M F, BOVIK A C. A statistical evaluation of recent full reference image quality assessment algorithms[J]. IEEE Transactions on Image Processing, 2006, 15(11): 3440-3451. DOI URL
[107]	WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612. DOI URL
[108]	MA C, YANG C Y, YANG X, et al. Learning a no-reference quality metric for single-image super-resolution[J]. Computer Vision and Image Understanding, 2017, 158: 1-16. DOI URL
[109]	ZHANG L, ZHANG L, MOU X, et al. FSIM: a feature similarity index for image quality assessment[J]. IEEE Transactions on Image Processing, 2011, 20(8): 2378-2386. DOI URL
[110]	HANG R, ISOLA P, EFROS A A, et al. The unreasonable effectiveness of deep features as a perceptual metric[C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Munich, Sep 8-14, 2018. Washington: IEEE Computer Society, 2018: 586-595.