Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network

计算机科学 相似性(几何) 人工智能 平均意见得分 残余物 图像(数学) 卷积神经网络 计算机视觉 像素 深度学习 鉴别器 模式识别(心理学) 算法 公制(单位) 探测器 电信 运营管理 经济
作者
Christian Ledig,Lucas Theis,Ferenc Huszár,José Caballero,Andrew Cunningham,Alejandro Acosta,Andrew P. Aitken,Alykhan Tejani,Johannes Totz,Zehan Wang,Wenzhe Shi
出处
期刊:Computer Vision and Pattern Recognition 卷期号:: 105-114 被引量:12510
标识
DOI:10.1109/cvpr.2017.19
摘要

Despite the breakthroughs in accuracy and speed of single image super-resolution using faster and deeper convolutional neural networks, one central problem remains largely unsolved: how do we recover the finer texture details when we super-resolve at large upscaling factors? The behavior of optimization-based super-resolution methods is principally driven by the choice of the objective function. Recent work has largely focused on minimizing the mean squared reconstruction error. The resulting estimates have high peak signal-to-noise ratios, but they are often lacking high-frequency details and are perceptually unsatisfying in the sense that they fail to match the fidelity expected at the higher resolution. In this paper, we present SRGAN, a generative adversarial network (GAN) for image super-resolution (SR). To our knowledge, it is the first framework capable of inferring photo-realistic natural images for 4x upscaling factors. To achieve this, we propose a perceptual loss function which consists of an adversarial loss and a content loss. The adversarial loss pushes our solution to the natural image manifold using a discriminator network that is trained to differentiate between the super-resolved images and original photo-realistic images. In addition, we use a content loss motivated by perceptual similarity instead of similarity in pixel space. Our deep residual network is able to recover photo-realistic textures from heavily downsampled images on public benchmarks. An extensive mean-opinion-score (MOS) test shows hugely significant gains in perceptual quality using SRGAN. The MOS scores obtained with SRGAN are closer to those of the original high-resolution images than to those obtained with any state-of-the-art method.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
刚刚
jerome完成签到 ,获得积分10
1秒前
科研通AI2S的应助被花花衬衫采纳,获得10
1秒前
julie发布了新的文献求助10
1秒前
慕青的应助被维C续航中采纳,获得10
1秒前
小马甲的应助被Qwertyuiop采纳,获得10
1秒前
CEO完成签到,获得积分10
2秒前
2秒前
风吹麦浪的应助被予秋采纳,获得10
2秒前
蛋蛋发布了新的文献求助10
2秒前
风吹麦浪的应助被予秋采纳,获得10
2秒前
2秒前
南故完成签到,获得积分10
3秒前
3秒前
gggx完成签到,获得积分10
3秒前
3秒前
田様的应助被LiClMn采纳,获得10
3秒前
4秒前
5秒前
科院er完成签到 ,获得积分10
5秒前
Hello的应助被无聊的寒香采纳,获得10
5秒前
顾矜的应助被生而追梦不止采纳,获得10
6秒前
6秒前
6秒前
6秒前
李李发布了新的文献求助10
6秒前
龙海完成签到 ,获得积分10
6秒前
6666的应助被zombie采纳,获得10
6秒前
7秒前
LCBR发布了新的文献求助10
7秒前
淡定的静枫完成签到,获得积分10
7秒前
小二郎的应助被julie采纳,获得10
8秒前
8秒前
8秒前
8秒前
artemis发布了新的文献求助30
8秒前
wyc完成签到 ,获得积分10
9秒前
IRONARMOUR完成签到,获得积分10
9秒前
ijiahe完成签到,获得积分10
9秒前
water完成签到,获得积分10
9秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
Social Psychology in the Real World 600
八旗制度の研究 600
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8181155
求助须知:如何正确求助?哪些是违规求助? 9571140
关于积分的说明 21816740
捐赠科研通 7777528
什么是DOI,文献DOI怎么找? 3433436
关于科研通互助平台的介绍 2557664
邀请新用户注册赠送积分活动 2479482