Meta-SR: A Magnification-Arbitrary Network for Super-Resolution Xuecai Hu, Haoyuan Mu, Xiangyu Zhang, Zilei Wang, Tieniu Tan, Jian Sun [ pdf ]
Blind Super-Resolution With Iterative Kernel Correction Jinjin Gu, Hannan Lu, Wangmeng Zuo, Chao Dong [ pdf ]
Camera Lens Super-Resolution Chang Chen, Zhiwei Xiong, Xinmei Tian, Zheng-Jun Zha, Feng Wu [ pdf ], [ supp ]
Deep Plug-And-Play Super-Resolution for Arbitrary Blur Kernels Kai Zhang, Wangmeng Zuo, Lei Zhang [ pdf ]
Towards Real Scene Super-Resolution With Raw Images Xiangyu Xu, Yongrui Ma, Wenxiu Sun [ pdf ]
ODE-Inspired Network Design for Single Image Super-Resolution Xiangyu He, Zitao Mo, Peisong Wang, Yang Liu, Mingyuan Yang, Jian Cheng [ pdf ], [ supp ]
Feedback Network for Image Super-Resolution Zhen Li, Jinglei Yang, Zheng Liu, Xiaomin Yang, Gwanggil Jeon, Wei Wu [ pdf ], [ supp ]
Recurrent Back-Projection Network for Video Super-Resolution Muhammad Haris, Gregory Shakhnarovich, Norimichi Ukita [ pdf ], [ supp ]
Image Super-Resolution by Neural Texture Transfer Zhifei Zhang, Zhaowen Wang, Zhe Lin, Hairong Qi [ pdf ]
Natural and Realistic Single Image Super-Resolution With Explicit Natural Manifold Discrimination Jae Woong Soh, Gu Yong Park, Junho Jo, Nam Ik Cho [ pdf ], [ supp ]
3D Appearance Super-Resolution With Deep Learning Yawei Li, Vagia Tsiminaki, Radu Timofte, Marc Pollefeys, Luc Van Gool [ pdf ], [ supp ]
Fast Spatio-Temporal Residual Network for Video Super-Resolution Sheng Li, Fengxiang He, Bo Du, Lefei Zhang, Yonghao Xu, Dacheng Tao [ pdf ], [ supp ]
Residual Networks for Light Field Image Super-Resolution Shuo Zhang, Youfang Lin, Hao Sheng [ pdf ]
Second-Order Attention Network for Single Image Super-Resolution Tao Dai, Jianrui Cai, Yongbing Zhang, Shu-Tao Xia, Lei Zhang [ pdf ], [ pdf ]
Hyperspectral Image Super-Resolution With Optimized RGB Guidance Ying Fu, Tao Zhang, Yinqiang Zheng, Debing Zhang, Hua Huang [ pdf ]
Learning Parallax Attention for Stereo Image Super-Resolution Longguang Wang, Yingqian Wang, Zhengfa Liang, Zaiping Lin, Jungang Yang, Wei An, Yulan Guo [ pdf ], [ supp ]
Face Super-resolution Guided by Facial Component Heatmaps Xin Yu, Basura Fernando, Bernard Ghanem, Fatih Porikli, Richard Hartley [ pdf ]
Image Super-Resolution Using Very Deep Residual Channel Attention Networks Yulun Zhang, Kunpeng Li, Kai Li, Lichen Wang, Bineng Zhong, Yun Fu [ pdf ]
Super-Resolution and Sparse View CT Reconstruction Guangming Zang, Mohamed Aly, Ramzi Idoughi, Peter Wonka, Wolfgang Heidrich [ pdf ]
Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network Namhyuk Ahn, Byungkon Kang, Kyung-Ah Sohn [ pdf ]
SRFeat: Single Image Super-Resolution with Feature Discrimination Seong-Jin Park, Hyeongseok Son, Sunghyun Cho, Ki-Sang Hong, Seungyong Lee [ pdf ]
To learn image super-resolution, use a GAN to learn how to do image degradation first Adrian Bulat, Jing Yang, Georgios Tzimiropoulos [ pdf ]
Multi-scale Residual Network for Image Super-Resolution Juncheng Li, Faming Fang, Kangfu Mei, Guixu Zhang [ pdf ]
Super-FAN: Integrated Facial Landmark Localization and Super-Resolution of Real-World Low Resolution Faces in Arbitrary Poses With GANs Adrian Bulat, Georgios Tzimiropoulos [ pdf ] [ Supp ]
Fight Ill-Posedness With Ill-Posedness: Single-Shot Variational Depth Super-Resolution From Shading Bjoern Haefner, Yvain Quéau, Thomas Möllenhoff, Daniel Cremers [ pdf ] [ Supp ]
Recovering Realistic Texture in Image Super-Resolution by Deep Spatial Feature Transform Xintao Wang, Ke Yu, Chao Dong, Chen Change Loy [ pdf ]
Fast and Accurate Single Image Super-Resolution via Information Distillation Network Zheng Hui, Xiumei Wang, Xinbo Gao [ pdf ]
Image Super-Resolution via Dual-State Recurrent Networks Wei Han, Shiyu Chang, Ding Liu, Mo Yu, Michael Witbrock, Thomas S. Huang [ pdf ]
Deep Back-Projection Networks for Super-Resolution Muhammad Haris, Gregory Shakhnarovich, Norimichi Ukita [ pdf ] [ Supp ]
Residual Dense Network for Image Super-Resolution Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhong, Yun Fu [ pdf ]
FSRNet: End-to-End Learning Face Super-Resolution With Facial Priors Yu Chen, Ying Tai, Xiaoming Liu, Chunhua Shen, Jian Yang [ pdf ]
Unsupervised Sparse Dirichlet-Net for Hyperspectral Image Super-Resolution Ying Qu, Hairong Qi, Chiman Kwan [ pdf ]
“Zero-Shot” Super-Resolution Using Deep Internal Learning Assaf Shocher, Nadav Cohen, Michal Irani [ pdf ]
Deep Video Super-Resolution Network Using Dynamic Upsampling Filters Without Explicit Motion Compensation Younghyun Jo, Seoung Wug Oh, Jaeyeon Kang, Seon Joo Kim [ pdf ] [ Supp ]
Learning a Single Convolutional Super-Resolution Network for Multiple Degradations Kai Zhang, Wangmeng Zuo, Lei Zhang [ pdf ]
Feature Super-Resolution: Make Machine See More Clearly Weimin Tan, Bo Yan, Bahetiyaer Bare [ pdf ]
Frame-Recurrent Video Super-Resolution Mehdi S. M. Sajjadi, Raviteja Vemulapalli, Matthew Brown [ pdf ]
Temporal Shape Super-Resolution by Intra-Frame Motion Encoding Using High-Fps Structured Light Yuki Shiba, Satoshi Ono, Ryo Furukawa, Shinsaku Hiura, Hiroshi Kawasaki [ pdf ] [ Supp ] [ video ]
Robust Video Super-Resolution With Learned Temporal Dynamics Ding Liu, Zhaowen Wang, Yuchen Fan, Xianming Liu, Zhangyang Wang, Shiyu Chang, Thomas Huang [ pdf ]
Detail-Revealing Deep Video Super-Resolution Xin Tao, Hongyun Gao, Renjie Liao, Jue Wang, Jiaya Jia [ pdf ] [ video ]
EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis Mehdi S. M. Sajjadi, Bernhard Scholkopf, Michael Hirsch [ pdf ] [ Supp ][ video ]
Joint Estimation of Camera Pose, Depth, Deblurring, and Super-Resolution From a Blurred Image Sequence Haesol Park, Kyoung Mu Lee [ pdf ] [ Supp ]
Image Super-Resolution Using Dense Skip Connections Tong Tong, Gen Li, Xiejie Liu, Qinquan Gao [ pdf ]
Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution Wei-Sheng Lai, Jia-Bin Huang, Narendra Ahuja, Ming-Hsuan Yang [ pdf ]
Image Super-Resolution via Deep Recursive Residual Network Ying Tai, Jian Yang, Xiaoming Liu [ pdf ]
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network Christian Ledig, Lucas Theis, Ferenc Huszar, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, Wenzhe Shi [ pdf ] [ poster ] [ video ]
Real-Time Video Super-Resolution With Spatio-Temporal Networks and Motion Compensation Jose Caballero, Christian Ledig, Andrew Aitken, Alejandro Acosta, Johannes Totz, Zehan Wang, Wenzhe Shi [ pdf ] [ poster ]
Hyperspectral Image Super-Resolution via Non-Local Sparse Tensor Factorization Renwei Dian, Leyuan Fang, Shutao Li [ pdf ] [ poster ]
Simultaneous Super-Resolution and Cross-Modality Synthesis of 3D Medical Images Using Weakly-Supervised Joint Convolutional Sparse Coding Yawen Huang, Ling Shao, Alejandro F. Frangi [ pdf ] [ poster ]
Reference Guided Deep Super-Resolution via Manifold Localized External Compensation Wenhan Yang, Sifeng Xia, Jiaying Liu, and Zongming Guo Accepted by IEEE Trans. on Circuit System for Video Technology (TCSVT), June 2018. [ project ]
Joint-Feature Guided Depth Map Super-Resolution With Face Priors Shuai Yang, Jiaying Liu, Yuming Fang, and Zongming Guo IEEE Trans. on Cybernetics (TCYB), , , , Jan. 2018. [ project ]
Deep Edge Guided Recurrent Residual Learning for Image Super-Resolution Wenhan Yang, Jiashi Feng, Jianchao Yang, Fang Zhao, Jiaying Liu, Zongming Guo and Shuicheng Yan IEEE Trans. on Image Processing (TIP), , , , Dec. 2017. [ project ]
Retrieval Compensated Group Structured Sparsity for Image Super-Resolution Jiaying Liu, Wenhan Yang, Xinfeng Zhang, and Zongming Guo IEEE Trans. on Multimedia (TMM), , , , Feb. 2017. [ project ]
ECCV2016 Accelerating the Super-Resolution Convolutional Neural Network Chao Dong, Chen Change Loy, Xiaoou Tang [ project ]
Perceptual Losses for Real-Time Style Transfer and Super-Resolution Justin Johnson, Alexandre Alahi, Li Fei-Fei [ pdf ]
End-to-End Image Super-Resolution via Deep and Shallow Convolutional Networks Yifan Wang, Lijun Wang, Hongyu Wang, Peihua Li [ pdf ]
Deeply-Recursive Convolutional Network for Image Super-Resolution Jiwon Kim, Jung Kwon Lee, Kyoung Mu Lee [ pdf ]
Accurate Image Super-Resolution Using Very Deep Convolutional Networks Jiwon Kim, Jung Kwon Lee, Kyoung Mu Lee [ pdf ]
Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network Wenzhe Shi, Jose Caballero, Ferenc Huszar, Johannes Totz, Andrew P. Aitken, Rob Bishop, Daniel Rueckert, Zehan Wang [ pdf ]
Image Super-Resolution Using Deep Convolutional Networks Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) SRCNN [ project ] [ pdf ] [ supplementary material ]
Neighborhood Regression for Edge-Preserving Image Super-Resolution Yanghao Li, Jiaying Liu, Wenhan Yang and Zongming Guo IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Brisbane, Australia, Apr. 2015. [ project ]
Learning a Deep Convolutional Network for Image Super-Resolution Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang ECCV 2014 SRCNN [ project ] [ pdf ]
Super-Resolution From a Single Image Daniel Glasner, Shai Bagon, Michal Irani ICCV 2009 [ project ] [ pdf ]
Video Super-Resolution Based on Spatial-Temporal Recurrent Neural Networks Wenhan Yang, Jiashi Feng, Guosen Xie, Jiaying Liu, Zongming Guo and Shuicheng Yan Computer Vision and Image Understanding (CVIU), , , March. 2018. [ project ]
Video Super-Resolution With Convolutional Neural Networks Armin Kappeler ; Seunghwan Yoo ; Qiqin Dai ; Aggelos K. Katsaggelos IEEE Transactions on Computational Imaging [ pdf ]
High Quality Remote Sensing Image Super-Resolution Using Deep Memory Connected Network Wenjia Xu; Guangluan XU; Yang Wang; Xian Sun; Daoyu Lin; Yirong WU IGARSS 2018 - 2018 IEEE International Geoscience and Remote SensingSymposium [ pdf ]
Domain Generalization是Domain Adaptation的一种特殊问题,在DA的基础上,DG的目的为对于 任意不可见 的目标域都有很强的鲁棒性。
各源域 及目标域 之间的Momment距离定义为源域与目标域的距离及各源域之间的距离。公式如下:
目标函数表示为 个源域的分类损失加上前述
S-T的对齐比较有效。
这是一篇ICCV2019的论文。 实际中,标注数据和非标注数据往往存在数量级的差距,标注数据的分布和非标注数据差距(可认为是真实分布)也很大。本文通过对抗学习拉齐标注数据和非标注数据的分布,此外参考mixup方法利用非标注数据尝试解决标注数据较少的问题。
如下,是本文的loss公式,公式分为两部分, 第一部分表示分类任务,第二部分表示对抗学习。
具体地,对抗损失如下, 为样本属于标注样本 还是非标注样本 的分类函数。通过对抗学习,可以将标注样本和非标注样本的特征 拉到同一空间。
此外,为了解决标注样本数量较少的问题,本文采用mixup方法利用非标注样本,公式如下,非标注样本的类别标签 由分类器生成。鉴别器的标签 也进行mixup。
下图是在SVHN数据集上的表现及跟STOA方法的对比。
整体流程如下:
Episodic的效果
李文
论文名称:Rich feature hierarchies for accurate object detection and semantic segmentation 提出时间:2014年 论文地址: 针对问题: 从Alexnet提出后,作者等人思考如何利用卷积网络来完成检测任务,即输入一张图,实现图上目标的定位(目标在哪)和分类(目标是什么)两个目标,并最终完成了RCNN网络模型。 创新点: RCNN提出时,检测网络的执行思路还是脱胎于分类网络。也就是深度学习部分仅完成输入图像块的分类工作。那么对检测任务来说如何完成目标的定位呢,作者采用的是Selective Search候选区域提取算法,来获得当前输入图上可能包含目标的不同图像块,再将图像块裁剪到固定的尺寸输入CNN网络来进行当前图像块类别的判断。 参考博客: 。 论文题目:OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks 提出时间:2014年 论文地址: 针对问题: 该论文讨论了,CNN提取到的特征能够同时用于定位和分类两个任务。也就是在CNN提取到特征以后,在网络后端组织两组卷积或全连接层,一组用于实现定位,输出当前图像上目标的最小外接矩形框坐标,一组用于分类,输出当前图像上目标的类别信息。也是以此为起点,检测网络出现基础主干网络(backbone)+分类头或回归头(定位头)的网络设计模式雏形。 创新点: 在这篇论文中还有两个比较有意思的点,一是作者认为全连接层其实质实现的操作和1x1的卷积是类似的,而且用1x1的卷积核还可以避免FC对输入特征尺寸的限制,那用1x1卷积来替换FC层,是否可行呢?作者在测试时通过将全连接层替换为1x1卷积核证明是可行的;二是提出了offset max-pooling,也就是对池化层输入特征不能整除的情况,通过进行滑动池化并将不同的池化层传递给后续网络层来提高效果。另外作者在论文里提到他的用法是先基于主干网络+分类头训练,然后切换分类头为回归头,再训练回归头的参数,最终完成整个网络的训练。图像的输入作者采用的是直接在输入图上利用卷积核划窗。然后在指定的每个网络层上回归目标的尺度和空间位置。 参考博客: 论文题目:Scalable Object Detection using Deep Neural Networks 提出时间:2014年 论文地址: 针对问题: 既然CNN网络提取的特征可以直接用于检测任务(定位+分类),作者就尝试将目标框(可能包含目标的最小外包矩形框)提取任务放到CNN中进行。也就是直接通过网络完成输入图像上目标的定位工作。 创新点: 本文作者通过将物体检测问题定义为输出多个bounding box的回归问题. 同时每个bounding box会输出关于是否包含目标物体的置信度, 使得模型更加紧凑和高效。先通过聚类获得图像中可能有目标的位置聚类中心,(800个anchor box)然后学习预测不考虑目标类别的二分类网络,背景or前景。用到了多尺度下的检测。 参考博客: 论文题目:DeepBox: Learning Objectness with Convolutional Networks 提出时间:2015年ICCV 论文地址: 主要针对的问题: 本文完成的工作与第三篇类似,都是对目标框提取算法的优化方案,区别是本文首先采用自底而上的方案来提取图像上的疑似目标框,然后再利用CNN网络提取特征对目标框进行是否为前景区域的排序;而第三篇为直接利用CNN网络来回归图像上可能的目标位置。创新点: 本文作者想通过CNN学习输入图像的特征,从而实现对输入网络目标框是否为真实目标的情况进行计算,量化每个输入框的包含目标的可能性值。 参考博客: 论文题目:AttentionNet: AggregatingWeak Directions for Accurate Object Detection 提出时间:2015年ICCV 论文地址: 主要针对的问题: 对检测网络的实现方案进行思考,之前的执行策略是,先确定输入图像中可能包含目标位置的矩形框,再对每个矩形框进行分类和回归从而确定目标的准确位置,参考RCNN。那么能否直接利用回归的思路从图像的四个角点,逐渐得到目标的最小外接矩形框和类别呢? 创新点: 通过从图像的四个角点,逐步迭代的方式,每次计算一个缩小的方向,并缩小指定的距离来使得逐渐逼近目标。作者还提出了针对多目标情况的处理方式。 参考博客: 论文题目:Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition 提出时间:2014年 论文地址: 针对问题: 如RCNN会将输入的目标图像块处理到同一尺寸再输入进CNN网络,在处理过程中就造成了图像块信息的损失。在实际的场景中,输入网络的目标尺寸很难统一,而网络最后的全连接层又要求输入的特征信息为统一维度的向量。作者就尝试进行不同尺寸CNN网络提取到的特征维度进行统一。创新点: 作者提出的SPPnet中,通过使用特征金字塔池化来使得最后的卷积层输出结果可以统一到全连接层需要的尺寸,在训练的时候,池化的操作还是通过滑动窗口完成的,池化的核宽高及步长通过当前层的特征图的宽高计算得到。原论文中的特征金字塔池化操作图示如下。 参考博客 : 论文题目:Object detection via a multi-region & semantic segmentation-aware CNN model 提出时间:2015年 论文地址: 针对问题: 既然第三篇论文multibox算法提出了可以用CNN来实现输入图像中待检测目标的定位,本文作者就尝试增加一些训练时的方法技巧来提高CNN网络最终的定位精度。创新点: 作者通过对输入网络的region进行一定的处理(通过数据增强,使得网络利用目标周围的上下文信息得到更精准的目标框)来增加网络对目标回归框的精度。具体的处理方式包括:扩大输入目标的标签包围框、取输入目标的标签中包围框的一部分等并对不同区域分别回归位置,使得网络对目标的边界更加敏感。这种操作丰富了输入目标的多样性,从而提高了回归框的精度。 参考博客 : 论文题目:Fast-RCNN 提出时间:2015年 论文地址: 针对问题: RCNN中的CNN每输入一个图像块就要执行一次前向计算,这显然是非常耗时的,那么如何优化这部分呢? 创新点: 作者参考了SPPNet(第六篇论文),在网络中实现了ROIpooling来使得输入的图像块不用裁剪到统一尺寸,从而避免了输入的信息丢失。其次是将整张图输入网络得到特征图,再将原图上用Selective Search算法得到的目标框映射到特征图上,避免了特征的重复提取。 参考博客 : 论文题目:DeepProposal: Hunting Objects by Cascading Deep Convolutional Layers 提出时间:2015年 论文地址: 主要针对的问题: 本文的作者观察到CNN可以提取到很棒的对输入图像进行表征的论文,作者尝试通过实验来对CNN网络不同层所产生的特征的作用和情况进行讨论和解析。 创新点: 作者在不同的激活层上以滑动窗口的方式生成了假设,并表明最终的卷积层可以以较高的查全率找到感兴趣的对象,但是由于特征图的粗糙性,定位性很差。相反,网络的第一层可以更好地定位感兴趣的对象,但召回率降低。 论文题目:Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks 提出时间:2015年NIPS 论文地址: 主要针对的问题: 由multibox(第三篇)和DeepBox(第四篇)等论文,我们知道,用CNN可以生成目标待检测框,并判定当前框为目标的概率,那能否将该模型整合到目标检测的模型中,从而实现真正输入端为图像,输出为最终检测结果的,全部依赖CNN完成的检测系统呢? 创新点: 将当前输入图目标框提取整合到了检测网络中,依赖一个小的目标框提取网络RPN来替代Selective Search算法,从而实现真正的端到端检测算法。 参考博客 :
318 浏览 4 回答
104 浏览 3 回答
288 浏览 2 回答
105 浏览 3 回答
171 浏览 5 回答
334 浏览 3 回答
268 浏览 3 回答
277 浏览 5 回答
341 浏览 4 回答
232 浏览 4 回答
328 浏览 3 回答
143 浏览 3 回答
159 浏览 4 回答
288 浏览 6 回答
175 浏览 4 回答