1、Second-order Attention Network for Single Image Super-Resolution
基於注意力網絡的改進，論文客觀指標最高
http://www4.comp.polyu.edu.hk/~cslzhang/paper/CVPR19-SAN.pdf
2、Image Super-Resolution by Neural Texture Transfer
比較填充的手段，思路很好
https://arxiv.org/pdf/1903.00834.pdf
3、Feedback Network for Image Super-Resolution
新思路，效果頂尖，值得深入
https://github.com/Paper99/SRFBN_CVPR19
4、Deep Plug-and-Play Super-Resolution for Arbitrary Blur Kernels
能應對不同下采樣方式。。。感覺很水
https://github.com/cszn/DPSR
5、Meta-SR: A Magnification-Arbitrary Network for Super-Resolution
任意比列超分
http://www.mingriqingbao.com/web/detail/forword/P/36079
6、Deep Back-Projection Networks for Super-Resolution
cvpr2018 ，18年贏了很多獎項
https://github.com/alterzero/DBPN-Pytorch
7、Recurrent Back-Projection Network for Video Super-Resolution
基於上面單張圖像做的改進，應用於視頻超分辨率
https://github.com/alterzero/RBPN-PyTorch
一下論文資源尚未公佈
8、Residual Networks for Light Field Image Super-Resolution
9、Hyperspectral Image Super-Resolution With Optimized RGB Guidance
10、ODE-Inspired Network Design for Single Image Super-Resolution
11、Natural and Realistic Single Image Super-Resolution With Explicit Natural Manifold Discrimination
12、Learning Parallax Attention for Stereo Image Super-Resolution
13、Fast Spatio-Temporal Residual Network for Video Super-Resolution
14、3D Appearance Super-Resolution With Deep Learning
15、Blind Super-Resolution With Iterative Kernel Correction
16、Towards Real Scene Super-Resolution With Raw Images