非宁静无以致远。
我们在之前两篇博客,深度学习应用到图像超分辨率重建1, 深度学习应用到图像超分辨率重建2已经介绍了一些图像超分辨的基础了， 下面我们继续分享一些最新的一些论文。之前只是想稍微简单介绍一下每一篇文章，但是写着写着发现写的越来越多，联想的就越来越多了。将所有的内容放到一篇文章可能显得有点冗杂，所以自己将每一篇文章单独写了一个博客，然后给了相应的传送门。
1. CrossNet: An End-to-end Reference-based Super Resolution Network using Cross-scale Warping (ECCV, 2018)
传送门：文献阅读:CrossNet: An End-to-end Reference-based Super Resolution Network using Cross-scale Warping
2. Multi-scale Residual Network for Image Super-Resolution (ECCV, 2018)
传送门: 文献阅读:Multi-scale Residual Network for Image Super-Resolution
3. Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network (ECCV, 2018)
传送门: 文献阅读:Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network
4. SRFeat: Single Image Super-Resolution with Feature Discrimination (ECCV, 2018)
传送门: 文献阅读:SRFeat: Single Image Super-Resolution with Feature Discrimination
5. Image Super-Resolution Using Very Deep Residual Channel Attention Networks (ECCV, 2018)
传送门: 文献阅读:Image Super-Resolution Using Very Deep Residual Channel Attention Networks
6. To learn image super-resolution, use a GAN to learn how to do image degradation first (ECCV, 2018)
传送门: 文献阅读:To learn image super-resolution, use a GAN to learn how to do image degradation first
7. ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (ECCV18 Workshops )
文章地址： https://arxiv.org/abs/1809.00219
作者的项目地址： https://github.com/xinntao/ESRGAN
8. Analyzing Perception-Distortion Tradeoff using Enhanced Perceptual Super-resolution Network（ECCV2018 Workshops）
文章地址：EPSR
作者的项目地址：https://github.com/subeeshvasu/2018_subeesh_epsr_eccvw
9. Video Super-Resolution Using Non-Simultaneous Fully Recurrent Convolutional Network (TIP 2019)
文章地址： 暂无
作者的项目地址：暂无
这里我们还介绍几篇写的比较好的博客:
- 知乎: 深度学习端到端超分辨率方法发展历程（二）
- 量子位：超分辨率技术如何发展？这6篇ECCV 18论文带你一次尽览
- 知乎: 超分辨率重建相关网站整理（持续更新）
- github: Single-Image-Super-Resolution
- 知乎：现在low-level vision的热点方向是什么？
图像超分辨技术发展的还是比较快的，不管是图像超分辨重建，图像去噪，图像去马赛克,这些都有着千丝万缕的关系。我们都可以将好的思想借鉴MRI重建当中。另外，最近的sr的项目基本上都是pytorch写的～