论文速读之SUNet、MAXIM、Restormer、MIRNet、SwinIR、HINet、MPRNet、CSRNet

Transformer系列：建议细读
Pre-trained image processing transformer
SwinIR: Image Restoration Using Swin Transformer 解读
Uformer: A General U-Shaped Transformer for Image Restoration 解读
U2Former: A Nested U-shaped Transformer for Image Restoration
Restormer: Efficient Transformer for High-Resolution Image Restoration 解读
On Efficient Transformer and Image Pre-training for Low-level Vision 解读
SUNet: Swin Transformer UNet for Image Denoising
Practical Blind Denoising via Swin-Conv-UNet and Data Synthesis

引用：https://www.youtube.com/channel/UCIV7WMa5O_TZKZZSzXcLmmw/videos，侵删

SUNet: Swin Transformer with UNet for Image Denoising

MAXIM: Multi-Axis MLP for Image Processing 有难度

Restormer: Efficient Transformer for High-Resolution Image Restoration

---

MIRNet: Learning Enriched Features for Real Image Restoration and Enhancement

DAU：参考CycleISP 2019

SKFF：
下采样和上采样：

---

SwinIR: Image Restoration Using Swin Transformer

Image Denoising on SIDD(真实图像)：

Image Denoising on DND(真实图像)：

SwinIR架构：灰度、彩色高斯去噪

---

HINet: Half Instance Normalization Network for Image Restoration

解读
Instance Normalization：

HINet架构：



代码：

---

MPRNet: Multi Stage Progressive Image Restoration CVPR2021Oral

网络架构：




最后一个阶段：无上下采样

---

CSRNet: Cross-Scale Residual Network: A General Framework for Image Super-Resolution, Denoising and Deblocking

该方法是非盲去噪：