Super-Resolution
- Aligned Structured Sparsity Learning for Efficient Image Super-Resolution (NeurIPS2021) (Pruning, Align ResBlock)
- Collapsible Linear Blocks for Super-Efficient Super Resolution (arXiv2021) (Reparameter, Gradient update of Reparameter, tensorflow)
- Edge-oriented Convolution Block for Real-time Super Resolution on Mobile Devices (ACM MM2021) (Reparameter, ACB-DBB-RepVGG, +below 0.1dB, torch)
- Interpreting Super-Resolution Networks with Local Attribution Maps (CVPR2021) (Attribution Analysis, LAM可视化，Diffusion Index-Gini index, Exploration很好)
- Finding Discriminative Filters for Specific Degradations in Blind Super-Resolution (NeurIPS2021) (Attribution Analysis, 滤波器针对性的可视化)
Restoration
- ISNAS-DIP: Image-Specific Neural Architecture Search for Deep Image Prior (CVPR2022) (NAS, DIP, Fourier, Image-Specific Model, Histogram Distance)
- Path-Restore: Learning Network Path Selection for Image Restoration (Tpami2021) (Reinforcement learning for Path Selection, Mask用于指导哪里简单哪里复杂)
Denoising
- Generative Adaptive Convolutions for Real-world Noisy Image Denoising (AAAI2022) (Fourier, Dynamic Conv)
参考：Paper Reading-list_zyr_freedom的博客-CSDN博客
|