Key-Graph Transformer for Image Restoration
Not on the map. Why?
Abstract (arXiv)
While it is crucial to capture global information for effective image restoration (IR), integrating such cues into transformer-based methods becomes computationally expensive, especially with high input resolution. Furthermore, the self-attention mechanism in transformers is prone to considering unnecessary global cues from unrelated objects or regions, introducing computational inefficiencies. In response to these challenges, we introduce the Key-Graph Transformer (KGT) in this paper. Specifically, KGT views patch features as graph nodes. The proposed Key-Graph Constructor efficiently forms a sparse yet representative Key-Graph by selectively connecting essential nodes instead of all the nodes. Then the proposed Key-Graph Attention is conducted under the guidance of the Key-Graph only among selected nodes with linear computational complexity within each window. Extensive experiments across 6 IR tasks confirm the proposed KGT's state-of-the-art performance, showcasing advancements both quantitatively and qualitatively.
Key facts. Cited by 3 papers in the map (data through 13 Oct 2026). Builds on 62 of its 95 references that are in the map. First posted on arXiv 4 Feb 2024.
Builds on
62 of 95 references.
References · 62
- 2017Attention Is All You Needcited by 60,106
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,043
- 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,111
- 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,516
- 2022FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenesscited by 3,637
- 2017Non-local Neural Networkscited by 2,291
- 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,222
- 2021SwinIR: Image Restoration Using Swin Transformercited by 1,876
- 2021Restormer: Efficient Transformer for High-Resolution Image Restorationcited by 1,622
- 2017Enhanced Deep Residual Networks for Single Image Super-Resolutioncited by 1,549
- 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,259
- 2020Pre-Trained Image Processing Transformercited by 892
38 more
- 2018Representation Learning on Graphs with Jumping Knowledge Networkscited by 877
- 2021Multi-Stage Progressive Image Restorationcited by 813
- 2016Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurringcited by 724
- 2022Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Modelcited by 580
- 2020Learning Continuous Image Representation with Local Implicit Image Functioncited by 552
- 2019Measuring and Relieving the Over-smoothing Problem for Graph Neural Networks from the Topological Viewcited by 545
- 2017FFDNet: Toward a Fast and Flexible Solution for CNN based Image Denoisingcited by 509
- 2020Plug-and-Play Image Restoration with Deep Denoiser Priorcited by 469
- 2022Activating More Pixels in Image Super-Resolution Transformercited by 435
- 2017DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networkscited by 366
- 2023ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shiftingcited by 364
- 2019DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Bettercited by 361
- 2018Scale-recurrent Network for Deep Image Deblurringcited by 332
- 2021Rethinking Coarse-to-Fine Approach in Single Image Deblurringcited by 324
- 2017Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphscited by 308
- 2022MAXIM: Multi-Axis MLP for Image Processingcited by 301
- 2020Single Image Super-Resolution via a Holistic Attention Networkcited by 265
- 2021TransWeather: Transformer-based Restoration of Images Degraded by Adverse Weather Conditionscited by 244
- 2019Residual Non-local Attention Networks for Image Restorationcited by 236
- 2023Implicit Diffusion Models for Continuous Super-Resolutioncited by 217
- 2019Deep Stacked Hierarchical Multi-patch Network for Image Deblurringcited by 205
- 2017Attentive Generative Adversarial Network for Raindrop Removal from a Single Imagecited by 204
- 2023Learning A Sparse Transformer Network for Effective Image Derainingcited by 204
- 2017DesnowNet: Context-Aware Deep Network for Snow Removalcited by 200
- 2022Structure-Aware Transformer for Graph Representation Learningcited by 191
- 2020Deblurring by Realistic Blurringcited by 188
- 2022Vision GNN: An Image is Worth Graph of Nodescited by 185
- 2023Image Restoration with Mean-Reverting Stochastic Differential Equationscited by 183
- 2020Human-Aware Motion Deblurringcited by 174
- 2018Non-Local Recurrent Network for Image Restorationcited by 160
- 2019Heavy Rain Image Restoration: Integrating Physics Model and Conditional Adversarial Learningcited by 157
- 2023Stabilizing Transformer Training by Preventing Attention Entropy Collapsecited by 157
- 2023Efficient and Explicit Modelling of Image Hierarchies for Image Restorationcited by 145
- 2022Signal Propagation in Transformers: Theoretical Perspectives and the Role of Rank Collapsecited by 138
- 2020Spatially-Attentive Patch-Hierarchical Network for Adaptive Motion Deblurringcited by 119
- 2019Multi-Temporal Recurrent Neural Networks For Progressive Non-Uniform Single Image Deblurring With Incremental Temporal Trainingcited by 112
- 2022Not too little, not too much: a theoretical analysis of graph (over)smoothingcited by 106
- 2022Cross Aggregation Transformer for Image Restorationcited by 92
50 most cited of 62 All 62 (JSON)
Built on by
Citing papers per year: 2025: 2; 2026: 1.
3 papers.