ECCV 2024 超分辨率（super-resolution）方向上接收论文总结

ECCV 2024

ECCV（欧洲计算机视觉国际会议）创办于 1887 年，每两年举办一次。它与 CVPR（每年一届）ICCV（每两年一届）并称计算机视觉方向的三大顶级会议。今年，是 ECCV 召开的第 18 届顶会，于 9 月 29 日-10 月 4 日在意大利米兰正式开幕。

现将超分辨率方向上接收的论文汇总如下，遗漏之处还请大家斧正。

图像超分

1 Accelerating Image Super-Resolution Networks with Pixel-Level Classification

• Paper: https://arxiv.org/abs/2407.21448
• Code: https://github.com/3587jjh/PCSR
• Keywords: Lightweight

2 AdaDiffSR: Adaptive Region-aware Dynamic acceleration Diffusion Model for Real-World Image Super-Resolution

• Paper: ECCV 2024
• Code:
• Keywords: Real-World, Diffusion

3 MTKD: Multi-Teacher Knowledge Distillation for Image Super-Resolution

• Paper: https://arxiv.org/abs/2404.09571
• Code:
• Keywords: Knowledge Distillation

4 You Only Need One Step: Fast Super-Resolution with Stable Diffusion via Scale Distillation

• Paper: https://arxiv.org/abs/2401.17258
• Code:
• Keywords: Diffusion, Lightweight

5 Pairwise Distance Distillation for Unsupervised Real-World Image Super-Resolution

• Paper: https://arxiv.org/abs/2407.07302
• Code: https://github.com/Yuehan717/PDD
• Keywords: Unsupervised, Real-World

6 Overcoming Distribution Mismatch in Quantizing Image Super-Resolution Networks

• Paper: https://arxiv.org/abs/2307.13337
• Code:
• Keywords: Quantizing

7 Learning Dual-Level Deformable Implicit Representation for Real-World Scale Arbitrary Super-Resolution

• Paper: https://arxiv.org/abs/2403.10925
• Code: https://github.com/nonozhizhiovo/RealArbiSR
• Keywords: Real-World, Scale Arbitrary

8 XPSR: Cross-modal Priors for Diffusion-based Image Super-Resolution

• Paper: https://arxiv.org/abs/2403.05049
• Code: https://github.com/qyp2000/XPSR
• Keywords: Diffusion

9 HiT-SR: Hierarchical Transformer for Efficient Image Super-Resolution

• Paper: https://arxiv.org/abs/2407.05878
• Code: https://github.com/XiangZ-0/HiT-SR
• Keywords: Transformer, Lightweight

10 SMFANet: A Lightweight Self-Modulation Feature Aggregation Network for Efficient Image Super-Resolution

• Paper: ECCV 2024
• Code: https://github.com/Zheng-MJ/SMFANet
• Keywords: Lightweight

11 Pixel-Aware Stable Diffusion for Realistic Image Super-Resolution and Personalized Stylization

• Paper: https://arxiv.org/abs/2308.14469
• Code: https://github.com/yangxy/PASD/
• Keywords: Diffusion
• Blog: 论文笔记：Pixel-Aware Stable Diffusion for Realistic Image Super-Resolution and Personalized Stylization

12 Spatially-Variant Degradation Model for Dataset-free Super-resolution

• Paper: https://arxiv.org/abs/2407.08252
• Code: https://github.com/shaojieguoECNU/SVDSR
• Keywords: Blind

13 Confidence-Based Iterative Generation for Real-World Image Super-Resolution

• Paper: ECCV 2024
• Code: https://github.com/USTC-JialunPeng/RealSRT
• Keywords: Real-World

视频超分

1 Event-Adapted Video Super-Resolution

• Paper: ECCV 2024
• Code:
• Keywords: Event

2 Kalman-Inspired Feature Propagation for Video Face Super-Resolution

• Paper: https://arxiv.org/abs/2408.05205
• Code: https://jnjaby.github.io/projects/KEEP
• Keywords: Face

3 Data Overfitting for On-Device Super-Resolution with Dynamic Algorithm and Compiler Co-Design

• Paper: https://arxiv.org/abs/2407.02813
• Code: https://github.com/coulsonlee/Dy-DCA-ECCV2024
• Keywords: Lightweight

4 Enhancing Perceptual Quality in Video Super-Resolution through Temporally-Consistent Detail Synthesis using Diffusion Models

• Paper: https://arxiv.org/abs/2311.15908
• Code: https://github.com/claudiom4sir/StableVSR
• Keywords: Diffusion

5 Motion-Guided Latent Diffusion for Temporally Consistent Real-world Video Super-resolution

• Paper: https://arxiv.org/abs/2312.00853
• Code: https://github.com/Zhang-Jing-Xuan/MGLD-VSR/
• Keywords: Real-world
• Blog: 用于实现真实世界时间一致性视频超分的运动引导潜在扩散模型

6 Arbitrary-Scale Video Super-Resolution with Structural and Textural Priors

• Paper: https://arxiv.org/abs/2407.09919
• Code: https://github.com/shangwei5/ST-AVSR
• Keywords: Arbitrary-Scale

7 RealViformer: Investigating Attention for Real-World Video Super-Resolution

• Paper: https://arxiv.org/abs/2407.13987
• Code: https://github.com/Yuehan717/RealViformer
• Keywords: Real-Worl

特殊场景

1 Adaptive Multi-modal Fusion of Spatially Variant Kernel Refinement with Diffusion Model for Blind Image Super-Resolution

• Paper: https://arxiv.org/abs/2403.05808
• Code:
• Keywords: Blind, Multi-modal, Diffusion

2 DCDM: Diffusion-Conditioned-Diffusion Model for Scene Text Image Super-Resolution

• Paper: ECCV 2024
• Code: https://github.com/shreygithub/DCDM
• Keywords: Diffusion, Scene Text

3 Learning Exhaustive Correlation for Spectral Super-Resolution: Where Spatial-Spectral Attention Meets Linear Dependence

• Paper: https://arxiv.org/abs/2312.12833
• Code:
• Keywords: Spectral

4 Contourlet Residual for Prompt Learning Enhanced Infrared Image Super-Resolution

• Paper: ECCV 2024
• Code: https://github.com/hey-it-s-me/CoRPLE
• Keywords: Infrared

5 UCIP: A Universal Framework for Compressed Image Super-Resolution using Dynamic Prompt

• Paper: https://arxiv.org/abs/2407.13108
• Code: https://lixinustc.github.io/UCIP.github.io
• Keywords: Compressed

6 OmniSSR: Zero-shot Omnidirectional Image Super-Resolution using Stable Diffusion Model

• Paper: https://arxiv.org/abs/2404.10312
• Code:
• Keywords: Omnidirectional

数据集

1 A New Dataset and Framework for Real-World Blurred Images Super-Resolution

• Paper: https://arxiv.org/abs/2407.14880
• Code: https://github.com/Imalne/PBaSR
• Keywords: Real-World, Blurred

总结

Diffusion 和 轻量化是本届ECCV 超分辨率方向接收较多的热点方向。

---

参考资料

1. ECCV 2024