Research map / Image generation / Editing and style transfer
Adverse Weather Removal with Codebook Priors
Tian Ye, Sixiang Chen, Jinbin Bai, Jun Shi, Chenghao Xue, Jingxia Jiang, Junjie Yin, Erkang Chen, Yun Liu
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use parallel decoder (new Dd plus frozen VQGAN decoder branch) instead of Dd only (ablation) | AWRCP · Outdoor-Rain | PSNR: 28.12 -> 31.39 | Improved |
| use parallel decoder instead of fixed pretrained VQGAN decoder (ablation) | AWRCP · Outdoor-Rain | PSNR: 23.01 -> 31.39 | Improved |
| use parallel decoder instead of finetuning pretrained VQGAN decoder (ablation) | AWRCP · Outdoor-Rain | PSNR: 27.23 -> 31.39 | Improved |
| use deformable cross-attention with sparse sampling instead of window cross-attention (ablation) | AWRCP · Outdoor-Rain | PSNR: 30.21 -> 31.39 | Improved |
| use deformable cross-attention instead of SFT fusion (ablation) | AWRCP · Outdoor-Rain | PSNR: 30.12 -> 31.39 | Improved |
| add hierarchical texture warping head (ablation) | AWRCP · Outdoor-Rain | PSNR: 30.91 -> 31.39 | Improved |
| add deformable convolution in texture warping module (ablation) | AWRCP · Outdoor-Rain | PSNR: 31.21 -> 31.39 | Improved |
| add channel and spatial attention in texture warping module (ablation) | AWRCP · Outdoor-Rain | PSNR / SSIM: PSNR 31.44 -> 31.39; SSIM 0.9312 -> 0.9329 | Mixed |
| unified AWRCP vs diffusion-based WeatherDiff64 on raindrop SSIM | AWRCP · RainDrop | SSIM: 0.9313 (IDT) vs 0.9314 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Continuous Adverse Weather Removal via Degradation-Aware Distillation (CVPR 2025 · similarity 0.87)
- AWRaCLe: All-Weather Image Restoration Using Visual In-Context Learning (AAAI 2025 · similarity 0.87)
- MODEM: A Morton-Order Degradation Estimation Mechanism for Adverse Weather Image Recovery (NeurIPS 2025 · similarity 0.84)
- Real-World Adverse Weather Image Restoration via Dual-Level Reinforcement Learning with High-Quality Cold Start (NeurIPS 2025 · similarity 0.84)
- Learning Real-World Image De-weathering with Imperfect Supervision (AAAI 2024 · similarity 0.84)
- Multi-axis Prompt and Multi-dimension Fusion Network for All-in-one Weather-degraded Image Restoration (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
