Research map / Image generation / Image restoration
Single Image Deraining: From Model-Based to Data-Driven and Beyond
Wenhan Yang, Robby T. Tan, Shiqi Wang, Yuming Fang, Jiaying Liu
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| compare deep-learning deraining (DDN) against model-based methods | DDN · Rain100L, Rain100H, Rain1400 | PSNR/SSIM: more than 3dB, 7dB, and 0.7dB on Rain100L, Rain100H, and Rain1400 | Improved |
| compare newest deep methods RESCAN, PReNet, JORDER-E | RESCAN, PReNet, JORDER-E · Synthetic deraining benchmarks | PSNR/SSIM: performance gaps are considerably close | No clear effect |
| use non-reference metrics to predict human preference | NIQE, PIQE, BRISQUE, etc. · Real rain images (20 images) | SROCC: 0.2216, 0.1473, 0.1864 for SROCC, KROCC, and PLCC | Hurt |
| add subjective MOS ranking vs quantitative metric ranking | 2019 deraining methods · Real rain images (20 images) | MOS: JORDER-E average MOS 3.1366 vs PReNet 2.3252 | Mixed |
| compare PReNet against heavier networks in size and speed | PReNet · 512x512 input | time and parameters: 0.11 seconds, 168,963 parameters vs HeavyRainRemoval 0.73 seconds, 40,627,038 | Improved |
| use GPU deep network instead of CPU sparse-coding method | JORDER-E · 512x512 input | time (seconds): 0.13 vs LP 1177.3 | Improved |
Closest papers by meaning
- Single Image Deraining: A Comprehensive Benchmark Analysis (CVPR 2019 · similarity 0.94)
- Not Just Streaks: Towards Ground Truth for Single Image Deraining (ECCV 2022 · similarity 0.93)
- EfficientDeRain: Learning Pixel-wise Dilation Filtering for High-Efficiency Single-Image Deraining (AAAI 2021 · similarity 0.93)
- UniRain: Unified Image Deraining with RAG-based Dataset Distillation and Multi-objective Reweighted Optimization (CVPR 2026 · similarity 0.92)
- Recurrent Multi-Frame Deraining: Combining Physics Guidance and Adversarial Learning (TPAMI 2021 · similarity 0.92)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.91)
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.
