Research map / Image generation / Image restoration
From Sky to the Ground: A Large-scale Benchmark and Simple Baseline Towards Real Rain Removal
Yun Guo, Xueyao Xiao, Yi Chang, Shumin Deng, Luxin Yan
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 |
|---|---|---|---|
| add cross-layer attention between rain and image streams (ablation) | SCD-Former · LHP-Rain | PSNR / SSIM: 33.92 / 0.9384 -> 34.33 / 0.9403 | Improved |
| train deraining models on LHP-Rain instead of SPA-data when testing on LHP-Rain | SCD-Former · cross-dataset test on LHP-Rain | PSNR (dB): A->C 29.41 vs C->C 34.33 | Improved |
| train on other real datasets and test on LHP-Rain | deraining methods · B->C and A->C | PSNR (dB): Rainy 29.97 vs A->C best 29.41 | Hurt |
| train on LHP-Rain and test on SPA-data | SCD-Former · SPA-data | PSNR (dB): 32.60 (rainy) -> 34.38 | Improved |
| remove subspace projection from RLRTR | RLRTR · LHP-Rain GT generation | visual rain residual | Hurt |
| remove affine transformation alignment from RLRTR | RLRTR · LHP-Rain GT generation | visual quality | Hurt |
| apply SCD-Former deraining before lane segmentation | LaneNet · LHP-Rain | Acc: 0.237 -> 0.449 | Improved |
| apply SCD-Former deraining before object detection | YOLOv5 · LHP-Rain | mAP: 0.543 -> 0.575 | Improved |
| use RLRTR for static haze in heavy rain | RLRTR · LHP-Rain GT generation | haze removal | No clear effect |
Closest papers by meaning
- Not Just Streaks: Towards Ground Truth for Single Image Deraining (ECCV 2022 · similarity 0.93)
- Spatial Attentive Single-Image Deraining With a High Quality Real Rain Dataset (CVPR 2019 · similarity 0.93)
- Robust Video Content Alignment and Compensation for Rain Removal in a CNN Framework (CVPR 2018 · similarity 0.91)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.90)
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution (AAAI 2025 · similarity 0.90)
- Towards Universal Rainy Image Restoration: Benchmark and Baseline (AAAI 2025 · similarity 0.89)
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.
