Research map / Detection and segmentation / CNN features
Beyond Monocular Deraining: Parallel Stereo Deraining Network Via Semantic Prior
Kaihao Zhang, Wenhan Luo, Yanjiang Yu, Wenqi Ren, Fang Zhao, Changsheng Li, Lin Ma, Wei Liu, Hongdong Li
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 semantic segmentation task to monocular deraining (ablation) | PRRNet(D) · RainKITTI2012 | PSNR: 30.71 -> 31.56 | Improved |
| add semantic-rethinking loop (ablation) | PRRNet(D+S) · RainKITTI2012 | PSNR: 31.56 -> 31.89 | Improved |
| add semantic-rethinking loop (ablation) | PRRNet(D+S) · RainKITTI2012 | SSIM: 0.928 -> 0.930 | Improved |
| use stereo images instead of monocular (ablation) | PRRNet · RainKITTI2012 | PSNR: 31.89 -> 33.01 | Improved |
| use enhanced version with semantic prior and parallel stereo network (monocular) (ablation) | PRRNet(monocular) · RainKITTI2012 | PSNR: 31.89 -> 32.38 | Improved |
| use enhanced version with semantic prior and parallel stereo network (stereo) (ablation) | PRRNet(stereo) · RainKITTI2012 | PSNR: 33.01 -> 34.13 | Improved |
| use enhanced version (stereo) (ablation) | PRRNet(stereo) · RainKITTI2012 | SSIM: 0.936 -> 0.947 | Improved |
| use enhanced version (stereo) (ablation) | PRRNet(stereo) · RainKITTI2015 | PSNR: 32.58 -> 33.83 | Improved |
| use EPRRNet stereo instead of best monocular baseline DeHRain | DeHRain · RainKITTI2012 | PSNR: 31.02 -> 34.13 | Improved |
| use EPRRNet monocular instead of DAF-Net | DAF-Net · RainCityscapes (monocular) | PSNR: 30.06 -> 31.11 | Improved |
| use EPRRNet monocular instead of DAF-Net | DAF-Net · RainCityscapes (monocular) | SSIM: 0.9530 -> 0.9741 | Improved |
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.
- RESCAN — Rescan: Inductive Instance Segmentation for Indoor RGBD Scans (ICCV 2019) · better on 1 of 1 shared result
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
- Beyond Monocular Deraining: Stereo Image Deraining via Semantic Understanding (ECCV 2020 · similarity 0.98)
- Close the Loop: A Unified Bottom-Up and Top-Down Paradigm for Joint Image Deraining and Segmentation (AAAI 2022 · similarity 0.91)
- Spatial Attentive Single-Image Deraining With a High Quality Real Rain Dataset (CVPR 2019 · similarity 0.90)
- Semantic Segmentation on Raindrop Degraded Images Using Two-Stage Dual Teacher-Student Learning (AAAI 2025 · similarity 0.90)
- Not Just Streaks: Towards Ground Truth for Single Image Deraining (ECCV 2022 · similarity 0.90)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · 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.
