Research map / Video / Video segmentation and tracking
Unsupervised Video Deraining with An Event Camera
Jin Wang, Wenming Weng, Yueyi Zhang, Zhiwei Xiong
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 event input to supervised deraining network (ablation) | Model#A · N-NTURain | PSNR: 38.41 -> 41.88 | Improved |
| add event input without cross-modal contrastive learning (unsupervised) (ablation) | Model#C · N-NTURain | PSNR: 29.45 -> 30.00 | No clear effect |
| add cross-modal contrastive learning to frame+event unsupervised network (ablation) | Model#D · N-NTURain | PSNR: 30.00 -> 37.30 | Improved |
| add inter-modal positive contrastive loss (ablation) | Full loss · N-NTURain | PSNR: 35.02 -> 37.30 | Improved |
| add inter-modal negative contrastive loss (ablation) | Full loss · N-NTURain | PSNR: 35.04 -> 37.30 | Improved |
| add event-domain intra-modal contrastive loss (ablation) | Full loss · N-NTURain | PSNR: 37.18 -> 37.30 | Improved |
| use asymmetric separation (encoder in frame, encoder+attention in event) instead of symmetric (ablation) | Separation module · N-NTURain | PSNR: 35.68 -> 37.30 | Improved |
| add cross-modal fusion with enhancement and negative suppression (ablation) | Deraining network · N-NTURain | PSNR: 34.32 -> 37.30 | Improved |
| also subtract the frame background feature in fusion (ablation) | Cross-modal fusion · N-NTURain | PSNR: 37.30 -> 35.40 | Hurt |
| use event-based unsupervised method instead of prior unsupervised methods | Unsupervised video deraining · N-NTURain | PSNR: 32.74 (SLDNet) -> 37.30 | Improved |
Closest papers by meaning
- Rethinking Video Rain Streak Removal: A New Synthesis Model and a Deraining Network with Video Rain Prior (ECCV 2022 · similarity 0.90)
- Semi-Supervised Video Deraining With Dynamical Rain Generator (CVPR 2021 · similarity 0.90)
- PRE-Mamba: A 4D State Space Model for Ultra-High-Frequent Event Camera Deraining (ICCV 2025 · similarity 0.90)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.90)
- Semi-Supervised State-Space Model with Dynamic Stacking Filter for Real-World Video Deraining (CVPR 2025 · similarity 0.88)
- NightRain: Nighttime Video Deraining via Adaptive-Rain-Removal and Adaptive-Correction (AAAI 2024 · similarity 0.87)
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.
