Research map / Architectures and efficiency / Attention and transformers
SmartAssign: Learning a Smart Knowledge Assignment Strategy for Deraining and Desnowing
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 gated knowledge filtering and task-targeted knowledge assignment rather than using conventional MTL (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 32.11 -> 33.07; 31.93 -> 32.50 | Helped |
| Add knowledge contrast (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 32.34 -> 33.07; 31.73 -> 32.50 | Helped |
| Add the gated CNN module to GLSF-ViT (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 32.46 -> 33.07; 32.03 -> 32.50 | Helped |
| Add joint rain-and-snow training instead of single-task training (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 32.19 -> 33.07; 32.20 -> 32.50 | Helped |
| Add the auxiliary rain/snow recognition task (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 33.01 -> 33.07; 32.41 -> 32.50 | Helped |
| Use both unique and shared knowledge rather than common knowledge alone (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 30.11 -> 33.07; 29.86 -> 32.50 | Helped |
| Use both unique and shared knowledge rather than unique knowledge alone (ablation) | SmartAssign · Rain1200 and CSD | PSNR: 31.43 -> 33.07; 31.65 -> 32.50 | Helped |
| Use SmartAssign for deraining and desnowing | SmartAssign · Rain-streak, Rain1200, Real-rain, Snow100K-L, SRRS, and CSD | PSNR: 33.01 -> 33.16; 32.91 -> 33.07; 36.04 -> 36.55; 25.66 -> 29.45; 28.03 -> 30.53; 30.10 -> 32.50 | Helped |
Closest papers by meaning
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.86)
- WeatherGFM: Learning a Weather Generalist Foundation Model via In-context Learning (ICLR 2025 · similarity 0.86)
- Semantic Segmentation in Multiple Adverse Weather Conditions with Domain Knowledge Retention (AAAI 2024 · similarity 0.84)
- Bidirectional Multi-Scale Implicit Neural Representations for Image Deraining (CVPR 2024 · similarity 0.83)
- Language-driven All-in-one Adverse Weather Removal (CVPR 2024 · similarity 0.83)
- DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer (NeurIPS 2025 · similarity 0.83)
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.