Research map / Label-efficient and robust learning / Recognition and long-tail
Curricular Contrastive Regularization for Physics-Aware Single Image Dehazing
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 |
|---|---|---|---|
| Replace the PA module with PDU (ablation) | FFA-Net · SOTS-indoor | PSNR: 1.91dB | Helped |
| Replace the PA module with FDU (ablation) | FFA-Net · SOTS-indoor | PSNR: 36.39 -> 36.59 | Helped |
| Use PDU instead of FDU (ablation) | FFA-Net · SOTS-indoor | PSNR: 1.71dB | Helped |
| Add non-consensual contrastive regularization (1:10) (ablation) | FFA-Net + PDU · SOTS-indoor | PSNR: 38.30 -> 41.32 | Helped |
| Use consensual instead of non-consensual contrastive negatives (ablation) | FFA-Net + PDU · SOTS-indoor | PSNR: 0.77dB | Helped |
| Add consensual contrastive regularization (1:7) without curriculum (ablation) | FFA-Net + PDU · SOTS-indoor | PSNR: 3.79dB | Helped |
| Add the curriculum strategy to consensual contrastive regularization (ablation) | FFA-Net + PDU + consensual CR · SOTS-indoor | PSNR: 0.57dB | Helped |
| Apply C²R instead of non-consensual CR (1:10) (ablation) | Five dehazing baselines · SOTS-indoor | PSNR: 0.93-2.10dB | Helped |
| Apply C²R instead of self-paced curriculum learning (ablation) | Five dehazing baselines · SOTS-indoor | PSNR: 0.71dB | Helped |
| Use C²R rather than consensual CR (ablation) | GDN · SOTS-indoor | SSIM: -0.0012 | Hurt |
| Use C²PNet rather than UDN | SOTS-indoor · SOTS-indoor | PSNR: 3.94dB | Helped |
| Use C²PNet rather than the strongest baseline | SOTS-outdoor · SOTS-outdoor | PSNR: at least 1.50dB | Helped |
Closest papers by meaning
- Learning Hazing to Dehazing: Towards Realistic Haze Generation for Real-World Image Dehazing (CVPR 2025 · similarity 0.87)
- Exploiting Diffusion Prior for Real-World Image Dehazing with Unpaired Training (AAAI 2025 · similarity 0.87)
- PEIE: Physics Embedded Illumination Estimation for Adaptive Dehazing (AAAI 2025 · similarity 0.86)
- PRISM: Rethinking Atmospheric Scattering Reconstruction as a Unified Understanding and Restoration Model for Real-world Dehazing (NeurIPS 2026 · similarity 0.86)
- Contrastive Learning for Low-Light Raw Denoising (Student Abstract) (AAAI 2024 · similarity 0.85)
- ODCR: Orthogonal Decoupling Contrastive Regularization for Unpaired Image Dehazing (CVPR 2024 · similarity 0.85)
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.