Research map / Image generation / Image restoration
Densely Connected Pyramid Dehazing Network
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 multi-level pyramid pooling to the densely connected encoder-decoder (ablation) | DCPDN transmission estimator · TestA and TestB | SSIM (transmission): 0.9555 -> 0.9652; 0.9033 -> 0.9109 | Helped |
| Add multi-level pyramid pooling to the densely connected encoder-decoder (ablation) | DCPDN · TestA and TestB | SSIM (dehazed image): 0.9252 -> 0.9402; 0.8474 -> 0.8503 | Helped |
| Add gradient loss to L2 loss (ablation) | DCPDN transmission estimator with pyramid pooling · TestA and TestB | SSIM (transmission): 0.9652 -> 0.9687; 0.9109 -> 0.9239 | Helped |
| Add gradient loss to L2 loss (ablation) | DCPDN · TestA and TestB | SSIM (dehazed image): 0.9402 -> 0.9489; 0.8503 -> 0.8582 | Helped |
| Use edge-preserving loss instead of L2 plus gradient loss (ablation) | DCPDN transmission estimator with pyramid pooling · TestA and TestB | SSIM (transmission): 0.9687 -> 0.9732; 0.9239 -> 0.9276 | Helped |
| Use edge-preserving loss instead of L2 plus gradient loss (ablation) | DCPDN · TestA and TestB | SSIM (dehazed image): 0.9489 -> 0.9530; 0.8582 -> 0.8652 | Helped |
| Add a joint discriminator to the edge-preserving-loss configuration (ablation) | DCPDN · TestA and TestB | SSIM (transmission): 0.9732 -> 0.9776; 0.9276 -> 0.9352 | Helped |
| Add a joint discriminator to the edge-preserving-loss configuration (ablation) | DCPDN · TestA and TestB | SSIM (dehazed image): 0.9530 -> 0.9560; 0.8652 -> 0.8746 | Helped |
| Jointly estimate transmission, atmospheric light, and the dehazed image by embedding the scattering model | DCPDN · TestA | SSIM (dehazed image): 0.8842 -> 0.9560 | Helped |
| Jointly estimate transmission, atmospheric light, and the dehazed image by embedding the scattering model | DCPDN · TestB | SSIM (dehazed image): 0.8325 -> 0.8746 | Helped |
Closest papers by meaning
- Real-world Image Dehazing with Coherence-based Pseudo Labeling and Cooperative Unfolding Network (NeurIPS 2024 · similarity 0.88)
- PEIE: Physics Embedded Illumination Estimation for Adaptive Dehazing (AAAI 2025 · similarity 0.86)
- Depth-Centric Dehazing and Depth-Estimation from Real-World Hazy Driving Video (AAAI 2025 · similarity 0.85)
- Depth Information Assisted Collaborative Mutual Promotion Network for Single Image Dehazing (CVPR 2024 · similarity 0.85)
- PRISM: Rethinking Atmospheric Scattering Reconstruction as a Unified Understanding and Restoration Model for Real-world Dehazing (NeurIPS 2026 · similarity 0.84)
- Dehaze-RetinexGAN: Real-World Image Dehazing via Retinex-based Generative Adversarial Network (AAAI 2025 · similarity 0.84)
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.