Research map / Image generation / Image restoration
Deep Learning for Seeing Through Window With Raindrops
Yuhui Quan, Shijie Deng, Yixin Chen, Hui Ji
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 the JPCA module (PA and CA) (ablation) | Base-1 · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 29.91 -> 30.86 | Improved |
| add the JPCA module (PA and CA) (ablation) | Base-1 · Qian et al. raindrop test set (58 pairs) | SSIM: 0.8967 -> 0.9263 | Improved |
| add the physical shape attention (PA) module on top of CA (ablation) | Base-3 · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 30.58 -> 30.86 | Improved |
| add the physical shape attention (PA) module on top of CA (ablation) | Base-3 · Qian et al. raindrop test set (58 pairs) | SSIM: 0.9224 -> 0.9263 | Improved |
| add the channel attention (CA) module on top of PA (ablation) | Base-2 · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 30.08 -> 30.86 | Improved |
| add the channel attention (CA) module on top of PA (ablation) | Base-2 · Qian et al. raindrop test set (58 pairs) | SSIM: 0.9154 -> 0.9263 | Improved |
| add PA module alone to the network without attention (ablation) | Base-1 · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 29.91 -> 30.08 | Improved |
| add CA module alone to the network without attention (ablation) | Base-1 · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 29.91 -> 30.58 | Improved |
| make the elliptic shape parameters learnable instead of fixed to ones (ablation) | Ours · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 0.21dB | Improved |
| use proposed attention CNN instead of AttentGAN | AttentGAN · Qian et al. raindrop test set (58 pairs) | PSNR (dB): 30.55 -> 30.86 | Improved |
| use proposed attention CNN instead of AttentGAN | AttentGAN · Qian et al. raindrop test set (58 pairs) | PSNR on luminance (dB): 31.57 -> 31.44 | Hurt |
| use proposed attention CNN instead of AttentGAN | AttentGAN · Qian et al. raindrop test set (58 pairs) | SSIM: 0.9023 -> 0.9263 | Improved |
Closest papers by meaning
- Depth-Attentional Features for Single-Image Rain Removal (CVPR 2019 · similarity 0.89)
- Removing Rain From Single Images via a Deep Detail Network (CVPR 2017 · similarity 0.89)
- A Model-Driven Deep Neural Network for Single Image Rain Removal (CVPR 2020 · similarity 0.89)
- Semantic Segmentation on Raindrop Degraded Images Using Two-Stage Dual Teacher-Student Learning (AAAI 2025 · similarity 0.88)
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution (AAAI 2025 · similarity 0.87)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.86)
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.
