Research map / Image generation / Image restoration
Single Image Reflection Removal With Physically-Based Training Images
Check 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 a priori loss using BT-net backtracking (ablation) | SP-net · SIR Wild | PSNR: 24.31 -> 25.54 | Helped |
| Add a priori loss using BT-net backtracking (ablation) | SP-net · SIR Wild | SSIM: 0.874 -> 0.905 | Helped |
| Add a priori loss using BT-net backtracking (ablation) | SP-net · Real 100 | PSNR: 20.86 -> 21.58 | Helped |
| Add a priori loss using BT-net backtracking (ablation) | SP-net · Real 100 | SSIM: 0.772 -> 0.789 | Helped |
| Add a priori loss using BT-net backtracking (ablation) | SP-net · Rendered Testset | PSNR: 27.34 -> 27.90 | Helped |
| Add a priori loss using BT-net backtracking (ablation) | SP-net · Rendered Testset | SSIM: 0.889 -> 0.894 | Helped |
| Use physically rendered training data instead of Wen-generated data to fine-tune CEILNet | CEILNet · SIR Wild | PSNR: 19.23 -> 22.51 | Helped |
| Use physically rendered training data instead of Wen-generated data to fine-tune Zhang | Zhang · SIR Wild | PSNR: 21.34 -> 23.18 | Helped |
| Use physically rendered training data instead of Wen-generated data to fine-tune CEILNet | CEILNet · Rendered Testset | PSNR: 20.23 -> 23.46 | Helped |
| Use physically rendered training data instead of Wen-generated data to fine-tune Zhang | Zhang · Rendered Testset | PSNR: 21.83 -> 24.43 | Helped |
| Apply reflection removal rather than leave the input unchanged | Ours · SIR Wild | PSNR: 25.89 -> 25.55 | Hurt |
| Apply reflection removal rather than leave the input unchanged | Ours · Real 100 | SSIM: 0.797 -> 0.789 | Hurt |
Closest papers by meaning
- Single Image Reflection Removal Beyond Linearity (CVPR 2019 · similarity 0.89)
- Flash-Split: 2D Reflection Removal with Flash Cues and Latent Diffusion Separation (CVPR 2025 · similarity 0.86)
- FIRM: Flexible Interactive Reflection ReMoval (AAAI 2025 · similarity 0.86)
- Multi-View Dynamic Reflection Prior for Video Glass Surface Detection (AAAI 2024 · similarity 0.85)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 2024 · similarity 0.84)
- Revisiting Single Image Reflection Removal In the Wild (CVPR 2024 · 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.