Research map / Image generation / Image restoration
Single Image Reflection Separation via Component Synergy
Qiming Hu, Xiaojie Guo
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 learnable residue term with residual-corrected reconstruction loss (vs none) (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 24.84 -> 25.40 | Improved |
| use linear additive reconstruction loss instead of residual-corrected loss (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 25.40 -> 22.21 | Hurt |
| add feature interaction between the two streams (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 24.81 -> 25.40 | Improved |
| use MuGI instead of YTMT interaction (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 24.89 -> 25.40 | Improved |
| use DSFNet pyramid encoder instead of HyperColumn (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 23.99 -> 25.40 | Improved |
| add feature encoder (semantic pyramid) (ablation) | DSRNet · SIR2 + Real20 average | PSNR: 24.75 -> 25.40 | Improved |
| train with extra data (Nature dataset and 13,700 synthetic pairs) | DSRNet · SIR2 + Real20 average | PSNR: 25.40 -> 25.75 | Improved |
Closest papers by meaning
- Learning to Jointly Generate and Separate Reflections (ICCV 2019 · similarity 0.90)
- Trash or Treasure? An Interactive Dual-Stream Strategy for Single Image Reflection Separation (NeurIPS 2021 · similarity 0.89)
- Single Image Reflection Removal Beyond Linearity (CVPR 2019 · similarity 0.89)
- Single Image Reflection Separation via Dual-Stream Interactive Transformers (NeurIPS 2024 · similarity 0.87)
- Revisiting Single Image Reflection Removal In the Wild (CVPR 2024 · similarity 0.87)
- FIRM: Flexible Interactive Reflection ReMoval (AAAI 2025 · 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.
