Research map / Image generation / Image restoration
Rethinking the CSC Model for Natural Images
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 |
|---|---|---|---|
| Use a strided CSC approximation with q=10 rather than standard CSC with q=1 (ablation) | CSC denoising · Set12, σ=15 | PSNR: 28.99 -> 31.46 | Helped |
| Use a strided CSC approximation with q=10 rather than standard CSC with q=1 (ablation) | CSC denoising · Set12, σ=25 | PSNR: 25.78 -> 28.88 | Helped |
| Use a strided CSC approximation with q=10 rather than standard CSC with q=1 (ablation) | CSC denoising · Set12, σ=50 | PSNR: 21.49 -> 25.56 | Helped |
| Use a strided CSC approximation with q=10 rather than standard CSC with q=1 (ablation) | CSC denoising · Set12, σ=75 | PSNR: 18.83 -> 23.66 | Helped |
| Use stride q=10 instead of patch averaging | CSC denoising · Set12, σ=15 | PSNR: 31.23 -> 31.46 | Helped |
| Use stride q=10 instead of patch averaging | CSC denoising · Set12, σ=25 | PSNR: 28.73 -> 28.88 | Helped |
| Use stride q=10 instead of patch averaging | CSC denoising · Set12, σ=50 | PSNR: 25.32 -> 25.56 | Helped |
| Use stride q=10 instead of patch averaging | CSC denoising · Set12, σ=75 | PSNR: 23.28 -> 23.66 | Helped |
| Use intermediate stride q=7 or q=8 rather than q=1 or q=11 (ablation) | CSCNet · BSD68, σ=25 | PSNR: 28.74; 28.9; 29.11 | Helped |
| Use CSCNet rather than DnCNN | grayscale denoising · BSD68, σ=15 and σ=25 | PSNR: 31.72 -> 31.57; 29.22 -> 29.11 | Hurt |
| Use CSCNet rather than FFDNet | grayscale denoising · BSD68, σ=50 and σ=75 | PSNR: 26.29 -> 26.24; 24.79 -> 24.77 | Hurt |
| Use CSCNet rather than the strongest listed baseline | color denoising · color-BSD68 | PSNR: 33.89 -> 33.83; 31.23 -> 31.18; 27.96 -> 28.00; 26.24 -> 26.32 | Mixed |
Closest papers by meaning
- A Local Block Coordinate Descent Algorithm for the CSC Model (CVPR 2019 · similarity 0.90)
- Beyond Matryoshka: Revisiting Sparse Coding for Adaptive Representation (ICML 2025 · similarity 0.86)
- SparseFormer: Sparse Visual Recognition via Limited Latent Tokens (ICLR 2024 · similarity 0.85)
- From Flat to Hierarchical: Extracting Sparse Representations with Matching Pursuit (NeurIPS 2025 · similarity 0.85)
- Sparse Bayesian Generative Modeling for Compressive Sensing (NeurIPS 2024 · similarity 0.85)
- Patch Rebirth: Fast and Transferable Model Inversion of Vision Transformers (NeurIPS 2026 · 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.