Research map / Image generation / Image restoration
Nonlocal Low-Rank Tensor Factor Analysis for Image Restoration
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 NLR-TFA instead of TVAL3 | image CS reconstruction · 8-image average, pseudo-radial CSr 0.02 | PSNR (dB): 21.15 -> 28.77 | Helped |
| use NLR-TFA instead of NLR-CS | image CS reconstruction · 8-image average, pseudo-radial CSr 0.02 | PSNR (dB): 17.28 -> 28.77 | Helped |
| use NLR-TFA instead of D-AMP | image CS reconstruction · 8-image average, pseudo-radial CSr 0.02 | PSNR (dB): 16.29 -> 28.77 | Helped |
| use NLR-TFA instead of BM3D-CS | image CS reconstruction · 8-image average, pseudo-radial CSr 0.02 | PSNR (dB): 16.80 -> 28.77 | Helped |
| use NLR-TFA instead of D-AMP | image CS reconstruction with Gaussian sensing · 8-image average, CSr 0.02 | PSNR (dB): 16.99 -> 25.04 | Helped |
| increase patch size from 4×4 to 6×6 (ablation) | NLR-TFA · Foreman, pseudo-radial CSr 0.02 | PSNR (dB): 34.01 -> 31.16 | Hurt |
| increase patch size from 4×4 to 8×8 (ablation) | NLR-TFA · Foreman, pseudo-radial CSr 0.02 | PSNR (dB): 34.01 -> 29.63 | Hurt |
| increase patch size from 4×4 to 10×10 (ablation) | NLR-TFA · Foreman, pseudo-radial CSr 0.02 | PSNR (dB): 34.01 -> 28.57 | Hurt |
| increase patch size from 4×4 to 12×12 (ablation) | NLR-TFA · Foreman, pseudo-radial CSr 0.02 | PSNR (dB): 34.01 -> 27.39 | Hurt |
| use low-rank tensor factorization of grouped patches within ADMM | NLR-TFA · noiseless pseudo-radial CS, CSr < 0.1 | PSNR (dB): +15.04 dB, +10.65 dB, +14.36 dB, and +14.58 dB | Helped |
| use rank-3 rather than rank-8 patch-group tensors (ablation) | low-rank tensor reconstruction · clean image tensor aggregation | PSNR (dB): 36.54 dB -> 37.09 dB | No clear effect |
| use pretrained convolutional inversion approximation instead of conjugate-gradient inner-loop updates | NLR-TFA with Gaussian sensing · random Gaussian CS | runtime | Helped |
Closest papers by meaning
- Robust High-Order Tensor Compressive Sensing Based on M-Estimators (AAAI 2026 · similarity 0.89)
- Non-Convex Tensor Recovery from Local Measurements (AAAI 2025 · similarity 0.89)
- Hyper-Laplacian Regularized Unidirectional Low-Rank Tensor Recovery for Multispectral Image Denoising (CVPR 2017 · similarity 0.88)
- Deep Rank-One Tensor Functional Factorization for Multi-Dimensional Data Recovery (AAAI 2025 · similarity 0.88)
- Low-Rank Tensor Completion With a New Tensor Nuclear Norm Induced by Invertible Linear Transforms (CVPR 2019 · similarity 0.87)
- OTLRM: Orthogonal Learning-based Low-Rank Metric for Multi-Dimensional Inverse Problems (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.