Research map / Image generation / Image restoration
LRDUN: A Low-Rank Deep Unfolding Network for Efficient Spectral Compressive Imaging
He Huang, Yujun Guo, Wei He
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 |
|---|---|---|---|
| replace full-HSI processing with low-rank basis/subspace decomposition (ablation) | Baseline-1 (3-stage) · KAIST simulation (3-stage ablation) | PSNR (dB): 38.16 -> 39.44 | Improved |
| replace full-HSI processing with low-rank decomposition (ablation) | Baseline-1 (3-stage) · KAIST simulation (3-stage ablation) | FLOPs (G): 26.95 -> 10.26 | Improved |
| add nuclear norm loss (ablation) | Baseline-1 · KAIST simulation (3-stage ablation) | PSNR (dB): 38.16 -> 37.66 | Hurt |
| add truncated SVD layer (ablation) | Baseline-1 · KAIST simulation (3-stage ablation) | PSNR (dB): 38.16 -> 37.93 | Hurt |
| add SP subspace prior (SP-1) (ablation) | Baseline-1 · KAIST simulation (3-stage ablation) | PSNR (dB): 38.16 -> 38.33 | Improved |
| add SP subspace prior (SP-2) (ablation) | Baseline-1 · KAIST simulation (3-stage ablation) | PSNR (dB): 38.16 -> 38.52 | Improved |
| increase feature dimension C beyond k (enable GFUM) (ablation) | LRDUN · KAIST simulation | PSNR (dB) | Improved |
| increase physical rank k with C=16 (ablation) | LRDUN · KAIST simulation | PSNR (dB) | Mixed |
| add SCAB module (ablation) | LRDUN-3stg · KAIST simulation (3-stage attention ablation) | PSNR (dB): 37.48 -> 39.44 | Improved |
| use window spatial attention instead of SCAB (ablation) | LRDUN-3stg · KAIST simulation (3-stage attention ablation) | PSNR (dB): 39.44 -> 39.13 | Hurt |
| use window spectral attention instead of SCAB (ablation) | LRDUN-3stg · KAIST simulation (3-stage attention ablation) | PSNR (dB): 39.44 -> 39.02 | Hurt |
| use HS-MSA instead of SCAB (ablation) | LRDUN-3stg · KAIST simulation (3-stage attention ablation) | PSNR (dB): 39.44 -> 39.31 | Hurt |
Closest papers by meaning
- Detail Matters: Mamba-Inspired Joint Unfolding Network for Snapshot Spectral Compressive Imaging (AAAI 2025 · similarity 0.91)
- Dual Prior Unfolding for Snapshot Compressive Imaging (CVPR 2024 · similarity 0.88)
- Untrained Neural Nets for Snapshot Compressive Imaging: Theory and Algorithms (NeurIPS 2024 · similarity 0.88)
- HUNet: Homotopy Unfolding Network for Image Compressive Sensing (CVPR 2025 · similarity 0.88)
- CPP-Net: Embracing Multi-Scale Feature Fusion into Deep Unfolding CP-PPA Network for Compressive Sensing (CVPR 2024 · similarity 0.88)
- Using Powerful Prior Knowledge of Diffusion Model in Deep Unfolding Networks for Image Compressive Sensing (CVPR 2025 · similarity 0.88)
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.
