Research map / Learning theory and algorithms / Matrix and tensor methods
A Statistical Experimental Design Method for Constructing Deterministic Sensing Matrices for Compressed Sensing
Youran Qi, Xu He, Tzu-Hsiang Hung, Peter Chien
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 |
|---|---|---|---|
| Use proposed back-elimination Walsh-Hadamard sub-matrix instead of DeVore's construction | Deterministic sensing matrix · Boat, Baboon, Peppers image reconstruction | PSNR (dB): 22.69 -> 26.42 dB (Boat); 18.79 -> 20.70 dB (Baboon); 25.88 -> 29.50 dB (Peppers) | Improved |
| Use proposed matrix instead of DeVore, Li, Yu-Zhao, Guo-Liu constructions | Deterministic sensing matrix · Coherence comparison (Tables 3-6) | coherence: 0.750 -> 0.375 (Li et al., r=2 s=3); 0.500 -> 0.298 (Yu and Zhao, q=7) | Improved |
| Use proposed matrix instead of DeVore's matrix (p=125 vs 128, n=25) | OMP reconstruction · Reconstruction probability, 5000 sparse signals per sparsity | reconstruction probability | Improved |
| Use proposed matrix instead of Guo and Liu's matrix (n=45) | OMP reconstruction · Reconstruction probability, p=252/256 | reconstruction probability | Improved |
| Use DeVore's deterministic matrix instead of random Gaussian | OMP reconstruction · Reconstruction probability, p=125 n=25 | reconstruction probability | No clear effect |
| Increase sampling rate n/p from 0.20 to 0.80 (Algorithm 1) | Proposed matrix, p=256 · Coherence vs n/p | coherence: 0.216 (n=51) -> 0.054 (n=205) | Improved |
| Increase number of columns p at n/p=0.50 | Proposed matrix · Coherence vs p | coherence: 0.250 (p=64) -> 0.156 (p=128) -> 0.109 (p=256) | Improved |
| Use weighted average Algorithm 2 instead of sequential Algorithm 1 (ablation) | Back-elimination · Coherence tables 1-2 | coherence: 0.263 -> 0.211 (p=128, n=38); 0.089 -> 0.111 (p=128, n=90) | Mixed |
| Use back-elimination column selection instead of random deletion (ablation) | Sub-Walsh Hadamard matrix · Minimum distance decrease | minimum distance of sub-code: decreasing rate of 1/2 per deletion asymptotically | Improved |
| Use proposed informed row selection instead of random UE(s2)-optimal row sampling (ablation) | Hadamard-based design · Coherence, 30 random UE(s2) designs | coherence | Improved |
Closest papers by meaning
- Two New Approaches to Compressed Sensing Exhibiting Both Robust Sparse Recovery and the Grouping Effect (JMLR 2017 · similarity 0.87)
- Hierarchical Multi-Stage Recovery Framework for Kronecker Compressed Sensing (ICLR 2026 · similarity 0.86)
- The Price of Sparsity: Sufficient Conditions for Sparse Recovery using Sparse and Sparsified Measurements (NeurIPS 2025 · similarity 0.86)
- Recovery Guarantees for Posterior Sampling of One-Bit Compressed Sensing (NeurIPS 2026 · similarity 0.86)
- Exponential Spectral Pursuit: An Effective Initialization Method for Sparse Phase Retrieval (ICML 2024 · similarity 0.84)
- Guarantees of a Preconditioned Subgradient Algorithm for Overparameterized Asymmetric Low-rank Matrix Recovery (ICML 2025 · 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.
