Research map / Image generation / Fast diffusion sampling
Adaptive Compressed Sensing with Diffusion-Based Posterior Sampling
Noam Elata, Tomer Michaeli, Michael Elad
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 |
|---|---|---|---|
| choose measurements adaptively with AdaSense instead of optimal non-adaptive PCA | DDRM restoration · CelebA-HQ 256, rank 192 | PSNR: 24.60 -> 26.20 | Improved |
| replace final DDRM sample with mean of posterior samples (ablation) | AdaSense · CelebA-HQ 256 | PSNR / LPIPS: PSNR 26.20 -> 26.87, LPIPS 0.1950 -> 0.2449 | Mixed |
| increase number of adaptive iterations N at fixed total measurements (ablation) | AdaSense · CelebA-HQ 256 | PSNR and LPIPS | Improved |
| use AdaSense-selected vertical columns instead of random or equi-spaced masks | FastMRI vertical R10 with DDRM · FastMRI knee | PSNR: 24.73 -> 27.01 | Improved |
| use AdaSense instead of Poisson disk mask with wavelet reconstruction at R400 | FastMRI general subsampling · FastMRI knee R400 Wavelet | PSNR: 18.22 -> 23.32 | Improved |
| AdaSense versus trained active acquisition methods | FastMRI R16-2L · FastMRI knee | PSNR: 27.70 (SS-DDQN) vs 27.51 | No clear effect |
| use many more posterior samples s beyond about 4/3 of r (ablation) | AdaSense · CelebA-HQ 256 (N=1, r=192) | restoration quality | No clear effect |
Closest papers by meaning
- Sampling Innovation-Based Adaptive Compressive Sensing (CVPR 2025 · similarity 0.87)
- Recovery Guarantees for Posterior Sampling of One-Bit Compressed Sensing (NeurIPS 2026 · similarity 0.87)
- Adapt and Diffuse: Sample-adaptive Reconstruction via Latent Diffusion Models (ICML 2024 · similarity 0.87)
- A Unified Model for Compressed Sensing MRI Across Undersampling Patterns (CVPR 2025 · similarity 0.87)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.87)
- Reconstruction-free Cascaded Adaptive Compressive Sensing (CVPR 2024 · similarity 0.87)
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.
