Research map / Image generation / Super-resolution
Super Resolved Imaging with Adaptive Optics
Robin Swanson, Esther Y. H. Lin, Masen Lamb, Suresh Sivanandam, Kiriakos N. Kutulakos
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 |
|---|---|---|---|
| jointly optimize deformable-mirror phase profiles with the CNN instead of training the CNN with no learned phase (ablation) | EDSR-based MISR · Simulated AO SR, 2x, N=4 | PSNR (dB): 33.72 -> 43.13 | Improved |
| learn per-exposure phases (Non-Modal) rather than classic tip/tilt raster shifts (ablation) | CNN reconstruction · Simulated AO SR, 2x, N=4 | PSNR (dB): 39.56 -> 43.13 | Improved |
| optimize tip/tilt shift values instead of fixed raster scan (ablation) | CNN reconstruction · Simulated AO SR, 2x, N=4 | PSNR (dB): 39.56 -> 39.67 | Improved |
| use Zernike modes instead of tip/tilt only (ablation) | CNN reconstruction · Simulated AO SR, 8x, N=4 | PSNR (dB): 25.47 -> 27.74 | Improved |
| replace CNN reconstruction with linear shift-and-add reconstruction (ablation) | learned phases · Simulated AO SR, 2x, N=4 | PSNR (dB): 43.13 -> 31.08 | Hurt |
| increase number of exposures from 2 to 8 (ablation) | CNN reconstruction with Non-Modal phase · Simulated AO SR, 4x | PSNR (dB): 31.17 -> 37.48 | Improved |
| increase number of exposures for linear reconstruction (ablation) | linear reconstruction · Simulated AO SR, 4x, Tip/Tilt | PSNR (dB): 26.11 -> 26.11 | No clear effect |
| learned phases with linear reconstruction at 8x (ablation) | linear reconstruction · Simulated AO SR, 8x, N=4 | PSNR (dB): 23.32 -> 23.33 | No clear effect |
| Zernike versus Non-Modal phase representation at higher exposure counts (ablation) | CNN reconstruction · Simulated AO SR, 4x, N=6 | PSNR (dB): 36.29 vs 36.24 | No clear effect |
Closest papers by meaning
- FluxFlow: Conservative Flow-Matching for Astronomical Image Super-Resolution (NeurIPS 2026 · similarity 0.84)
- STAR: A Benchmark for Astronomical Star Fields Super-Resolution (NeurIPS 2025 · similarity 0.82)
- Broadband Wide Field of View Imaging with Computational Mirrors (ECCV 2026 · similarity 0.82)
- A New Statistical Model of Star Speckles for Learning to Detect and Characterize Exoplanets in Direct Imaging Observations (CVPR 2025 · similarity 0.81)
- WaveMo: Learning Wavefront Modulations to See Through Scattering (CVPR 2024 · similarity 0.79)
- Continuous Optical Zooming: A Benchmark for Arbitrary-Scale Image Super-Resolution in Real World (CVPR 2024 · similarity 0.79)
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.
