Hidden
SAR — Hidden Tier
(3 scenes)Fully blind server-side evaluation — no data download.
What you get
No data downloadable. Algorithm runs server-side on hidden measurements.
How to use
Package algorithm as Docker container / Python script. Submit via link.
What to submit
Containerized algorithm accepting y + H, outputting x_hat + corrected spec.
Parameter Specifications
🔒
True spec hidden — blind evaluation, only ranges available.
| Parameter | Spec Range | Unit |
|---|---|---|
| motion_error | -1.4 – 4.6 | cm |
| phase_error | -0.21 – 0.69 | rad |
| range_cell_migration | -0.35 – 1.15 | cells |
Hidden Tier Leaderboard
| # | Method | Score | PSNR | SSIM | Consistency | Trust | Source |
|---|---|---|---|---|---|---|---|
| 1 | DiffusionSAR + gradient | 0.676 | 27.68 | 0.874 | 0.75 | ✓ Certified | Wei et al., NeurIPS 2024 |
| 2 | SARFormer + gradient | 0.626 | 24.88 | 0.798 | 0.79 | ✓ Certified | Li et al., CVPR 2024 |
| 3 | SAR-CAM + gradient | 0.624 | 24.86 | 0.798 | 0.78 | ✓ Certified | Cross-attention SAR, 2024 |
| 4 | Lee Filter + gradient | 0.607 | 24.07 | 0.771 | 0.79 | ✓ Certified | Lee, IEEE TGRS 1980 |
| 5 | Range-Doppler + gradient | 0.600 | 23.52 | 0.751 | 0.82 | ✓ Certified | SAR signal processing baseline |
| 6 | ScoreSAR + gradient | 0.582 | 23.23 | 0.74 | 0.77 | ✓ Certified | Johnson et al., ECCV 2025 |
| 7 | PanSharpener++ + gradient | 0.551 | 21.32 | 0.66 | 0.87 | ✓ Certified | Zhang et al., ICCV 2024 |
| 8 | SAR-DRN + gradient | 0.529 | 20.68 | 0.631 | 0.85 | ✓ Certified | Zhang et al., RS 2018 |
| 9 | SAR-BM3D + gradient | 0.516 | 20.28 | 0.612 | 0.84 | ✓ Certified | Parrilli et al., IEEE TGRS 2012 |
| 10 | SAR-ResNet + gradient | 0.508 | 20.7 | 0.632 | 0.74 | ✓ Certified | Chen et al., IEEE TGRS 2022 |
| 11 | SARDenoiserViT + gradient | 0.488 | 19.33 | 0.566 | 0.84 | ✓ Certified | Wang et al., ICCV 2024 |
| 12 | Chirp Scaling + gradient | 0.473 | 19.44 | 0.572 | 0.75 | ✓ Certified | Raney et al., IEEE TGRS 1994 |
| 13 | Matched Filter + gradient | 0.447 | 18.38 | 0.519 | 0.78 | ✓ Certified | Standard SAR focusing |
Dataset
Scenes: 3
Scoring Formula
0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)
PSNR: 40%
SSIM: 40%
Consistency: 20%