Public
TEM — Public Tier
(3 scenes)Full-access development tier with all data visible.
What you get
Measurements (y), ideal forward operator (H), spec ranges, ground truth (x_true), and true mismatch spec.
How to use
Load HDF5 → compare reconstruction vs x_true → check consistency → iterate.
What to submit
Reconstructed signals (x_hat) and corrected spec as HDF5.
Parameter Specifications
✓
True spec visible — use these exact values for Scenario III oracle reconstruction.
| Parameter | Spec Range | True Value | Unit |
|---|---|---|---|
| defocus | -50.0 – 100.0 | 25.0 | nm |
| Cs | -0.01 – 0.02 | 0.005 | mm |
| beam_tilt | -0.5 – 1.0 | 0.25 | mrad |
InverseNet Baseline Scores
Method: CPU_baseline — Mismatch parameter: nominal
Scenario I (Ideal)
17.98 dB
SSIM 0.2184
Scenario II (Mismatch)
9.14 dB
SSIM 0.1941
Scenario III (Oracle)
14.57 dB
SSIM 0.1284
Per-scene breakdown (4 scenes)
| Scene | PSNR I | SSIM I | PSNR II | SSIM II | PSNR III | SSIM III |
|---|---|---|---|---|---|---|
| scene_00 | 19.63 | 0.3153 | 9.25 | 0.1281 | 14.96 | 0.0996 |
| scene_01 | 17.05 | 0.2055 | 7.74 | 0.1774 | 17.90 | 0.2570 |
| scene_02 | 18.00 | 0.2280 | 12.48 | 0.4695 | 7.30 | 0.0359 |
| scene_03 | 17.23 | 0.1248 | 7.10 | 0.0014 | 18.11 | 0.1212 |
Public Tier Leaderboard
| # | Method | Score | PSNR | SSIM | Consistency | Trust | Source |
|---|---|---|---|---|---|---|---|
| 1 | SwinIR + gradient | 0.787 | 32.38 | 0.947 | 0.92 | ✓ Certified | Liang et al., ICCVW 2021 |
| 2 | Noise2Void + gradient | 0.756 | 30.09 | 0.918 | 0.94 | ✓ Certified | Krull et al., CVPR 2019 |
| 3 | BM3D + gradient | 0.677 | 26.39 | 0.843 | 0.88 | ✓ Certified | Dabov et al., IEEE TIP 2007 |
| 4 | Wiener Filter + gradient | 0.582 | 22.3 | 0.703 | 0.89 | ✓ Certified | Analytical baseline |
Visible Data Fields
y
H_ideal
spec_ranges
x_true
true_spec
Dataset
Format: HDF5
Scenes: 3
Scoring Formula
0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)
PSNR: 40%
SSIM: 40%
Consistency: 20%