Public
Cathodoluminescence (CL) Imaging — Public Tier
(5 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 |
|---|---|---|---|
| beam_current_drift | -1.0 – 2.0 | 0.5 | - |
| collection_efficiency_variation | -4.0 – 8.0 | 2.0 | spatial |
| spectral_calibration_error | -0.4 – 0.8 | 0.2 | nm |
| carbon_contamination | -2.0 – 4.0 | 1.0 | signalloss |
InverseNet Baseline Scores
Method: CPU_baseline — Mismatch parameter: nominal
Scenario I (Ideal)
33.04 dB
SSIM 0.8313
Scenario II (Mismatch)
22.80 dB
SSIM 0.4030
Scenario III (Oracle)
24.56 dB
SSIM 0.4982
Per-scene breakdown (4 scenes)
| Scene | PSNR I | SSIM I | PSNR II | SSIM II | PSNR III | SSIM III |
|---|---|---|---|---|---|---|
| scene_00 | 32.90 | 0.8330 | 22.40 | 0.4133 | 24.52 | 0.5154 |
| scene_01 | 32.93 | 0.8361 | 22.59 | 0.4422 | 24.50 | 0.5349 |
| scene_02 | 33.03 | 0.8183 | 23.28 | 0.3763 | 24.65 | 0.4644 |
| scene_03 | 33.32 | 0.8377 | 22.94 | 0.3802 | 24.59 | 0.4779 |
Public Tier Leaderboard
| # | Method | Score | PSNR | SSIM | Consistency | Trust | Source |
|---|---|---|---|---|---|---|---|
| 1 | Restormer-CL + gradient | 0.847 | 36.66 | 0.977 | 0.95 | ✓ Certified | Zamir et al., CVPR 2022 (CL adapted) |
| 2 | DiffusionEM + gradient | 0.843 | 37.51 | 0.98 | 0.88 | ✓ Certified | Gao et al., Nat. Methods 2024 (EM adapted) |
| 3 | PINN-CL + gradient | 0.828 | 35.09 | 0.968 | 0.95 | ✓ Certified | Raissi et al., J. Comput. Phys. 2019 (CL) |
| 4 | SwinIR-CL + gradient | 0.810 | 34.79 | 0.966 | 0.88 | ✓ Certified | Liang et al., ICCV 2021 (CL adapted) |
| 5 | U-Net-CL + gradient | 0.796 | 32.94 | 0.952 | 0.93 | ✓ Certified | Ronneberger et al., MICCAI 2015 (CL adapted) |
| 6 | CARE-CL + gradient | 0.792 | 33.34 | 0.956 | 0.88 | ✓ Certified | Weigert et al., Nat. Methods 2018 (CL adapted) |
| 7 | DnCNN-CL + gradient | 0.731 | 28.92 | 0.899 | 0.91 | ✓ Certified | Zhang et al., IEEE TIP 2017 (CL adapted) |
| 8 | Richardson-Lucy + gradient | 0.680 | 25.95 | 0.831 | 0.94 | ✓ Certified | Richardson, J. Opt. Soc. Am. 1972 |
| 9 | Wiener-CL + gradient | 0.593 | 22.73 | 0.72 | 0.89 | ✓ Certified | Castleman, Digital Image Processing, 1996 |
Visible Data Fields
y
H_ideal
spec_ranges
x_true
true_spec
Dataset
Format: HDF5
Scenes: 5
Scoring Formula
0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)
PSNR: 40%
SSIM: 40%
Consistency: 20%