Public
X-ray Crystallography — 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 |
|---|---|---|---|
| crystal_orientation_error | -0.1 – 0.2 | 0.05 | deg |
| beam_divergence | 0.08 – 0.14 | 0.11 | mrad |
| absorption_correction | 0.98 – 1.04 | 1.01 | - |
| scale_factor | 0.99 – 1.02 | 1.005 | - |
InverseNet Baseline Scores
Method: CPU_baseline — Mismatch parameter: nominal
Scenario I (Ideal)
22.46 dB
SSIM 0.5793
Scenario II (Mismatch)
21.44 dB
SSIM 0.5212
Scenario III (Oracle)
23.94 dB
SSIM 0.6362
Per-scene breakdown (4 scenes)
| Scene | PSNR I | SSIM I | PSNR II | SSIM II | PSNR III | SSIM III |
|---|---|---|---|---|---|---|
| scene_00 | 22.27 | 0.5891 | 21.75 | 0.5153 | 23.37 | 0.6088 |
| scene_01 | 23.10 | 0.5905 | 22.85 | 0.5697 | 25.69 | 0.6947 |
| scene_02 | 21.24 | 0.5088 | 20.63 | 0.4941 | 23.17 | 0.6135 |
| scene_03 | 23.23 | 0.6286 | 20.51 | 0.5056 | 23.53 | 0.6277 |
Public Tier Leaderboard
| # | Method | Score | PSNR | SSIM | Consistency | Trust | Source |
|---|---|---|---|---|---|---|---|
| 1 | CrystFormer + gradient | 0.745 | 29.87 | 0.915 | 0.9 | ✓ Certified | Crystallographic phasing transformer, 2024 |
| 2 | DL-Phase + gradient | 0.710 | 27.99 | 0.881 | 0.89 | ✓ Certified | Jumper et al., Nature 2021 |
| 3 | SHELXD + gradient | 0.651 | 24.74 | 0.794 | 0.93 | ✓ Certified | Sheldrick, Acta Cryst. D 2010 |
| 4 | Molecular Replacement + gradient | 0.497 | 19.24 | 0.562 | 0.9 | ✓ Certified | McCoy et al., J. Appl. Cryst. 2007 |
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%