Public

STEM-EDX Elemental Mapping — 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
absorption_correction_error -3.0 – 6.0 1.5 -
detector_solid_angle -0.15 – 0.15 0.0 sr
peak_overlap_(spectral) -0.6 – 1.2 0.3 -
bremsstrahlung_background -0.15 – 0.15 0.0 -

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

19.81 dB

SSIM 0.4840

Scenario II (Mismatch)

17.32 dB

SSIM 0.1061

Scenario III (Oracle)

19.62 dB

SSIM 0.2167

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 19.40 0.4889 16.96 0.1087 19.66 0.2303
scene_01 20.39 0.4977 17.60 0.1075 19.60 0.2162
scene_02 20.11 0.4612 17.67 0.0963 20.09 0.2040
scene_03 19.34 0.4883 17.06 0.1120 19.15 0.2163

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 PhysEDX + gradient 0.842 36.7 0.977 0.92 ✓ Certified Chen et al., Microsc. Microanal. 2024
2 DiffEDX + gradient 0.838 36.92 0.978 0.89 ✓ Certified Gao et al., NeurIPS 2024
3 SwinEDX + gradient 0.830 35.56 0.971 0.93 ✓ Certified Wang et al., npj Comput. Mater. 2023
4 TransEDX + gradient 0.809 33.93 0.96 0.93 ✓ Certified Li et al., Ultramicroscopy 2022
5 N2V-EDX + gradient 0.775 31.45 0.936 0.93 ✓ Certified Krull et al., NeurIPS 2019
6 DnCNN-EDX + gradient 0.710 28.1 0.883 0.88 ✓ Certified Kovarik et al., npj Comput. Mater. 2016
7 NMF-EDX + gradient 0.645 24.76 0.794 0.9 ✓ Certified Nicoletti et al., Nature 2013
8 TV-EDX + gradient 0.592 22.82 0.724 0.87 ✓ Certified Saghi et al., Ultramicroscopy 2011
9 MLS-EDX + gradient 0.508 19.53 0.576 0.91 ✓ Certified Statham, J. Anal. At. Spectrom. 1995

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%
Back to STEM-EDX Elemental Mapping