X-ray Fluorescence (XRF) Imaging

X-ray Fluorescence (XRF) Imaging

Join the Competition Contribute to Benchmark

Standard reconstruction benchmark — forward model perfectly known, no calibration needed. Score = 0.5 × clip((PSNR−15)/30, 0, 1) + 0.5 × SSIM

# Method Score PSNR (dB) SSIM Source
🥇 SpectraFormer 0.784 34.0 0.935 ✓ Certified Spectral unmixing transformer, 2024
🥈 XRF-UNet 0.733 32.0 0.900 ✓ Certified Anunziata et al., 2022
🥉 PnP-BM3D 0.617 28.0 0.800 ✓ Certified Danielyan et al., 2012
4 FP-Quantify 0.498 24.5 0.680 ✓ Certified Sole et al., 2007

Dataset: PWM Benchmark (4 algorithms)

Blind Reconstruction Challenge — forward model has unknown mismatch, must calibrate from data. Score = 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)

# Method Overall Score Public
PSNR / SSIM
 Dev
PSNR / SSIM
 Hidden
PSNR / SSIM
 Trust Source
🥇 SpectraFormer + gradient 0.739
0.769
31.43 dB / 0.936
0.745
29.86 dB / 0.915
0.703
28.71 dB / 0.895
✓ Certified Spectral unmixing transformer, 2024
🥈 XRF-UNet + gradient 0.645
0.733
29.06 dB / 0.901
0.653
25.53 dB / 0.818
0.550
21.5 dB / 0.668
✓ Certified Anunziata et al., X-Ray Spectrom. 2022
🥉 PnP-BM3D + gradient 0.590
0.666
25.98 dB / 0.831
0.564
22.37 dB / 0.706
0.539
21.38 dB / 0.663
✓ Certified Danielyan et al., 2012
4 FP-Quantify + gradient 0.552
0.571
21.82 dB / 0.682
0.574
22.29 dB / 0.702
0.512
20.63 dB / 0.629
✓ Certified Sole et al., Spectrochim. Acta B 2007

Complete score requires all 3 tiers (Public + Dev + Hidden).

Join the competition →
Scoring: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖) PSNR 40% · SSIM 40% · Consistency 20%
Public 3 scenes

Full-access development tier with all data visible.

What you get & how to use

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.

Public Leaderboard
# Method Score PSNR SSIM
1 SpectraFormer + gradient 0.769 31.43 0.936
2 XRF-UNet + gradient 0.733 29.06 0.901
3 PnP-BM3D + gradient 0.666 25.98 0.831
4 FP-Quantify + gradient 0.571 21.82 0.682
Spec Ranges (4 parameters)
Parameter Min Max Unit
excitation_energy_drift -0.01 0.02 keV
detector_resolution 126.0 138.0 eV
matrix_absorption 0.97 1.06 -
beam_spot_size 0.8 1.4 um
View Details →
Dev 3 scenes

Blind evaluation tier — no ground truth available.

What you get & how to use

What you get: Measurements (y), ideal forward operator (H), and spec ranges only.

How to use: Apply your pipeline from the Public tier. Use consistency as self-check.

What to submit: Reconstructed signals and corrected spec. Scored server-side.

Dev Leaderboard
# Method Score PSNR SSIM
1 SpectraFormer + gradient 0.745 29.86 0.915
2 XRF-UNet + gradient 0.653 25.53 0.818
3 FP-Quantify + gradient 0.574 22.29 0.702
4 PnP-BM3D + gradient 0.564 22.37 0.706
Spec Ranges (4 parameters)
Parameter Min Max Unit
excitation_energy_drift -0.012 0.018 keV
detector_resolution 125.2 137.2 eV
matrix_absorption 0.964 1.054 -
beam_spot_size 0.76 1.36 um
Submit Reconstruction View Details →
Hidden 3 scenes

Fully blind server-side evaluation — no data download.

What you get & how to use

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.

Hidden Leaderboard
# Method Score PSNR SSIM
1 SpectraFormer + gradient 0.703 28.71 0.895
2 XRF-UNet + gradient 0.550 21.5 0.668
3 PnP-BM3D + gradient 0.539 21.38 0.663
4 FP-Quantify + gradient 0.512 20.63 0.629
Spec Ranges (4 parameters)
Parameter Min Max Unit
excitation_energy_drift -0.007 0.023 keV
detector_resolution 127.2 139.2 eV
matrix_absorption 0.979 1.069 -
beam_spot_size 0.86 1.46 um
Submit Algorithm View Details →

Blind Reconstruction Challenge

Challenge

Given measurements with unknown mismatch and spec ranges (not exact params), reconstruct the original signal. A method must be evaluated on all three tiers for a complete score. Scored on a composite metric: 0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖).

Input

Measurements y, ideal forward model H, spec ranges

Output

Reconstructed signal x̂

Spec DAG — Forward Model Pipeline

M → R → D

M Modulation
R Rotation
D Detector

Mismatch Parameters

Symbol Parameter Description Nominal Perturbed
e_e excitation_energy_drift Excitation energy drift (keV) 0.0 0.01
d_r detector_resolution Detector resolution (eV) 130.0 134.0
m_a matrix_absorption Matrix absorption (-) 1.0 1.03
b_s beam_spot_size Beam spot size (um) 1.0 1.2

Credits System

40%
Platform Profit Pool
Revenue allocated to benchmark rewards
30%
Winner Share
Top algorithm receives from pool
$100
Min Withdrawal
Minimum payout threshold
Spec Primitives Reference (11 primitives)
P Propagation

Free-space or medium propagation kernel (Fresnel, Rayleigh-Sommerfeld).

M Mask / Modulation

Spatial or spatio-temporal amplitude modulation (coded aperture, SLM pattern).

Π Projection

Geometric projection operator (Radon transform, fan-beam, cone-beam).

F Fourier Sampling

Sampling in the Fourier / k-space domain (MRI, ptychography).

C Convolution

Shift-invariant convolution with a point-spread function (PSF).

Σ Summation / Integration

Summation along a physical dimension (spectral, temporal, angular).

D Detector

Sensor readout with gain g and noise model η (Gaussian, Poisson, mixed).

S Structured Illumination

Patterned illumination (block, Hadamard, random) applied to the scene.

W Wavelength Dispersion

Spectral dispersion element (prism, grating) with shift α and aperture a.

R Rotation / Motion

Sample or gantry rotation (CT, electron tomography).

Λ Wavelength Selection

Spectral filter or monochromator selecting a wavelength band.